diff --git a/CMakeLists.txt b/CMakeLists.txt
index c2686243e0fee4ddaa329b138ff04c1da3bcc4c4..dda5b08694d2c1ccfccccec4fe53b2bdb053dc91 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -44,7 +44,6 @@ set (CMAKE_CXX_EXTENSIONS OFF)
add_definitions (-DIPPL_LINUX -DIPPL_STRINGSTREAM)
add_definitions (-DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DONT_POOL)
add_definitions (-DIPPL_USE_XDIV_RNG -DPETE_BITWISE_COPY)
-add_definitions (-DIPPL_HAS_TEMPLATED_COMPLEX)
add_definitions (-DIPPL_USE_PARTIAL_SPECIALIZATION)
add_definitions (-Drestrict=__restrict__ -DNOCTAssert)
diff --git a/ippl/src/AppTypes/dcomplex.h b/ippl/src/AppTypes/dcomplex.h
index 8334a2291d479701ca426af0815d52b395d81c89..6a38639f0e4075770af297f7ab1ed3afa137849d 100644
--- a/ippl/src/AppTypes/dcomplex.h
+++ b/ippl/src/AppTypes/dcomplex.h
@@ -23,8 +23,6 @@
// include standard complex header file
#include <complex>
-#ifdef IPPL_HAS_TEMPLATED_COMPLEX
-
// KAI and others have a templated complex class
#ifdef IPPL_USE_SINGLE_PRECISION
typedef std::complex<float> dcomplex;
@@ -32,20 +30,4 @@ typedef std::complex<float> dcomplex;
typedef std::complex<double> dcomplex;
#endif
-typedef std::complex<float> fComplex;
-
-#else
-
-// This assumes that all other compilers have the old non-templated
-// complex type which is like complex<double> in the draft standard.
-typedef std::complex dcomplex;
-
-#endif // IPPL_HAS_TEMPLATED_COMPLEX
-
#endif // DCOMPLEX_H
-
-/***************************************************************************
- * $RCSfile: dcomplex.h,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:24 $
- * IPPL_VERSION_ID: $Id: dcomplex.h,v 1.1.1.1 2003/01/23 07:40:24 adelmann Exp $
- ***************************************************************************/
\ No newline at end of file
diff --git a/ippl/src/FFT/CMakeLists.txt b/ippl/src/FFT/CMakeLists.txt
index f39436f450d3017717f12cdeef055b77c83404eb..da9c15401efc216a3f6c437e79f8340f8cd79e17 100644
--- a/ippl/src/FFT/CMakeLists.txt
+++ b/ippl/src/FFT/CMakeLists.txt
@@ -1,22 +1,20 @@
-SET (_SRCS
- fftpack.cpp
+set (_SRCS
+ fftpack.cpp
)
-SET (_HDRS
- f2c.h
+set (_HDRS
FFTBase.hpp
FFTBase.h
FFT.hpp
FFT.h
fftpack_FFT.h
- SCSL_FFT.h
)
-INCLUDE_DIRECTORIES (
+include_directories (
${CMAKE_CURRENT_SOURCE_DIR}
-)
+ )
-ADD_IPPL_SOURCES (${_SRCS})
-ADD_IPPL_HEADERS (${_HDRS})
+add_ippl_sources (${_SRCS})
+add_ippl_headers (${_HDRS})
-INSTALL (FILES ${_HDRS} DESTINATION include/FFT)
+#install (FILES ${_HDRS} DESTINATION include/FFT)
diff --git a/ippl/src/FFT/FFT.h b/ippl/src/FFT/FFT.h
index d02cf22fd541ea2207a8339599b2a4d5865506b5..7eb051e2334e0f704b8382fd6e78f6a85813c5a4 100644
--- a/ippl/src/FFT/FFT.h
+++ b/ippl/src/FFT/FFT.h
@@ -1,21 +1,23 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
-
-//--------------------------------------------------------------------------
-// Class FFT
-//--------------------------------------------------------------------------
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
+/**
+ The FFT class performs complex-to-complex, real-to-complex, or sine
+ transforms on IPPL Fields. FFT is templated on the type of transform
+ to be performed, the dimensionality of the Field to transform, and the
+ floating-point precision type of the Field (float or double).
+*/
#ifndef IPPL_FFT_FFT_H
#define IPPL_FFT_FFT_H
-// include files
#include "FFT/FFTBase.h"
#ifdef IPPL_DKS
@@ -29,12 +31,6 @@ template <class T, unsigned Dim> class BareField;
template <class T, unsigned Dim> class LField;
-/**
- The FFT class performs complex-to-complex, real-to-complex, or sine
- transforms on IPPL Fields. FFT is templated on the type of transform
- to be performed, the dimensionality of the Field to transform, and the
- floating-point precision type of the Field (float or double).
-*/
/**
Tag classes for CC type of Fourier transforms
@@ -55,7 +51,6 @@ class SineTransform {};
template <class Transform, size_t Dim, class T>
class FFT : public FFTBase<Dim,T> {};
-
/**
complex-to-complex FFT class
*/
@@ -64,239 +59,237 @@ class FFT<CCTransform,Dim,T> : public FFTBase<Dim,T> {
private:
#ifdef IPPL_DKS
- DKSOPAL base;
+ DKSOPAL base;
#endif
public:
- // typedefs
- typedef FieldLayout<Dim> Layout_t;
- typedef std::complex<T> Complex_t;
- typedef BareField<Complex_t,Dim> ComplexField_t;
- typedef LField<Complex_t,Dim> ComplexLField_t;
- typedef typename FFTBase<Dim,T>::Domain_t Domain_t;
-
- /** Create a new FFT object with the given domain for the input Field.
- Specify which dimensions to transform along.
- Optional argument compressTemps indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& cdomain, const bool transformTheseDims[Dim],
- const bool& compressTemps=false);
-
- /**
- Create a new FFT object of type CCTransform, with a
- given domain. Default case of transforming along all dimensions.
- Note this was formerly in the .cpp file, but the IBM linker
- could not find it!
- */
- FFT(const Domain_t& cdomain, const bool& compressTemps=false)
+ typedef FieldLayout<Dim> Layout_t;
+ typedef std::complex<T> Complex_t;
+ typedef BareField<Complex_t,Dim> ComplexField_t;
+ typedef LField<Complex_t,Dim> ComplexLField_t;
+ typedef typename FFTBase<Dim,T>::Domain_t Domain_t;
+
+ /** Create a new FFT object with the given domain for the input Field.
+ Specify which dimensions to transform along.
+ Optional argument compressTemps indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& cdomain, const bool transformTheseDims[Dim],
+ const bool& compressTemps=false);
+
+ /**
+ Create a new FFT object of type CCTransform, with a
+ given domain. Default case of transforming along all dimensions.
+ Note this was formerly in the .cpp file, but the IBM linker
+ could not find it!
+ */
+FFT(const Domain_t& cdomain, const bool& compressTemps=false)
: FFTBase<Dim,T>(FFT<CCTransform,Dim,T>::ccFFT, cdomain,compressTemps) {
- // construct array of axis lengths
- int lengths[Dim];
- size_t d;
- for (d=0; d<Dim; ++d)
- lengths[d] = cdomain[d].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int transformTypes[Dim];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- for (d=0; d<Dim; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all transforms are complex-to-complex
- normFact /= lengths[d];
- }
+ // construct array of axis lengths
+ int lengths[Dim];
+ size_t d;
+ for (d=0; d<Dim; ++d)
+ lengths[d] = cdomain[d].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int transformTypes[Dim];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ for (d=0; d<Dim; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all transforms are complex-to-complex
+ normFact /= lengths[d];
+ }
#ifdef IPPL_DKS
#ifdef IPPL_DKS_OPENCL
- INFOMSG("Init DKS base opencl" << endl);
- base.setAPI("OpenCL", 6);
- base.setDevice("-gpu", 4);
- base.initDevice();
+ INFOMSG("Init DKS base opencl" << endl);
+ base.setAPI("OpenCL", 6);
+ base.setDevice("-gpu", 4);
+ base.initDevice();
#endif
#ifdef IPPL_DKS_CUDA
- INFOMSG("Init DKS base cuda" << endl);
- base.setAPI("Cuda", 4);
- base.setDevice("-gpu", 4);
- base.initDevice();
+ INFOMSG("Init DKS base cuda" << endl);
+ base.setAPI("Cuda", 4);
+ base.setDevice("-gpu", 4);
+ base.initDevice();
#endif
#ifdef IPPL_DKS_MIC
- INFOMSG("Init DKS base MIC" << endl);
- base.setAPI("OpenMP", 6);
- base.setDevice("-mic", 4);
- base.initDevice();
+ INFOMSG("Init DKS base MIC" << endl);
+ base.setAPI("OpenMP", 6);
+ base.setDevice("-mic", 4);
+ base.initDevice();
#endif
#endif
-
- // set up FFT Engine
- this->getEngine().setup(Dim, transformTypes, lengths);
- // set up the temporary fields
- setup();
- }
-
-
- // Destructor
- ~FFT(void);
-
- /** Do the FFT: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- User provides separate input and output fields
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, ComplexField_t& f, ComplexField_t& g,
- const bool& constInput=false);
- /**
- invoke using string for direction name
- */
- void transform(const char* directionName, ComplexField_t& f,
- ComplexField_t& g, const bool& constInput=false);
-
- /** overloaded versions which perform the FFT "in place"
- */
- void transform(int direction, ComplexField_t& f);
-
- void transform(const char* directionName, ComplexField_t& f) {
- // invoke in-place transform function using direction name string
- int direction = this->getDirection(directionName);
-
- // Check domain of incoming Field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
-
- // Common loop iterate and other vars:
- size_t d;
- int idim; // idim loops over the number of transform dims.
- int begdim, enddim; // beginning and end of transform dim loop
- size_t nTransformDims = this->numTransformDims();
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
-
- // Loop over the dimensions be transformed:
- begdim = (direction == +1) ? 0 : (nTransformDims-1);
- enddim = (direction == +1) ? nTransformDims : -1;
- for (idim = begdim; idim != enddim; idim += direction) {
+ // set up FFT Engine
+ this->getEngine().setup(Dim, transformTypes, lengths);
+ // set up the temporary fields
+ setup();
+ }
+
+
+ // Destructor
+ ~FFT(void);
+
+ /** Do the FFT: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ User provides separate input and output fields
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, ComplexField_t& f, ComplexField_t& g,
+ const bool& constInput=false);
+ /**
+ invoke using string for direction name
+ */
+ void transform(const char* directionName, ComplexField_t& f,
+ ComplexField_t& g, const bool& constInput=false);
+
+ /** overloaded versions which perform the FFT "in place"
+ */
+ void transform(int direction, ComplexField_t& f);
+
+ void transform(const char* directionName, ComplexField_t& f) {
+ // invoke in-place transform function using direction name string
+ int direction = this->getDirection(directionName);
+
+ // Check domain of incoming Field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int idim; // idim loops over the number of transform dims.
+ int begdim, enddim; // beginning and end of transform dim loop
+ size_t nTransformDims = this->numTransformDims();
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
+
+ // Loop over the dimensions be transformed:
+ begdim = (direction == +1) ? 0 : (nTransformDims-1);
+ enddim = (direction == +1) ? nTransformDims : -1;
+ for (idim = begdim; idim != enddim; idim += direction) {
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == begdim) {
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into f
- if (idim == enddim-direction) {
- // get domain for comparison
- const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
- (in_dom[0].length() == last_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
- }
- else if (idim == enddim-direction && temp != &f) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using f instead
-
- // transpose and permute to Field with transform dim first
- f[in_dom] = (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *temp = 0;
- temp = &f; // Field* management aid
- }
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == begdim) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into f
+ if (idim == enddim-direction) {
+ // get domain for comparison
+ const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
+ (in_dom[0].length() == last_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+ else if (idim == enddim-direction && temp != &f) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using f instead
+
+ // transpose and permute to Field with transform dim first
+ f[in_dom] = (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *temp = 0;
+ temp = &f; // Field* management aid
+ }
- // Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(idim, direction, localdata);
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
- } // loop over all the LFields
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(idim, direction, localdata);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
- } // loop over all transformed dimensions
+ } // loop over all transformed dimensions
- // skip final assignment and compress if we used f as final temporary
- if (temp != &f) {
+ // skip final assignment and compress if we used f as final temporary
+ if (temp != &f) {
- // Now assign back into original Field, and compress last temp's storage:
- f[in_dom] = (*temp)[temp->getLayout().getDomain()];
- if (this->compressTemps()) *temp = 0;
+ // Now assign back into original Field, and compress last temp's storage:
+ f[in_dom] = (*temp)[temp->getLayout().getDomain()];
+ if (this->compressTemps()) *temp = 0;
- }
+ }
- // Normalize:
- if (direction == +1)
- f *= Complex_t(this->getNormFact(), 0.0);
- return;
- }
+ // Normalize:
+ if (direction == +1)
+ f *= Complex_t(this->getNormFact(), 0.0);
+ return;
+ }
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
-
- /**
- How the temporary field's are laid out; these are computed from the
- input Field's domain. This will be allocated as an array of FieldLayouts
- with nTransformDims elements. Each is SERIAL along the zeroth dimension
- and the axes are permuted so that the transform direction is first
- */
- Layout_t** tempLayouts_m;
-
- /** The array of temporary fields, one for each transform direction
- These use the corresponding tempLayouts.
- */
- ComplexField_t** tempFields_m;
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
+
+ /**
+ How the temporary field's are laid out; these are computed from the
+ input Field's domain. This will be allocated as an array of FieldLayouts
+ with nTransformDims elements. Each is SERIAL along the zeroth dimension
+ and the axes are permuted so that the transform direction is first
+ */
+ Layout_t** tempLayouts_m;
+
+ /** The array of temporary fields, one for each transform direction
+ These use the corresponding tempLayouts.
+ */
+ ComplexField_t** tempFields_m;
};
@@ -307,14 +300,14 @@ private:
template <size_t Dim, class T>
inline void
FFT<CCTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& f,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<CCTransform,Dim,T>::ComplexField_t& f,
+ typename FFT<CCTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
@@ -326,71 +319,71 @@ class FFT<CCTransform,1U,T> : public FFTBase<1U,T> {
public:
- // typedefs
- typedef FieldLayout<1U> Layout_t;
- typedef std::complex<T> Complex_t;
- typedef BareField<Complex_t,1U> ComplexField_t;
- typedef LField<Complex_t,1U> ComplexLField_t;
- typedef typename FFTBase<1U,T>::Domain_t Domain_t;
-
- // Constructors:
-
- /** Create a new FFT object with the given domain for the input Field.
- Specify which dimensions to transform along.
- Optional argument compressTemps indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& cdomain, const bool transformTheseDims[1U],
- const bool& compressTemps=false);
- /** Create a new FFT object with the given domain for the input Field.
- Transform along all dimensions.
- Optional argument compressTemps indicates whether or not to compress
- temporary Fields in between uses.
-
- */
- FFT(const Domain_t& cdomain, const bool& compressTemps=false);
-
- // Destructor
- ~FFT(void);
-
- /** Do the FFT: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- User provides separate input and output fields
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, ComplexField_t& f, ComplexField_t& g,
- const bool& constInput=false);
- /**
- invoke using string for direction name
- */
- void transform(const char* directionName, ComplexField_t& f,
- ComplexField_t& g, const bool& constInput=false);
-
- /**
- overloaded versions which perform the FFT "in place"
- */
- void transform(int direction, ComplexField_t& f);
- void transform(const char* directionName, ComplexField_t& f);
+ // typedefs
+ typedef FieldLayout<1U> Layout_t;
+ typedef std::complex<T> Complex_t;
+ typedef BareField<Complex_t,1U> ComplexField_t;
+ typedef LField<Complex_t,1U> ComplexLField_t;
+ typedef typename FFTBase<1U,T>::Domain_t Domain_t;
+
+ // Constructors:
+
+ /** Create a new FFT object with the given domain for the input Field.
+ Specify which dimensions to transform along.
+ Optional argument compressTemps indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& cdomain, const bool transformTheseDims[1U],
+ const bool& compressTemps=false);
+ /** Create a new FFT object with the given domain for the input Field.
+ Transform along all dimensions.
+ Optional argument compressTemps indicates whether or not to compress
+ temporary Fields in between uses.
+
+ */
+ FFT(const Domain_t& cdomain, const bool& compressTemps=false);
+
+ // Destructor
+ ~FFT(void);
+
+ /** Do the FFT: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ User provides separate input and output fields
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, ComplexField_t& f, ComplexField_t& g,
+ const bool& constInput=false);
+ /**
+ invoke using string for direction name
+ */
+ void transform(const char* directionName, ComplexField_t& f,
+ ComplexField_t& g, const bool& constInput=false);
+
+ /**
+ overloaded versions which perform the FFT "in place"
+ */
+ void transform(int direction, ComplexField_t& f);
+ void transform(const char* directionName, ComplexField_t& f);
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
- /**
- The temporary field layout
- */
- Layout_t* tempLayouts_m;
+ /**
+ The temporary field layout
+ */
+ Layout_t* tempLayouts_m;
- /**
- The temporary field
- */
- ComplexField_t* tempFields_m;
+ /**
+ The temporary field
+ */
+ ComplexField_t* tempFields_m;
};
@@ -403,14 +396,14 @@ private:
template <class T>
inline void
FFT<CCTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<CCTransform,1U,T>::ComplexField_t& f,
- typename FFT<CCTransform,1U,T>::ComplexField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& f,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -419,12 +412,12 @@ FFT<CCTransform,1U,T>::transform(
template <class T>
inline void
FFT<CCTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<CCTransform,1U,T>::ComplexField_t& f)
+ const char* directionName,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& f)
{
- int dir = this->getDirection(directionName);
- transform(dir, f);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f);
+ return;
}
@@ -438,109 +431,109 @@ private:
public:
- // typedefs
- typedef FieldLayout<Dim> Layout_t;
- typedef BareField<T,Dim> RealField_t;
- typedef LField<T,Dim> RealLField_t;
- typedef std::complex<T> Complex_t;
- typedef BareField<Complex_t,Dim> ComplexField_t;
- typedef LField<Complex_t,Dim> ComplexLField_t;
- typedef typename FFTBase<Dim, T>::Domain_t Domain_t;
-
- // Constructors:
-
- /** Create a new FFT object with the given domains for input/output Fields
- Specify which dimensions to transform along.
- Optional argument compress indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool transformTheseDims[Dim], const bool& compressTemps=false);
-
- /**
- Same as above, but transform all dims:
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool& compressTemps=false, int serialAxes = 1);
-
- // Destructor
- ~FFT(void);
-
- /** real-to-complex FFT: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- Supply a second BareField to store the output.
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, RealField_t& f, ComplexField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, RealField_t& f,
- ComplexField_t& g, const bool& constInput=false);
-
- /** real-to-complex FFT on GPU: transfer the real field to GPU execute FFT
- return the pointer to memory on GPU where complex results are stored
- */
+ // typedefs
+ typedef FieldLayout<Dim> Layout_t;
+ typedef BareField<T,Dim> RealField_t;
+ typedef LField<T,Dim> RealLField_t;
+ typedef std::complex<T> Complex_t;
+ typedef BareField<Complex_t,Dim> ComplexField_t;
+ typedef LField<Complex_t,Dim> ComplexLField_t;
+ typedef typename FFTBase<Dim, T>::Domain_t Domain_t;
+
+ // Constructors:
+
+ /** Create a new FFT object with the given domains for input/output Fields
+ Specify which dimensions to transform along.
+ Optional argument compress indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool transformTheseDims[Dim], const bool& compressTemps=false);
+
+ /**
+ Same as above, but transform all dims:
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool& compressTemps=false, int serialAxes = 1);
+
+ // Destructor
+ ~FFT(void);
+
+ /** real-to-complex FFT: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ Supply a second BareField to store the output.
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, RealField_t& f, ComplexField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, RealField_t& f,
+ ComplexField_t& g, const bool& constInput=false);
+
+ /** real-to-complex FFT on GPU: transfer the real field to GPU execute FFT
+ return the pointer to memory on GPU where complex results are stored
+ */
#ifdef IPPL_DKS
- void transformDKSRC(int direction, RealField_t &f, void* real_ptr, void* comp_ptr,
- DKSOPAL &dksbase, int streamId = -1, const bool& constInput=false);
+ void transformDKSRC(int direction, RealField_t &f, void* real_ptr, void* comp_ptr,
+ DKSOPAL &dksbase, int streamId = -1, const bool& constInput=false);
#endif
- /** complex-to-real FFT
- Same as above, but with input and output field types reversed.
- */
- void transform(int direction, ComplexField_t& f, RealField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, ComplexField_t& f,
- RealField_t& g, const bool& constInput=false);
+ /** complex-to-real FFT
+ Same as above, but with input and output field types reversed.
+ */
+ void transform(int direction, ComplexField_t& f, RealField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, ComplexField_t& f,
+ RealField_t& g, const bool& constInput=false);
- /** complex-to-real FFT on GPU: pass pointer to GPU memory where complex field
- is stored, do the inverse FFT and transfer real field back to host memory
- */
+ /** complex-to-real FFT on GPU: pass pointer to GPU memory where complex field
+ is stored, do the inverse FFT and transfer real field back to host memory
+ */
#ifdef IPPL_DKS
- void transformDKSCR(int direction, RealField_t& g, void* real_ptr, void* comp_ptr,
- DKSOPAL &dksbase, int streamId = -1, const bool& constInput=false);
+ void transformDKSCR(int direction, RealField_t& g, void* real_ptr, void* comp_ptr,
+ DKSOPAL &dksbase, int streamId = -1, const bool& constInput=false);
#endif
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
-
- /** How the temporary fields are laid out; these are computed from the
- input Field's domain. This will be allocated as an array of FieldLayouts
- with nTransformDims elements. Each is SERIAL along the zeroth dimension
- and the axes are permuted so that the transform direction is first
- */
- Layout_t** tempLayouts_m;
-
- /**
- extra layout for the one real Field needed
- */
- Layout_t* tempRLayout_m;
-
- /** The array of temporary fields, one for each transform direction
- These use the corresponding tempLayouts.
- */
- ComplexField_t** tempFields_m;
-
- /**
- We need one real internal Field in this case.
- */
- RealField_t* tempRField_m;
-
- /**
- domain of the resulting complex fields
- const Domain_t& complexDomain_m;
- */
- Domain_t complexDomain_m;
-
- /**
- number of axes to make serial
- */
- int serialAxes_m;
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
+
+ /** How the temporary fields are laid out; these are computed from the
+ input Field's domain. This will be allocated as an array of FieldLayouts
+ with nTransformDims elements. Each is SERIAL along the zeroth dimension
+ and the axes are permuted so that the transform direction is first
+ */
+ Layout_t** tempLayouts_m;
+
+ /**
+ extra layout for the one real Field needed
+ */
+ Layout_t* tempRLayout_m;
+
+ /** The array of temporary fields, one for each transform direction
+ These use the corresponding tempLayouts.
+ */
+ ComplexField_t** tempFields_m;
+
+ /**
+ We need one real internal Field in this case.
+ */
+ RealField_t* tempRField_m;
+
+ /**
+ domain of the resulting complex fields
+ const Domain_t& complexDomain_m;
+ */
+ Domain_t complexDomain_m;
+
+ /**
+ number of axes to make serial
+ */
+ int serialAxes_m;
};
// Inline function definitions
@@ -551,14 +544,14 @@ private:
template <size_t Dim, class T>
inline void
FFT<RCTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<RCTransform,Dim,T>::RealField_t& f,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<RCTransform,Dim,T>::RealField_t& f,
+ typename FFT<RCTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -567,14 +560,14 @@ FFT<RCTransform,Dim,T>::transform(
template <size_t Dim, class T>
inline void
FFT<RCTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& f,
- typename FFT<RCTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<RCTransform,Dim,T>::ComplexField_t& f,
+ typename FFT<RCTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
@@ -586,87 +579,87 @@ class FFT<RCTransform,1U,T> : public FFTBase<1U,T> {
public:
- // typedefs
- typedef FieldLayout<1U> Layout_t;
- typedef BareField<T,1U> RealField_t;
- typedef LField<T,1U> RealLField_t;
- typedef std::complex<T> Complex_t;
- typedef BareField<Complex_t,1U> ComplexField_t;
- typedef LField<Complex_t,1U> ComplexLField_t;
- typedef typename FFTBase<1U,T>::Domain_t Domain_t;
-
- // Constructors:
-
- /**
- Create a new FFT object with the given domains for input/output Fields
- Specify which dimensions to transform along.
- Optional argument compress indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool transformTheseDims[1U], const bool& compressTemps=false);
- /**
- Same as above, but transform all dims:
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool& compressTemps=false);
-
- /**
- Destructor
- */
- ~FFT(void);
-
- /**
- real-to-complex FFT: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- Supply a second BareField to store the output.
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, RealField_t& f, ComplexField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, RealField_t& f,
- ComplexField_t& g, const bool& constInput=false);
-
- /**
- complex-to-real FFT
- Same as above, but with input and output field types reversed.
- */
- void transform(int direction, ComplexField_t& f, RealField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, ComplexField_t& f,
- RealField_t& g, const bool& constInput=false);
+ // typedefs
+ typedef FieldLayout<1U> Layout_t;
+ typedef BareField<T,1U> RealField_t;
+ typedef LField<T,1U> RealLField_t;
+ typedef std::complex<T> Complex_t;
+ typedef BareField<Complex_t,1U> ComplexField_t;
+ typedef LField<Complex_t,1U> ComplexLField_t;
+ typedef typename FFTBase<1U,T>::Domain_t Domain_t;
+
+ // Constructors:
+
+ /**
+ Create a new FFT object with the given domains for input/output Fields
+ Specify which dimensions to transform along.
+ Optional argument compress indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool transformTheseDims[1U], const bool& compressTemps=false);
+ /**
+ Same as above, but transform all dims:
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool& compressTemps=false);
+
+ /**
+ Destructor
+ */
+ ~FFT(void);
+
+ /**
+ real-to-complex FFT: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ Supply a second BareField to store the output.
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, RealField_t& f, ComplexField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, RealField_t& f,
+ ComplexField_t& g, const bool& constInput=false);
+
+ /**
+ complex-to-real FFT
+ Same as above, but with input and output field types reversed.
+ */
+ void transform(int direction, ComplexField_t& f, RealField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, ComplexField_t& f,
+ RealField_t& g, const bool& constInput=false);
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
-
- /**
- The temporary field layout
- */
- Layout_t* tempLayouts_m;
-
- /**
- The temporary field
- */
- ComplexField_t* tempFields_m;
-
- /**
- Real field layout
- */
- Layout_t* tempRLayout_m;
-
- /**
- We need one real internal Field in this case.
- domain of the resulting complex fields
- */
- // const Domain_t& complexDomain_m;
- Domain_t complexDomain_m;
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
+
+ /**
+ The temporary field layout
+ */
+ Layout_t* tempLayouts_m;
+
+ /**
+ The temporary field
+ */
+ ComplexField_t* tempFields_m;
+
+ /**
+ Real field layout
+ */
+ Layout_t* tempRLayout_m;
+
+ /**
+ We need one real internal Field in this case.
+ domain of the resulting complex fields
+ */
+ // const Domain_t& complexDomain_m;
+ Domain_t complexDomain_m;
};
/**
@@ -675,14 +668,14 @@ private:
template <class T>
inline void
FFT<RCTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<RCTransform,1U,T>::RealField_t& f,
- typename FFT<RCTransform,1U,T>::ComplexField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<RCTransform,1U,T>::RealField_t& f,
+ typename FFT<RCTransform,1U,T>::ComplexField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -691,14 +684,14 @@ FFT<RCTransform,1U,T>::transform(
template <class T>
inline void
FFT<RCTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<RCTransform,1U,T>::ComplexField_t& f,
- typename FFT<RCTransform,1U,T>::RealField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<RCTransform,1U,T>::ComplexField_t& f,
+ typename FFT<RCTransform,1U,T>::RealField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -709,138 +702,138 @@ class FFT<SineTransform,Dim,T> : public FFTBase<Dim,T> {
public:
- // typedefs
- typedef FieldLayout<Dim> Layout_t;
- typedef BareField<T,Dim> RealField_t;
- typedef LField<T,Dim> RealLField_t;
- typedef std::complex<T> Complex_t;
- typedef BareField<Complex_t,Dim> ComplexField_t;
- typedef LField<Complex_t,Dim> ComplexLField_t;
- typedef typename FFTBase<Dim,T>::Domain_t Domain_t;
-
- /** Constructor for doing sine transform(s) followed by RC FFT
- Create a new FFT object with the given domains for input/output Fields
- Specify which dimensions to transform along.
- Also specify which of these are sine transforms
- Optional argument compress indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool transformTheseDims[Dim],
- const bool sineTransformDims[Dim], const bool& compressTemps=false);
- /**
- Same as above, but transform all dims:
- */
- FFT(const Domain_t& rdomain, const Domain_t& cdomain,
- const bool sineTransformDims[Dim], const bool& compressTemps=false);
- /**
- Separate constructors for doing only sine transforms
- Create a new FFT object with the given domain for input/output Field
- Specify which dimensions to transform along.
- Optional argument compress indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& rdomain, const bool sineTransformDims[Dim],
- const bool& compressTemps=false);
- /**
- Same as above, but transform all dims:
- */
- FFT(const Domain_t& rdomain, const bool& compressTemps=false);
-
- ~FFT(void);
-
- /**
- These transforms are for combinations of sine transforms and RC FFTs
-
- Do the FFT: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- Supply a second BareField to store the output.
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, RealField_t& f, ComplexField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, RealField_t& f,
- ComplexField_t& g, const bool& constInput=false);
-
- /**
- complex-to-real FFT, followed by sine transform(s)
- Same as above, but with input and output field types reversed.
- */
- void transform(int direction, ComplexField_t& f, RealField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, ComplexField_t& f,
- RealField_t& g, const bool& constInput=false);
-
- /**
- These transforms are for doing sine transforms only
- sine transform: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- Supply a second BareField to store the output.
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, RealField_t& f, RealField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, RealField_t& f,
- RealField_t& g, const bool& constInput=false);
-
- /**
- In-place version of real-to-real transform
- */
- void transform(int direction, RealField_t& f);
- void transform(const char* directionName, RealField_t& f);
+ // typedefs
+ typedef FieldLayout<Dim> Layout_t;
+ typedef BareField<T,Dim> RealField_t;
+ typedef LField<T,Dim> RealLField_t;
+ typedef std::complex<T> Complex_t;
+ typedef BareField<Complex_t,Dim> ComplexField_t;
+ typedef LField<Complex_t,Dim> ComplexLField_t;
+ typedef typename FFTBase<Dim,T>::Domain_t Domain_t;
+
+ /** Constructor for doing sine transform(s) followed by RC FFT
+ Create a new FFT object with the given domains for input/output Fields
+ Specify which dimensions to transform along.
+ Also specify which of these are sine transforms
+ Optional argument compress indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool transformTheseDims[Dim],
+ const bool sineTransformDims[Dim], const bool& compressTemps=false);
+ /**
+ Same as above, but transform all dims:
+ */
+ FFT(const Domain_t& rdomain, const Domain_t& cdomain,
+ const bool sineTransformDims[Dim], const bool& compressTemps=false);
+ /**
+ Separate constructors for doing only sine transforms
+ Create a new FFT object with the given domain for input/output Field
+ Specify which dimensions to transform along.
+ Optional argument compress indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& rdomain, const bool sineTransformDims[Dim],
+ const bool& compressTemps=false);
+ /**
+ Same as above, but transform all dims:
+ */
+ FFT(const Domain_t& rdomain, const bool& compressTemps=false);
+
+ ~FFT(void);
+
+ /**
+ These transforms are for combinations of sine transforms and RC FFTs
+
+ Do the FFT: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ Supply a second BareField to store the output.
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, RealField_t& f, ComplexField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, RealField_t& f,
+ ComplexField_t& g, const bool& constInput=false);
+
+ /**
+ complex-to-real FFT, followed by sine transform(s)
+ Same as above, but with input and output field types reversed.
+ */
+ void transform(int direction, ComplexField_t& f, RealField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, ComplexField_t& f,
+ RealField_t& g, const bool& constInput=false);
+
+ /**
+ These transforms are for doing sine transforms only
+ sine transform: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ Supply a second BareField to store the output.
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, RealField_t& f, RealField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, RealField_t& f,
+ RealField_t& g, const bool& constInput=false);
+
+ /**
+ In-place version of real-to-real transform
+ */
+ void transform(int direction, RealField_t& f);
+ void transform(const char* directionName, RealField_t& f);
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
- /**
- which dimensions are sine transformed
- */
- bool sineTransformDims_m[Dim];
+ /**
+ which dimensions are sine transformed
+ */
+ bool sineTransformDims_m[Dim];
- /**
- number of sine transforms to perform
- */
- size_t numSineTransforms_m;
+ /**
+ number of sine transforms to perform
+ */
+ size_t numSineTransforms_m;
- /**
- layouts for temporary Fields: SERIAL along the zeroth dimension, with
- the axes are permuted so that the transform direction is first
+ /**
+ layouts for temporary Fields: SERIAL along the zeroth dimension, with
+ the axes are permuted so that the transform direction is first
- layouts for the temporary complex Fields
- */
+ layouts for the temporary complex Fields
+ */
- Layout_t** tempLayouts_m;
+ Layout_t** tempLayouts_m;
- /**
- layouts for the temporary real Fields
- */
- Layout_t** tempRLayouts_m;
+ /**
+ layouts for the temporary real Fields
+ */
+ Layout_t** tempRLayouts_m;
- /** The array of temporary complex Fields
- These use the corresponding tempLayouts.
- */
- ComplexField_t** tempFields_m;
+ /** The array of temporary complex Fields
+ These use the corresponding tempLayouts.
+ */
+ ComplexField_t** tempFields_m;
- /** The array of temporary real Fields
- These use the corresponding tempRLayouts.
- */
- RealField_t** tempRFields_m;
+ /** The array of temporary real Fields
+ These use the corresponding tempRLayouts.
+ */
+ RealField_t** tempRFields_m;
- /**
- domain of the resulting complex Field for real-to-complex transform
- */
- const Domain_t* complexDomain_m;
+ /**
+ domain of the resulting complex Field for real-to-complex transform
+ */
+ const Domain_t* complexDomain_m;
};
/**
@@ -849,14 +842,14 @@ private:
template <size_t Dim, class T>
inline void
FFT<SineTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,Dim,T>::RealField_t& f,
- typename FFT<SineTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<SineTransform,Dim,T>::RealField_t& f,
+ typename FFT<SineTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -865,14 +858,14 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
inline void
FFT<SineTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,Dim,T>::ComplexField_t& f,
- typename FFT<SineTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<SineTransform,Dim,T>::ComplexField_t& f,
+ typename FFT<SineTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -881,14 +874,14 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
inline void
FFT<SineTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,Dim,T>::RealField_t& f,
- typename FFT<SineTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<SineTransform,Dim,T>::RealField_t& f,
+ typename FFT<SineTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -897,12 +890,12 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
inline void
FFT<SineTransform,Dim,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,Dim,T>::RealField_t& f)
+ const char* directionName,
+ typename FFT<SineTransform,Dim,T>::RealField_t& f)
{
- int dir = this->getDirection(directionName);
- transform(dir, f);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f);
+ return;
}
@@ -914,64 +907,64 @@ class FFT<SineTransform,1U,T> : public FFTBase<1U,T> {
public:
- typedef FieldLayout<1U> Layout_t;
- typedef BareField<T,1U> RealField_t;
- typedef LField<T,1U> RealLField_t;
- typedef typename FFTBase<1U,T>::Domain_t Domain_t;
-
- /**
- Constructors for doing only sine transforms
- Create a new FFT object with the given domain for input/output Field
- specify which dimensions to transform along.
- Optional argument compress indicates whether or not to compress
- temporary Fields in between uses.
- */
- FFT(const Domain_t& rdomain, const bool sineTransformDims[1U],
- const bool& compressTemps=false);
- /**
- Same as above, but transform all dims:
- */
- FFT(const Domain_t& rdomain, const bool& compressTemps=false);
-
-
- ~FFT(void);
-
- /**
- sine transform: specify +1 or -1 to indicate forward or inverse
- transform, or specify the user-defined name string for the direction.
- Supply a second BareField to store the output.
- optional argument constInput indicates whether or not to treat the
- input Field argument f as const. If not, we can use it as a temporary
- in order to avoid an additional data transpose.
- */
- void transform(int direction, RealField_t& f, RealField_t& g,
- const bool& constInput=false);
- void transform(const char* directionName, RealField_t& f,
- RealField_t& g, const bool& constInput=false);
-
- /**
- In-place version of real-to-real transform
- */
- void transform(int direction, RealField_t& f);
- void transform(const char* directionName, RealField_t& f);
+ typedef FieldLayout<1U> Layout_t;
+ typedef BareField<T,1U> RealField_t;
+ typedef LField<T,1U> RealLField_t;
+ typedef typename FFTBase<1U,T>::Domain_t Domain_t;
+
+ /**
+ Constructors for doing only sine transforms
+ Create a new FFT object with the given domain for input/output Field
+ specify which dimensions to transform along.
+ Optional argument compress indicates whether or not to compress
+ temporary Fields in between uses.
+ */
+ FFT(const Domain_t& rdomain, const bool sineTransformDims[1U],
+ const bool& compressTemps=false);
+ /**
+ Same as above, but transform all dims:
+ */
+ FFT(const Domain_t& rdomain, const bool& compressTemps=false);
+
+
+ ~FFT(void);
+
+ /**
+ sine transform: specify +1 or -1 to indicate forward or inverse
+ transform, or specify the user-defined name string for the direction.
+ Supply a second BareField to store the output.
+ optional argument constInput indicates whether or not to treat the
+ input Field argument f as const. If not, we can use it as a temporary
+ in order to avoid an additional data transpose.
+ */
+ void transform(int direction, RealField_t& f, RealField_t& g,
+ const bool& constInput=false);
+ void transform(const char* directionName, RealField_t& f,
+ RealField_t& g, const bool& constInput=false);
+
+ /**
+ In-place version of real-to-real transform
+ */
+ void transform(int direction, RealField_t& f);
+ void transform(const char* directionName, RealField_t& f);
private:
- /**
- setup performs all the initializations necessary after the transform
- directions have been specified.
- */
- void setup(void);
+ /**
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
+ */
+ void setup(void);
- /**
- The temporary real Field layout
- */
- Layout_t* tempRLayouts_m;
+ /**
+ The temporary real Field layout
+ */
+ Layout_t* tempRLayouts_m;
- /**
- The temporary real Field
- */
- RealField_t* tempRFields_m;
+ /**
+ The temporary real Field
+ */
+ RealField_t* tempRFields_m;
};
@@ -981,14 +974,14 @@ private:
template <class T>
inline void
FFT<SineTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,1U,T>::RealField_t& f,
- typename FFT<SineTransform,1U,T>::RealField_t& g,
- const bool& constInput)
+ const char* directionName,
+ typename FFT<SineTransform,1U,T>::RealField_t& f,
+ typename FFT<SineTransform,1U,T>::RealField_t& g,
+ const bool& constInput)
{
- int dir = this->getDirection(directionName);
- transform(dir, f, g, constInput);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f, g, constInput);
+ return;
}
/**
@@ -997,20 +990,19 @@ FFT<SineTransform,1U,T>::transform(
template <class T>
inline void
FFT<SineTransform,1U,T>::transform(
- const char* directionName,
- typename FFT<SineTransform,1U,T>::RealField_t& f)
+ const char* directionName,
+ typename FFT<SineTransform,1U,T>::RealField_t& f)
{
- int dir = this->getDirection(directionName);
- transform(dir, f);
- return;
+ int dir = this->getDirection(directionName);
+ transform(dir, f);
+ return;
}
#include "FFT/FFT.hpp"
#endif // IPPL_FFT_FFT_H
-/***************************************************************************
- * $RCSfile: FFT.h,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
- * IPPL_VERSION_ID: $Id: FFT.h,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
- ***************************************************************************/
-
-
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/FFT.hpp b/ippl/src/FFT/FFT.hpp
index 4c29efba63ad1d1d27867d1461a5dc7bc6af4318..a7a422ba57a5c6d2079d42cead8bed6595d6dea3 100644
--- a/ippl/src/FFT/FFT.hpp
+++ b/ippl/src/FFT/FFT.hpp
@@ -1,28 +1,17 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- * This program was prepared by PSI.
- * All rights in the program are reserved by PSI.
- * Neither PSI nor the author(s)
- * makes any warranty, express or implied, or assumes any liability or
- * responsibility for the use of this software
- *
- *
- ***************************************************************************/
-
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
-
-// include files
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
+/**
+ Implementations for FFT constructor/destructor and transforms
+*/
+
#include "FFT/FFT.h"
#include "FieldLayout/FieldLayout.h"
#include "Field/BareField.h"
@@ -34,129 +23,96 @@
#define FFTDBG(x)
#endif
-
-/**
- FFT.cpp: implementations for FFT constructor/destructor and transforms
-*/
-
-
//=============================================================================
// FFT CCTransform Constructors
//=============================================================================
/**
- Create a new FFT object of type CCTransform, with a
- given domain. Also specify which dimensions to transform along.
+ Create a new FFT object of type CCTransform, with a
+ given domain. Also specify which dimensions to transform along.
*/
template <size_t Dim, class T>
FFT<CCTransform,Dim,T>::FFT(
- const typename FFT<CCTransform,Dim,T>::Domain_t& cdomain,
- const bool transformTheseDims[Dim],
- const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<CCTransform,Dim,T>::ccFFT, cdomain,
- transformTheseDims, compressTemps)
+ const typename FFT<CCTransform,Dim,T>::Domain_t& cdomain,
+ const bool transformTheseDims[Dim],
+ const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<CCTransform,Dim,T>::ccFFT, cdomain,
+ transformTheseDims, compressTemps)
{
-/*
-#ifdef IPPL_DKS
-#ifdef IPPL_DKS_OPENCL
- INFOMSG("Init DKS base opencl" << endl);
- base.setAPI("OpenCL", 6);
- base.setDevice("-gpu", 4);
- base.initDevice();
-
-#endif
-
-#ifdef IPPL_DKS_CUDA
- INFOMSG("Init DKS base cuda" << endl);
- base.setAPI("Cuda", 4);
- base.setDevice("-gpu", 4);
- base.initDevice();
-#endif
-
-#ifdef IPPL_DKS_MIC
- INFOMSG("Init DKS base MIC" << endl);
- base.setAPI("OpenMP", 6);
- base.setDevice("-mic", 4);
- base.initDevice();
-#endif
-#endif
-*/
-
- // construct array of axis lengths
- size_t nTransformDims = this->numTransformDims();
- int* lengths = new int[nTransformDims];
- size_t d;
- for (d=0; d<nTransformDims; ++d)
- lengths[d] = cdomain[this->activeDimension(d)].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int* transformTypes = new int[nTransformDims];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- for (d=0; d<nTransformDims; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all transforms are complex-to-complex
- normFact /= lengths[d];
- }
-
- // set up FFT Engine
- this->getEngine().setup(nTransformDims, transformTypes, lengths);
- delete [] transformTypes;
- delete [] lengths;
- // set up the temporary fields
- setup();
+ // construct array of axis lengths
+ size_t nTransformDims = this->numTransformDims();
+ int* lengths = new int[nTransformDims];
+ size_t d;
+ for (d=0; d<nTransformDims; ++d)
+ lengths[d] = cdomain[this->activeDimension(d)].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int* transformTypes = new int[nTransformDims];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ for (d=0; d<nTransformDims; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all transforms are complex-to-complex
+ normFact /= lengths[d];
+ }
+
+ // set up FFT Engine
+ this->getEngine().setup(nTransformDims, transformTypes, lengths);
+ delete [] transformTypes;
+ delete [] lengths;
+ // set up the temporary fields
+ setup();
}
-
/**
- setup performs all the initializations necessary after the transform
- directions have been specified.
+ setup performs all the initializations necessary after the transform
+ directions have been specified.
*/
template <size_t Dim, class T>
void
FFT<CCTransform,Dim,T>::setup(void)
{
- // Tau profiling
-
-
- size_t d, activeDim;
- size_t nTransformDims = this->numTransformDims();
- // Set up the arrays of temporary Fields and FieldLayouts:
- e_dim_tag serialParallel[Dim]; // Specifies SERIAL, PARALLEL dims in temp
- // make zeroth dimension always SERIAL
- serialParallel[0] = SERIAL;
- // all other dimensions parallel
- for (d=1; d<Dim; ++d)
- serialParallel[d] = PARALLEL;
-
- tempLayouts_m = new Layout_t*[nTransformDims];
- tempFields_m = new ComplexField_t*[nTransformDims];
-
- // loop over transform dimensions
- for (size_t dim=0; dim<nTransformDims; ++dim) {
- // get number of dimension to be transformed
- activeDim = this->activeDimension(dim);
- // Get input Field's domain
- const Domain_t& ndic = this->getDomain();
- // make new domain with permuted Indexes, activeDim first
- Domain_t ndip;
- ndip[0] = ndic[activeDim];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = activeDim + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = ndic[nextDim];
- }
- // generate temporary field layout
- tempLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
- // generate temporary Field
- tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) (*tempFields_m[dim]).Uncompress();
- }
+ // Tau profiling
+
- return;
+ size_t d, activeDim;
+ size_t nTransformDims = this->numTransformDims();
+ // Set up the arrays of temporary Fields and FieldLayouts:
+ e_dim_tag serialParallel[Dim]; // Specifies SERIAL, PARALLEL dims in temp
+ // make zeroth dimension always SERIAL
+ serialParallel[0] = SERIAL;
+ // all other dimensions parallel
+ for (d=1; d<Dim; ++d)
+ serialParallel[d] = PARALLEL;
+
+ tempLayouts_m = new Layout_t*[nTransformDims];
+ tempFields_m = new ComplexField_t*[nTransformDims];
+
+ // loop over transform dimensions
+ for (size_t dim=0; dim<nTransformDims; ++dim) {
+ // get number of dimension to be transformed
+ activeDim = this->activeDimension(dim);
+ // Get input Field's domain
+ const Domain_t& ndic = this->getDomain();
+ // make new domain with permuted Indexes, activeDim first
+ Domain_t ndip;
+ ndip[0] = ndic[activeDim];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = activeDim + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = ndic[nextDim];
+ }
+ // generate temporary field layout
+ tempLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
+ // generate temporary Field
+ tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) (*tempFields_m[dim]).Uncompress();
+ }
+
+ return;
}
//-----------------------------------------------------------------------------
@@ -166,22 +122,22 @@ FFT<CCTransform,Dim,T>::setup(void)
template <size_t Dim, class T>
FFT<CCTransform,Dim,T>::~FFT(void) {
- // Tau profiling
-
- /*
- #ifdef IPPL_OPENCL
- base.ocl_cleanUp();
- #endif
- */
-
- // delete arrays of temporary fields and field layouts
- size_t nTransformDims = this->numTransformDims();
- for (size_t d=0; d<nTransformDims; ++d) {
- delete tempFields_m[d];
- delete tempLayouts_m[d];
- }
- delete [] tempFields_m;
- delete [] tempLayouts_m;
+ // Tau profiling
+
+ /*
+ #ifdef IPPL_OPENCL
+ base.ocl_cleanUp();
+ #endif
+ */
+
+ // delete arrays of temporary fields and field layouts
+ size_t nTransformDims = this->numTransformDims();
+ for (size_t d=0; d<nTransformDims; ++d) {
+ delete tempFields_m[d];
+ delete tempLayouts_m[d];
+ }
+ delete [] tempFields_m;
+ delete [] tempLayouts_m;
}
@@ -192,313 +148,247 @@ FFT<CCTransform,Dim,T>::~FFT(void) {
template <size_t Dim, class T>
void
FFT<CCTransform,Dim,T>::transform(
- int direction,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& f,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<CCTransform,Dim,T>::ComplexField_t& f,
+ typename FFT<CCTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true);
-
- // Common loop iterate and other vars:
- size_t d;
- int idim; // idim loops over the number of transform dims.
- int begdim, enddim; // beginning and end of transform dim loop
- size_t nTransformDims = this->numTransformDims();
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
-
- // Loop over the dimensions be transformed:
- begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
- enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
- for (idim = begdim; idim != enddim; idim += direction) {
-
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == begdim && !constInput) {
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into g
- if (idim == enddim-direction) {
- // get the domain for comparison
- const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
- (out_dom[0].length() == last_dom[0].length()) &&
- (out_layout.getDistribution(0) == SERIAL) &&
- (g.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
- }
- else if (idim == enddim-direction && temp != &g) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using g instead
-
- // transpose and permute to Field with transform dim first
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = &g; // Field* management aid
- }
-
- // Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(idim, direction, localdata);
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
- } // loop over all transformed dimensions
-
- // skip final assignment and compress if we used g as final temporary
- if (temp != &g) {
-
- // Now assign into output Field, and compress last temp's storage:
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- if (this->compressTemps() && temp != &f) *temp = 0;
-
- }
-
- // Normalize:
- if (direction == +1)
- g *= Complex_t(this->getNormFact(), 0.0);
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true);
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int idim; // idim loops over the number of transform dims.
+ int begdim, enddim; // beginning and end of transform dim loop
+ size_t nTransformDims = this->numTransformDims();
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
+
+ // Loop over the dimensions be transformed:
+ begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
+ enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
+ for (idim = begdim; idim != enddim; idim += direction) {
+
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == begdim && !constInput) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into g
+ if (idim == enddim-direction) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
+ (out_dom[0].length() == last_dom[0].length()) &&
+ (out_layout.getDistribution(0) == SERIAL) &&
+ (g.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+ else if (idim == enddim-direction && temp != &g) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using g instead
+
+ // transpose and permute to Field with transform dim first
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = &g; // Field* management aid
+ }
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(idim, direction, localdata);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used g as final temporary
+ if (temp != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
+ if (this->compressTemps() && temp != &f) *temp = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1)
+ g *= Complex_t(this->getNormFact(), 0.0);
- return;
+ return;
}
-//-----------------------------------------------------------------------------
-// "in-place" FFT; specify +1 or -1 to indicate forward or inverse transform.
-//-----------------------------------------------------------------------------
-/*
-#ifdef IPPL_DKS
-template <unsigned Dim, class T>
-void
-FFT<CCTransform, Dim, T>::transform(
- int direction,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& f)
-{
-
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert(this->checkDomain(this->getDomain(),in_dom));
-
- //Field* for temp Field management:
- ComplexField_t* temp = &f;
- //Local work array to get data from ComplexField
- Complex_t* localdata;
-
- //long total_size = in_dom.size();
-
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- l_i = temp->begin_if();
-
- //get the field
- ComplexLField_t* ldf = (*l_i).second.get();
- //make sure we are uncomplressed
- ldf->Uncompress();
- int N[3] = {ldf->size(0), ldf->size(1), ldf->size(2)};
-
- localdata = ldf->getP();
-
- // DKS part
- int ierr;
- void *mem_ptr;
- int size = N[0]*N[1]*N[2];
-
- base.setupFFT(3, N);
- mem_ptr = base.allocateMemory<Complex_t>(size, ierr);
- ierr = base.writeData<Complex_t>(mem_ptr, localdata, size);
-
- if (direction == 1) {
- base.callFFT(mem_ptr, 3, N);
- base.callNormalizeFFT(mem_ptr, 3, N);
- } else {
- base.callIFFT(mem_ptr, 3, N);
- //base.callNormalizeFFT(mem_ptr, 3, N);
- }
-
-
- base.readData<Complex_t>(mem_ptr, localdata, size);
- base.freeMemory<Complex_t>(mem_ptr, size);
-
- //assign back to the original Field
- if (temp != &f) {
- f[in_dom] = (*temp)[temp->getLayout().getDomain()];
- if(this->compressTemps()) *temp = 0;
- }
-
- return;
-}
-#else
-*/
template <size_t Dim, class T>
void
FFT<CCTransform,Dim,T>::transform(
- int direction,
- typename FFT<CCTransform,Dim,T>::ComplexField_t& f)
+ int direction,
+ typename FFT<CCTransform,Dim,T>::ComplexField_t& f)
{
- // indicate we're doing another FFT
- // INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
-
- // Common loop iterate and other vars:
- size_t d;
- int idim; // idim loops over the number of transform dims.
- int begdim, enddim; // beginning and end of transform dim loop
- size_t nTransformDims = this->numTransformDims();
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
-
- // Loop over the dimensions be transformed:
- begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
- enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
- for (idim = begdim; idim != enddim; idim += direction) {
-
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == begdim) {
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into f
- if (idim == enddim-direction) {
- // get domain for comparison
- const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
- (in_dom[0].length() == last_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
- }
- else if (idim == enddim-direction && temp != &f) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using f instead
-
- // transpose and permute to Field with transform dim first
- f[in_dom] = (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *temp = 0;
- temp = &f; // Field* management aid
- }
-
- // Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(idim, direction, localdata);
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
-
- } // loop over all transformed dimensions
-
- // skip final assignment and compress if we used f as final temporary
- if (temp != &f) {
-
- // Now assign back into original Field, and compress last temp's storage:
- f[in_dom] = (*temp)[temp->getLayout().getDomain()];
- if (this->compressTemps()) *temp = 0;
-
- }
-
- // Normalize:
- if (direction == +1)
- f *= Complex_t(this->getNormFact(), 0.0);
+ // indicate we're doing another FFT
+ // INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int idim; // idim loops over the number of transform dims.
+ int begdim, enddim; // beginning and end of transform dim loop
+ size_t nTransformDims = this->numTransformDims();
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
+
+ // Loop over the dimensions be transformed:
+ begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
+ enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
+ for (idim = begdim; idim != enddim; idim += direction) {
+
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == begdim) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into f
+ if (idim == enddim-direction) {
+ // get domain for comparison
+ const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
+ (in_dom[0].length() == last_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<CCTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+ else if (idim == enddim-direction && temp != &f) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using f instead
+
+ // transpose and permute to Field with transform dim first
+ f[in_dom] = (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *temp = 0;
+ temp = &f; // Field* management aid
+ }
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(idim, direction, localdata);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used f as final temporary
+ if (temp != &f) {
+
+ // Now assign back into original Field, and compress last temp's storage:
+ f[in_dom] = (*temp)[temp->getLayout().getDomain()];
+ if (this->compressTemps()) *temp = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1)
+ f *= Complex_t(this->getNormFact(), 0.0);
- return;
+ return;
}
-//#endif
//=============================================================================
// 1D FFT CCTransform Constructors
@@ -511,32 +401,32 @@ FFT<CCTransform,Dim,T>::transform(
template <class T>
FFT<CCTransform,1U,T>::FFT(
- const typename FFT<CCTransform,1U,T>::Domain_t& cdomain,
- const bool transformTheseDims[1U], const bool& compressTemps)
- : FFTBase<1U,T>(FFT<CCTransform,1U,T>::ccFFT, cdomain,
- transformTheseDims, compressTemps)
+ const typename FFT<CCTransform,1U,T>::Domain_t& cdomain,
+ const bool transformTheseDims[1U], const bool& compressTemps)
+: FFTBase<1U,T>(FFT<CCTransform,1U,T>::ccFFT, cdomain,
+ transformTheseDims, compressTemps)
{
- // Tau profiling
+ // Tau profiling
- size_t nTransformDims = 1U;
- // get axis length
- int length;
- length = cdomain[0].length();
+ size_t nTransformDims = 1U;
+ // get axis length
+ int length;
+ length = cdomain[0].length();
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::ccFFT; // all transforms are complex-to-complex
- T& normFact = this->getNormFact();
- normFact = 1.0 / length;
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::ccFFT; // all transforms are complex-to-complex
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / length;
- // set up FFT Engine
- this->getEngine().setup(nTransformDims, &transformType, &length);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(nTransformDims, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -546,30 +436,27 @@ FFT<CCTransform,1U,T>::FFT(
template <class T>
FFT<CCTransform,1U,T>::FFT(
- const typename FFT<CCTransform,1U,T>::Domain_t& cdomain,
- const bool& compressTemps)
- : FFTBase<1U,T>(FFT<CCTransform,1U,T>::ccFFT, cdomain, compressTemps)
+ const typename FFT<CCTransform,1U,T>::Domain_t& cdomain,
+ const bool& compressTemps)
+: FFTBase<1U,T>(FFT<CCTransform,1U,T>::ccFFT, cdomain, compressTemps)
{
- // Tau profiling
-
-
-
+ // Tau profiling
- // get axis length
- int length;
- length = cdomain[0].length();
+ // get axis length
+ int length;
+ length = cdomain[0].length();
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::ccFFT; // all transforms are complex-to-complex
- T& normFact = this->getNormFact();
- normFact = 1.0 / length;
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::ccFFT; // all transforms are complex-to-complex
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / length;
- // set up FFT Engine
- this->getEngine().setup(1U, &transformType, &length);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(1U, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -581,21 +468,18 @@ template <class T>
void
FFT<CCTransform,1U,T>::setup(void)
{
- // Tau profiling
+ // Tau profiling
+ // Get input Field's domain
+ const Domain_t& ndic = this->getDomain();
+ // generate temporary field layout
+ tempLayouts_m = new Layout_t(ndic[0], PARALLEL, 1);
+ // generate temporary Field
+ tempFields_m = new ComplexField_t(*tempLayouts_m);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) tempFields_m->Uncompress();
-
-
- // Get input Field's domain
- const Domain_t& ndic = this->getDomain();
- // generate temporary field layout
- tempLayouts_m = new Layout_t(ndic[0], PARALLEL, 1);
- // generate temporary Field
- tempFields_m = new ComplexField_t(*tempLayouts_m);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) tempFields_m->Uncompress();
-
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -605,14 +489,14 @@ FFT<CCTransform,1U,T>::setup(void)
template <class T>
FFT<CCTransform,1U,T>::~FFT(void) {
- // Tau profiling
+ // Tau profiling
- // delete temporary fields and field layouts
- delete tempFields_m;
- delete tempLayouts_m;
+ // delete temporary fields and field layouts
+ delete tempFields_m;
+ delete tempLayouts_m;
}
@@ -623,107 +507,107 @@ FFT<CCTransform,1U,T>::~FFT(void) {
template <class T>
void
FFT<CCTransform,1U,T>::transform(
- int direction,
- typename FFT<CCTransform,1U,T>::ComplexField_t& f,
- typename FFT<CCTransform,1U,T>::ComplexField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& f,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- // INCIPPLSTAT(incFFTs);
+ // indicate we're doing another FFT
+ // INCIPPLSTAT(incFFTs);
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true);
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true);
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
- // Now do the serial transforms along this dimension:
+ // Now do the serial transforms along this dimension:
- // get temp domain for comparison
- const Domain_t& temp_dom = tempLayouts_m->getDomain();
+ // get temp domain for comparison
+ const Domain_t& temp_dom = tempLayouts_m->getDomain();
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (!constInput) {
- // check that zeroth axis is the same, has one vnode,
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (!constInput) {
+ // check that zeroth axis is the same, has one vnode,
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
+ (in_dom[0].length() == temp_dom[0].length()) &&
+ (in_layout.numVnodes() == 1) &&
+ (f.getGC() == FFT<CCTransform,1U,T>::nullGC) );
+ }
+
+ bool skipFinal;
+ // we might be able
+ // to skip the last temporary and transpose right into g
+
+ // check that zeroth axis is the same, has one vnode
// and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
- (in_dom[0].length() == temp_dom[0].length()) &&
- (in_layout.numVnodes() == 1) &&
- (f.getGC() == FFT<CCTransform,1U,T>::nullGC) );
- }
-
- bool skipFinal;
- // we might be able
- // to skip the last temporary and transpose right into g
-
- // check that zeroth axis is the same, has one vnode
- // and that there are no guard cells
- skipFinal = ( (out_dom[0].sameBase(temp_dom[0])) &&
- (out_dom[0].length() == temp_dom[0].length()) &&
- (out_layout.numVnodes() == 1) &&
- (g.getGC() == FFT<CCTransform,1U,T>::nullGC) );
-
- if (!skipTranspose) {
- // assign to Field with proper layout
- (*tempFields_m) = (*temp);
- temp = tempFields_m; // Field* management aid
- }
- if (skipFinal) {
- // we can skip the last temporary field
- // so do the transpose here using g instead
-
- // assign to Field with proper layout
- g = (*temp);
+ skipFinal = ( (out_dom[0].sameBase(temp_dom[0])) &&
+ (out_dom[0].length() == temp_dom[0].length()) &&
+ (out_layout.numVnodes() == 1) &&
+ (g.getGC() == FFT<CCTransform,1U,T>::nullGC) );
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = &g; // Field* management aid
- }
+ if (!skipTranspose) {
+ // assign to Field with proper layout
+ (*tempFields_m) = (*temp);
+ temp = tempFields_m; // Field* management aid
+ }
+ if (skipFinal) {
+ // we can skip the last temporary field
+ // so do the transpose here using g instead
+ // assign to Field with proper layout
+ g = (*temp);
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = &g; // Field* management aid
+ }
- // should be only one LField!
- typename ComplexField_t::const_iterator_if l_i = temp->begin_if();
- if (l_i != temp->end_if()) {
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
- // Do the 1D FFT:
- this->getEngine().callFFT(0, direction, localdata);
- }
+ // should be only one LField!
+ typename ComplexField_t::const_iterator_if l_i = temp->begin_if();
+ if (l_i != temp->end_if()) {
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+ // Do the 1D FFT:
+ this->getEngine().callFFT(0, direction, localdata);
+ }
- // skip final assignment and compress if we used g as final temporary
- if (temp != &g) {
- // Now assign into output Field, and compress last temp's storage:
- g = (*temp);
- if (this->compressTemps() && temp != &f) *temp = 0;
- }
+ // skip final assignment and compress if we used g as final temporary
+ if (temp != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g = (*temp);
+ if (this->compressTemps() && temp != &f) *temp = 0;
- // Normalize:
- if (direction == +1)
- g *= Complex_t(this->getNormFact(), 0.0);
+ }
+
+ // Normalize:
+ if (direction == +1)
+ g *= Complex_t(this->getNormFact(), 0.0);
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -733,79 +617,79 @@ FFT<CCTransform,1U,T>::transform(
template <class T>
void
FFT<CCTransform,1U,T>::transform(
- int direction,
- typename FFT<CCTransform,1U,T>::ComplexField_t& f)
+ int direction,
+ typename FFT<CCTransform,1U,T>::ComplexField_t& f)
{
- // indicate we're doing another FFT
- // INCIPPLSTAT(incFFTs);
+ // indicate we're doing another FFT
+ // INCIPPLSTAT(incFFTs);
- // Check domain of incoming Field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
+ // Check domain of incoming Field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
- // Now do the serial transforms along this dimension:
+ // Now do the serial transforms along this dimension:
- // get domain for comparison
- const Domain_t& temp_dom = tempLayouts_m->getDomain();
+ // get domain for comparison
+ const Domain_t& temp_dom = tempLayouts_m->getDomain();
- bool skipTranspose;
- // we might be able
- // to skip the transpose into the first temporary Field
+ bool skipTranspose;
+ // we might be able
+ // to skip the transpose into the first temporary Field
- // check that zeroth axis is the same, has one vnode,
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
- (in_dom[0].length() == temp_dom[0].length()) &&
- (in_layout.numVnodes() == 1) &&
- (f.getGC() == FFT<CCTransform,1U,T>::nullGC) );
+ // check that zeroth axis is the same, has one vnode,
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
+ (in_dom[0].length() == temp_dom[0].length()) &&
+ (in_layout.numVnodes() == 1) &&
+ (f.getGC() == FFT<CCTransform,1U,T>::nullGC) );
- if (!skipTranspose) {
- // assign to Field with proper layout
- (*tempFields_m) = (*temp);
- temp = tempFields_m; // Field* management aid
- }
+ if (!skipTranspose) {
+ // assign to Field with proper layout
+ (*tempFields_m) = (*temp);
+ temp = tempFields_m; // Field* management aid
+ }
- // should be only one LField!
- typename ComplexField_t::const_iterator_if l_i = temp->begin_if();
- if (l_i != temp->end_if()) {
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
+ // should be only one LField!
+ typename ComplexField_t::const_iterator_if l_i = temp->begin_if();
+ if (l_i != temp->end_if()) {
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
- // Do the 1D FFT:
- this->getEngine().callFFT(0, direction, localdata);
- }
+ // Do the 1D FFT:
+ this->getEngine().callFFT(0, direction, localdata);
+ }
- // skip final assignment and compress if we used f as final temporary
- if (temp != &f) {
+ // skip final assignment and compress if we used f as final temporary
+ if (temp != &f) {
- // Now assign back into original Field, and compress last temp's storage:
- f = (*temp);
- if (this->compressTemps()) *temp = 0;
+ // Now assign back into original Field, and compress last temp's storage:
+ f = (*temp);
+ if (this->compressTemps()) *temp = 0;
- }
+ }
- // Normalize:
- if (direction == +1)
- f *= Complex_t(this->getNormFact(), 0.0);
+ // Normalize:
+ if (direction == +1)
+ f *= Complex_t(this->getNormFact(), 0.0);
- return;
+ return;
}
@@ -823,75 +707,38 @@ FFT<CCTransform,1U,T>::transform(
template <size_t Dim, class T>
FFT<RCTransform,Dim,T>::FFT(
- const typename FFT<RCTransform,Dim,T>::Domain_t& rdomain,
- const typename FFT<RCTransform,Dim,T>::Domain_t& cdomain,
- const bool transformTheseDims[Dim], const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<RCTransform,Dim,T>::rcFFT, rdomain,
- transformTheseDims, compressTemps),
+ const typename FFT<RCTransform,Dim,T>::Domain_t& rdomain,
+ const typename FFT<RCTransform,Dim,T>::Domain_t& cdomain,
+ const bool transformTheseDims[Dim], const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<RCTransform,Dim,T>::rcFFT, rdomain,
+ transformTheseDims, compressTemps),
complexDomain_m(cdomain), serialAxes_m(1)
{
-/*
-#ifdef IPPL_DKS
-#ifdef IPPL_DKS_OPENCL
- INFOMSG("Init DKS base opencl" << endl);
- //base = DKSBase();
- base.setAPI("OpenCL", 6);
- base.setDevice("-gpu", 4);
- base.initDevice();
-
-#endif
-
-#ifdef IPPL_DKS_CUDA
- INFOMSG("Init DKS base cuda" << endl);
- base.setAPI("Cuda", 4);
- base.setDevice("-gpu", 4);
- base.initDevice();
-
- //create a stream for fft execution other than default
- base.createStream(fftStreamId);
-
-#endif
-
-#ifdef IPPL_DKS_MIC
- INFOMSG("Init DKS base MIC" << endl);
- base.setAPI("OpenMP", 6);
- base.setDevice("-mic", 4);
- base.initDevice();
-
- int dimsize[Dim];
- for (int d=0; d<Dim; ++d)
- dimsize[d] = rdomain[d].length();
-
- base.setupFFTRC(Dim, dimsize);
- base.setupFFTCR(Dim, dimsize,1./(dimsize[0]*dimsize[1]*dimsize[2]));
-#endif
-#endif
-*/
- // construct array of axis lengths
- size_t nTransformDims = this->numTransformDims();
- int* lengths = new int[nTransformDims];
- size_t d;
- for (d=0; d<nTransformDims; ++d)
- lengths[d] = rdomain[this->activeDimension(d)].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int* transformTypes = new int[nTransformDims];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- transformTypes[0] = FFTBase<Dim,T>::rcFFT; // first transform is real-to-complex
- normFact /= lengths[0];
- for (d=1; d<nTransformDims; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all other transforms are complex-to-complex
- normFact /= lengths[d];
- }
-
- // set up FFT Engine
- this->getEngine().setup(nTransformDims, transformTypes, lengths);
- delete [] transformTypes;
- delete [] lengths;
-
- // set up the temporary fields
- setup();
+ // construct array of axis lengths
+ size_t nTransformDims = this->numTransformDims();
+ int* lengths = new int[nTransformDims];
+ size_t d;
+ for (d=0; d<nTransformDims; ++d)
+ lengths[d] = rdomain[this->activeDimension(d)].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int* transformTypes = new int[nTransformDims];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ transformTypes[0] = FFTBase<Dim,T>::rcFFT; // first transform is real-to-complex
+ normFact /= lengths[0];
+ for (d=1; d<nTransformDims; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all other transforms are complex-to-complex
+ normFact /= lengths[d];
+ }
+
+ // set up FFT Engine
+ this->getEngine().setup(nTransformDims, transformTypes, lengths);
+ delete [] transformTypes;
+ delete [] lengths;
+
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -901,74 +748,37 @@ FFT<RCTransform,Dim,T>::FFT(
template <size_t Dim, class T>
FFT<RCTransform,Dim,T>::FFT(
- const typename FFT<RCTransform,Dim,T>::Domain_t& rdomain,
- const typename FFT<RCTransform,Dim,T>::Domain_t& cdomain,
- const bool& compressTemps,
- int serialAxes)
- : FFTBase<Dim,T>(FFT<RCTransform,Dim,T>::rcFFT, rdomain, compressTemps),
+ const typename FFT<RCTransform,Dim,T>::Domain_t& rdomain,
+ const typename FFT<RCTransform,Dim,T>::Domain_t& cdomain,
+ const bool& compressTemps,
+ int serialAxes)
+: FFTBase<Dim,T>(FFT<RCTransform,Dim,T>::rcFFT, rdomain, compressTemps),
complexDomain_m(cdomain), serialAxes_m(serialAxes)
{
- // Tau profiling
-
-/*
-#ifdef IPPL_DKS
-#ifdef IPPL_DKS_OPENCL
- INFOMSG("Init DKS base opencl" << endl);
- base.setAPI("OpenCL", 6);
- base.setDevice("-gpu", 4);
- base.initDevice();
-
-#endif
-
-#ifdef IPPL_DKS_CUDA
- INFOMSG("Init DKS base cuda" << endl);
- base.setAPI("Cuda", 4);
- base.setDevice("-gpu", 4);
- base.initDevice();
- base.setupFFT(0, NULL);
-
- base.createStream(fftStreamId);
-#endif
-
+ // Tau profiling
-#ifdef IPPL_DKS_MIC
- INFOMSG("Init DKS base MIC" << endl);
- base.setAPI("OpenMP", 6);
- base.setDevice("-mic", 4);
- base.initDevice();
-//BENI: Setup MIC for RC FFT and CR FFT (creates the different handles)
+ // construct array of axis lengths
+ int lengths[Dim];
+ size_t d;
+ for (d=0; d<Dim; ++d)
+ lengths[d] = rdomain[d].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int transformTypes[Dim];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ transformTypes[0] = FFTBase<Dim,T>::rcFFT; // first transform is real-to-complex
+ normFact /= lengths[0];
+ for (d=1; d<Dim; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all other transforms are complex-to-complex
+ normFact /= lengths[d];
+ }
- int dimsize[Dim];
- for (int d=0; d<Dim; ++d)
- dimsize[d] = rdomain[d].length();
+ // set up FFT Engine
+ this->getEngine().setup(Dim, transformTypes, lengths);
- base.setupFFTRC(Dim, dimsize);
- base.setupFFTCR(Dim, dimsize,1./(dimsize[0]*dimsize[1]*dimsize[2]));
-#endif
-#endif
-*/
- // construct array of axis lengths
- int lengths[Dim];
- size_t d;
- for (d=0; d<Dim; ++d)
- lengths[d] = rdomain[d].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int transformTypes[Dim];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- transformTypes[0] = FFTBase<Dim,T>::rcFFT; // first transform is real-to-complex
- normFact /= lengths[0];
- for (d=1; d<Dim; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // all other transforms are complex-to-complex
- normFact /= lengths[d];
- }
-
- // set up FFT Engine
- this->getEngine().setup(Dim, transformTypes, lengths);
-
- // set up the temporary fields
- setup();
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -980,133 +790,133 @@ template <size_t Dim, class T>
void
FFT<RCTransform,Dim,T>::setup(void) {
- // Tau profiling
+ // Tau profiling
- PAssert_GT(serialAxes_m, 0);
- PAssert_LT((size_t) serialAxes_m, Dim);
+ PAssert_GT(serialAxes_m, 0);
+ PAssert_LT((size_t) serialAxes_m, Dim);
- size_t d, d2, activeDim;
- size_t nTransformDims = this->numTransformDims();
+ size_t d, d2, activeDim;
+ size_t nTransformDims = this->numTransformDims();
- // Set up the arrays of temporary Fields and FieldLayouts:
+ // Set up the arrays of temporary Fields and FieldLayouts:
- // make first dimension(s) always SERIAL, all other dimensions parallel
- // for the real FFT; make first serialAxes_m axes serial for others
- e_dim_tag serialParallel[Dim];
- e_dim_tag NserialParallel[Dim];
- for (d=0; d < Dim; ++d) {
- serialParallel[d] = (d == 0 ? SERIAL : PARALLEL);
- NserialParallel[d] = (d < (size_t) serialAxes_m ? SERIAL : PARALLEL);
- }
-
- // check that domain lengths agree between real and complex domains
- const Domain_t& domain = this->getDomain();
- activeDim = this->activeDimension(0);
- bool match = true;
- for (d=0; d<Dim; ++d) {
- if (d == activeDim) {
- // real array length n, complex array length n/2+1
- if ( complexDomain_m[d].length() !=
- (domain[d].length()/2 + 1) ) match = false;
- }
- else {
- // real and complex arrays should be same length for all other dims
- if (complexDomain_m[d].length() != domain[d].length()) match = false;
- }
- }
- PInsist(match,
- "Domains provided for real and complex Fields are incompatible!");
-
- // allocate arrays of temp fields and layouts for complex fields
- tempLayouts_m = new Layout_t*[nTransformDims];
- tempFields_m = new ComplexField_t*[nTransformDims];
-
- // set up the single temporary real field, with first dim serial, others par
-
- // make new domains with permuted Indexes, activeDim first
- Domain_t ndip;
- Domain_t ndipc;
- ndip[0] = domain[activeDim];
- ndipc[0] = complexDomain_m[activeDim];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = activeDim + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = domain[nextDim];
- ndipc[d] = complexDomain_m[nextDim];
- }
-
- // generate layout and object for temporary real field
- tempRLayout_m = new Layout_t(ndip, serialParallel, this->transVnodes());
- tempRField_m = new RealField_t(*tempRLayout_m);
-
- // generate layout and object for first temporary complex Field
- tempLayouts_m[0] = new Layout_t(ndipc, serialParallel, this->transVnodes());
- tempFields_m[0] = new ComplexField_t(*tempLayouts_m[0]);
-
- // determine the order in which dimensions will be transposed. Put
- // the transposed dims first, and the others at the end.
- int fftorder[Dim], tmporder[Dim];
- int nofft = nTransformDims;
- for (d=0; d < nTransformDims; ++d)
- fftorder[d] = this->activeDimension(d);
- for (d=0; d < Dim; ++d) {
- // see if the dth dimension is one to transform
- bool active = false;
- for (d2=0; d2 < nTransformDims; ++d2) {
- if (this->activeDimension(d2) == d) {
- active = true;
- break;
- }
+ // make first dimension(s) always SERIAL, all other dimensions parallel
+ // for the real FFT; make first serialAxes_m axes serial for others
+ e_dim_tag serialParallel[Dim];
+ e_dim_tag NserialParallel[Dim];
+ for (d=0; d < Dim; ++d) {
+ serialParallel[d] = (d == 0 ? SERIAL : PARALLEL);
+ NserialParallel[d] = (d < (size_t) serialAxes_m ? SERIAL : PARALLEL);
}
- if (!active)
- // no it is not; put it at the bottom of list
- fftorder[nofft++] = d;
- }
-
- // But since the first FFT is done on a S,[P,P,...] field, permute
- // the order of this to get the first activeDimension at the end.
- nofft = fftorder[0];
- for (d=0; d < (Dim - 1); ++d)
- fftorder[d] = fftorder[d+1];
- fftorder[Dim-1] = nofft;
-
- // now construct the remaining temporary complex fields
-
- // loop through and create actual permuted layouts, and also fields
- size_t dim = 1; // already have one temp field
- while (dim < nTransformDims) {
-
- int sp;
- for (sp=0; sp < serialAxes_m && dim < nTransformDims; ++sp, ++dim) {
-
- // make new domain with permuted Indexes
- for (d=0; d < Dim; ++d)
- ndip[d] = complexDomain_m[fftorder[d]];
-
- // generate layout and object for temporary complex Field
- tempLayouts_m[dim] = new Layout_t(ndip, NserialParallel, this->transVnodes());
- tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
-
- // permute the fft order for the first 'serialAxes_m' axes
- if (serialAxes_m > 1) {
- tmporder[0] = fftorder[0];
- for (d=0; d < (size_t) (serialAxes_m-1); ++d)
- fftorder[d] = fftorder[d+1];
- fftorder[serialAxes_m - 1] = tmporder[0];
- }
+ // check that domain lengths agree between real and complex domains
+ const Domain_t& domain = this->getDomain();
+ activeDim = this->activeDimension(0);
+ bool match = true;
+ for (d=0; d<Dim; ++d) {
+ if (d == activeDim) {
+ // real array length n, complex array length n/2+1
+ if ( complexDomain_m[d].length() !=
+ (domain[d].length()/2 + 1) ) match = false;
+ }
+ else {
+ // real and complex arrays should be same length for all other dims
+ if (complexDomain_m[d].length() != domain[d].length()) match = false;
+ }
}
+ PInsist(match,
+ "Domains provided for real and complex Fields are incompatible!");
+
+ // allocate arrays of temp fields and layouts for complex fields
+ tempLayouts_m = new Layout_t*[nTransformDims];
+ tempFields_m = new ComplexField_t*[nTransformDims];
- // now, permute ALL the axes by serialAxes_m steps, to get the next
- // set of axes in the first n serial slots
- for (d=0; d < Dim; ++d)
- tmporder[d] = fftorder[d];
- for (d=0; d < Dim; ++d)
- fftorder[d] = tmporder[(d + serialAxes_m) % Dim];
- }
+ // set up the single temporary real field, with first dim serial, others par
+
+ // make new domains with permuted Indexes, activeDim first
+ Domain_t ndip;
+ Domain_t ndipc;
+ ndip[0] = domain[activeDim];
+ ndipc[0] = complexDomain_m[activeDim];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = activeDim + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = domain[nextDim];
+ ndipc[d] = complexDomain_m[nextDim];
+ }
+
+ // generate layout and object for temporary real field
+ tempRLayout_m = new Layout_t(ndip, serialParallel, this->transVnodes());
+ tempRField_m = new RealField_t(*tempRLayout_m);
+
+ // generate layout and object for first temporary complex Field
+ tempLayouts_m[0] = new Layout_t(ndipc, serialParallel, this->transVnodes());
+ tempFields_m[0] = new ComplexField_t(*tempLayouts_m[0]);
+
+ // determine the order in which dimensions will be transposed. Put
+ // the transposed dims first, and the others at the end.
+ int fftorder[Dim], tmporder[Dim];
+ int nofft = nTransformDims;
+ for (d=0; d < nTransformDims; ++d)
+ fftorder[d] = this->activeDimension(d);
+ for (d=0; d < Dim; ++d) {
+ // see if the dth dimension is one to transform
+ bool active = false;
+ for (d2=0; d2 < nTransformDims; ++d2) {
+ if (this->activeDimension(d2) == d) {
+ active = true;
+ break;
+ }
+ }
+
+ if (!active)
+ // no it is not; put it at the bottom of list
+ fftorder[nofft++] = d;
+ }
+
+ // But since the first FFT is done on a S,[P,P,...] field, permute
+ // the order of this to get the first activeDimension at the end.
+ nofft = fftorder[0];
+ for (d=0; d < (Dim - 1); ++d)
+ fftorder[d] = fftorder[d+1];
+ fftorder[Dim-1] = nofft;
+
+ // now construct the remaining temporary complex fields
+
+ // loop through and create actual permuted layouts, and also fields
+ size_t dim = 1; // already have one temp field
+ while (dim < nTransformDims) {
+
+ int sp;
+ for (sp=0; sp < serialAxes_m && dim < nTransformDims; ++sp, ++dim) {
+
+ // make new domain with permuted Indexes
+ for (d=0; d < Dim; ++d)
+ ndip[d] = complexDomain_m[fftorder[d]];
+
+ // generate layout and object for temporary complex Field
+ tempLayouts_m[dim] = new Layout_t(ndip, NserialParallel, this->transVnodes());
+ tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
+
+ // permute the fft order for the first 'serialAxes_m' axes
+ if (serialAxes_m > 1) {
+ tmporder[0] = fftorder[0];
+ for (d=0; d < (size_t) (serialAxes_m-1); ++d)
+ fftorder[d] = fftorder[d+1];
+ fftorder[serialAxes_m - 1] = tmporder[0];
+ }
+ }
+
+ // now, permute ALL the axes by serialAxes_m steps, to get the next
+ // set of axes in the first n serial slots
+ for (d=0; d < Dim; ++d)
+ tmporder[d] = fftorder[d];
+ for (d=0; d < Dim; ++d)
+ fftorder[d] = tmporder[(d + serialAxes_m) % Dim];
+ }
}
@@ -1117,19 +927,19 @@ FFT<RCTransform,Dim,T>::setup(void) {
template <size_t Dim, class T>
FFT<RCTransform,Dim,T>::~FFT(void) {
- // Tau profiling
+ // Tau profiling
- // delete temporary fields and layouts
- size_t nTransformDims = this->numTransformDims();
- for (size_t d=0; d<nTransformDims; ++d) {
- delete tempFields_m[d];
- delete tempLayouts_m[d];
- }
- delete [] tempFields_m;
- delete [] tempLayouts_m;
- delete tempRField_m;
- delete tempRLayout_m;
+ // delete temporary fields and layouts
+ size_t nTransformDims = this->numTransformDims();
+ for (size_t d=0; d<nTransformDims; ++d) {
+ delete tempFields_m[d];
+ delete tempLayouts_m[d];
+ }
+ delete [] tempFields_m;
+ delete [] tempLayouts_m;
+ delete tempRField_m;
+ delete tempRLayout_m;
}
@@ -1145,677 +955,386 @@ FFT<RCTransform,Dim,T>::~FFT(void) {
template <size_t Dim, class T>
void
FFT<RCTransform,Dim,T>::transformDKSRC(
- int direction,
- typename FFT<RCTransform,Dim,T>::RealField_t& f,
- void* real_ptr,
- void* comp_ptr,
- DKSOPAL &dksbase,
- int streamId,
- const bool& constInput)
+ int direction,
+ typename FFT<RCTransform,Dim,T>::RealField_t& f,
+ void* real_ptr,
+ void* comp_ptr,
+ DKSOPAL &dksbase,
+ int streamId,
+ const bool& constInput)
{
- //check the domain of incoming field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
-
- PAssert_EQ( this->checkDomain(this->getDomain(), in_dom), true);
-
- size_t nTransformDims = this->numTransformDims();
-
- //*** using tempRField_m and transposing f field ***//
- /*
- RealField_t* tempR = tempRField_m;
+ //check the domain of incoming field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
- if (!constInput) {
- // see if we can use input field f as a temporary
- bool skipTemp = true;
+ PAssert_EQ( this->checkDomain(this->getDomain(), in_dom), true);
- // more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempRLayout_m) ) {
- skipTemp = false;
- } else {
- // make sure distributions match
- for (unsigned d=0; d<Dim; ++d)
- if (in_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- // make sure vnode counts match
- if (in_layout.numVnodes() != tempRLayout_m->numVnodes())
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
+ size_t nTransformDims = this->numTransformDims();
- // if we can skip using this temporary, set the tempr pointer to the
- // original incoming field. otherwise, it will stay pointing at the
- // temporary real field, and we'll need to do a transpose of the data
- // from the original into the temporary.
- if (skipTemp)
- tempR = &f;
- }
-
- // if we're not using input as a temporary ...
- if (tempR != &f) {
- // transpose and permute to real field with transform dim first
- (*tempR)[tempR->getDomain()] = f[in_dom];
- }
- */
- //*** just use f field as is and keep decomposition as defined in input file ***//
- RealField_t* tempR = &f;
-
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- rl_i = tempR->begin_if();
-
- // get the lfields
- RealLField_t* rldf = (*rl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
- // get the raw data pointers
- T* localreal = rldf->getP();
-
- /** get global dimensions of real domain and local dimensions of real subdomain
- calc global dimensions of complex subdomain */
- int NR_l[Dim], NR_g[Dim], NC_g[Dim];
- for (size_t d = 0; d < Dim; d++) {
- NR_l[d] = (int)rldf->size(d);
- NR_g[d] = (int)tempR->getDomain()[d].length();
- NC_g[d] = NR_g[d];
- }
- NC_g[0] = (NC_g[0] / 2) + 1;
-
- //get global and local domain sizes
- int sizereal = NR_l[0]*NR_l[1]*NR_l[2];
- int totalreal = tempR->getDomain().size();
- //int totalcomp = NC_g[0]*NC_g[1]*NC_g[2];
-
- //local vnodes get starting position for real field subdomains
- int *idx = new int[Ippl::getNodes()];
- int *idy = new int[Ippl::getNodes()];
- int *idz = new int[Ippl::getNodes()];
- for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position for real field subdomains
- for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
-
- if (Ippl::myNode() == 0) {
-
- //if only one node is working do dksbase write otherwise use cuda aware mpi
- if (Ippl::getNodes() > 1) {
-
- if (streamId == -1) {
- //gather data from different mpi processes directly into gpu buffer
- dksbase.gather3DData( real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l,
- idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
- } else {
- //gather data using CUDA IPC for async data transfer
- dksbase.gather3DDataAsync<T>( real_ptr, localreal, NR_g, NR_l, id, streamId);
- //sync needed to wait for data transfer to finish
- dksbase.syncDevice();
- MPI_Barrier(Ippl::getComm());
- }
+ //*** just use f field as is and keep decomposition as defined in input file ***//
+ RealField_t* tempR = &f;
- } else {
- //write real data to device
- dksbase.writeDataAsync<T>(real_ptr, localreal, totalreal, streamId);
- //dksbase.writeData<T>(real_ptr, localreal, totalreal);
- }
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ rl_i = tempR->begin_if();
- //call real to complex fft
- dksbase.callR2CFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g, streamId);
+ // get the lfields
+ RealLField_t* rldf = (*rl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ // get the raw data pointers
+ T* localreal = rldf->getP();
- //normalize fft
- if (direction == +1)
- dksbase.callNormalizeFFT(comp_ptr, nTransformDims, (int*) NC_g, streamId);
+ /** get global dimensions of real domain and local dimensions of real subdomain
+ calc global dimensions of complex subdomain */
+ int NR_l[Dim], NR_g[Dim], NC_g[Dim];
+ for (size_t d = 0; d < Dim; d++) {
+ NR_l[d] = (int)rldf->size(d);
+ NR_g[d] = (int)tempR->getDomain()[d].length();
+ NC_g[d] = NR_g[d];
+ }
+ NC_g[0] = (NC_g[0] / 2) + 1;
+
+ //get global and local domain sizes
+ int sizereal = NR_l[0]*NR_l[1]*NR_l[2];
+ int totalreal = tempR->getDomain().size();
+ //int totalcomp = NC_g[0]*NC_g[1]*NC_g[2];
+
+ //local vnodes get starting position for real field subdomains
+ int *idx = new int[Ippl::getNodes()];
+ int *idy = new int[Ippl::getNodes()];
+ int *idz = new int[Ippl::getNodes()];
+ for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
+ Domain_t tmp = (*i_s).second->getDomain();
+ int node = (*i_s).second->getNode();
+ idx[node] = tmp[0].min();
+ idy[node] = tmp[1].min();
+ idz[node] = tmp[2].min();
+ }
+
+ //remote vnodes get starting position for real field subdomains
+ for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
+ Domain_t tmp = (*remote).second->getDomain();
+ int node = (*remote).second->getNode();
+ idx[node] = tmp[0].min();
+ idy[node] = tmp[1].min();
+ idz[node] = tmp[2].min();
+ }
+
+ int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
+
+ if (Ippl::myNode() == 0) {
+
+ //if only one node is working do dksbase write otherwise use cuda aware mpi
+ if (Ippl::getNodes() > 1) {
+
+ if (streamId == -1) {
+ //gather data from different mpi processes directly into gpu buffer
+ dksbase.gather3DData( real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l,
+ idx, idy, idz,
+ Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ } else {
+ //gather data using CUDA IPC for async data transfer
+ dksbase.gather3DDataAsync<T>( real_ptr, localreal, NR_g, NR_l, id, streamId);
+ //sync needed to wait for data transfer to finish
+ dksbase.syncDevice();
+ MPI_Barrier(Ippl::getComm());
+ }
+
+ } else {
+ //write real data to device
+ dksbase.writeDataAsync<T>(real_ptr, localreal, totalreal, streamId);
+ //dksbase.writeData<T>(real_ptr, localreal, totalreal);
+ }
+
+ //call real to complex fft
+ dksbase.callR2CFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g, streamId);
+
+ //normalize fft
+ if (direction == +1)
+ dksbase.callNormalizeFFT(comp_ptr, nTransformDims, (int*) NC_g, streamId);
- } else {
- if (streamId == -1) {
- //send data via gatherv to gpu controled by root process
- dksbase.gather3DData( NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
} else {
- //transfer data to device memory
- dksbase.gather3DDataAsync<T>( real_ptr, localreal, NR_g, NR_l, id, streamId);
- //sync needed to wait for data transfer to finish
- dksbase.syncDevice();
- MPI_Barrier(Ippl::getComm());
- }
-
- }
- /* end dks part */
-
- // finish timing the whole mess
+ if (streamId == -1) {
+ //send data via gatherv to gpu controled by root process
+ dksbase.gather3DData( NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
+ Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ } else {
+ //transfer data to device memory
+ dksbase.gather3DDataAsync<T>( real_ptr, localreal, NR_g, NR_l, id, streamId);
+ //sync needed to wait for data transfer to finish
+ dksbase.syncDevice();
+ MPI_Barrier(Ippl::getComm());
+ }
+
+ }
+ /* end dks part */
+
+ // finish timing the whole mess
}
#endif
-/*
-#ifdef IPPL_DKS
-template <unsigned Dim, class T>
+template <size_t Dim, class T>
void
FFT<RCTransform,Dim,T>::transform(
- int direction,
- typename FFT<RCTransform,Dim,T>::RealField_t& f,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<RCTransform,Dim,T>::RealField_t& f,
+ typename FFT<RCTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- // time the whole mess
- static IpplTimings::TimerRef tottimer=IpplTimings::getTimer("RC-total-gpu");
- IpplTimings::startTimer(tottimer);
+ FFTDBG(Inform tmsg("FFT-RC-forward"));
- // check domain of incoming fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
+ // time the whole mess
- PAssert( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(complexDomain_m,out_dom) );
+ // indicate we're doing another fft
+ // incipplstat(incffts);
- RealField_t* tempR = tempRField_m; // field* management aid
+ // check domain of incoming fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
- if (!constInput) {
- // see if we can use input field f as a temporary
- bool skipTemp = true;
- // more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempRLayout_m) ) {
- skipTemp = false;
- } else {
- // make sure distributions match
- for (unsigned d=0; d<Dim; ++d)
- if (in_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- // make sure vnode counts match
- if (in_layout.numVnodes() != tempRLayout_m->numVnodes())
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(complexDomain_m,out_dom), true);
- // if we can skip using this temporary, set the tempr pointer to the
- // original incoming field. otherwise, it will stay pointing at the
- // temporary real field, and we'll need to do a transpose of the data
- // from the original into the temporary.
- if (skipTemp)
- tempR = &f;
- }
+ // common loop iterate and other vars:
+ size_t d;
+ size_t idim; // idim loops over the number of transform dims.
+ size_t nTransformDims = this->numTransformDims();
- // if we're not using input as a temporary ...
- if (tempR != &f) {
- // transpose and permute to real field with transform dim first
- (*tempR)[tempR->getDomain()] = f[in_dom];
- }
+ // handle first rc transform separately
+ idim = 0;
- // field* for temp field management:
- ComplexField_t* temp = tempFields_m[0];
+ RealField_t* tempR = tempRField_m; // field* management aid
+ if (!constInput) {
+ // see if we can use input field f as a temporary
+ bool skipTemp = true;
- // see if we can put final result directly into g. this is useful if
- // we're doing just a 1d fft of one dimension of a multi-dimensional field.
+ // more rigorous match required here; check that layouts are identical
+ if ( !(in_layout == *tempRLayout_m) ) {
+ skipTemp = false;
+ } else {
+ // make sure distributions match
+ for (d=0; d<Dim; ++d)
+ if (in_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
+ skipTemp = false;
- unsigned nTransformDims = this->numTransformDims();
+ // make sure vnode counts match
+ if (in_layout.numVnodes() != tempRLayout_m->numVnodes())
+ skipTemp = false;
- if (nTransformDims == 1) { // only a single rc transform
- bool skipTemp = true;
+ // also make sure there are no guard cells
+ if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
+ skipTemp = false;
+ }
- // more rigorous match required here; check that layouts are identical
- if (!(out_layout == *tempLayouts_m[0])) {
- skipTemp = false;
- } else {
- for (unsigned d=0; d<Dim; ++d)
- if (out_layout.getDistribution(d) != tempLayouts_m[0]->getDistribution(d))
- skipTemp = false;
+ // if we can skip using this temporary, set the tempr pointer to the
+ // original incoming field. otherwise, it will stay pointing at the
+ // temporary real field, and we'll need to do a transpose of the data
+ // from the original into the temporary.
+ if (skipTemp)
+ tempR = &f;
+ }
- if ( out_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
- skipTemp = false;
+ // if we're not using input as a temporary ...
+ if (tempR != &f) {
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- // if we can skip using the temporary, set the pointer to the output
- // field for the first fft to the second provided field (g)
- if (skipTemp)
- temp = &g;
- }
- }
-
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- rl_i = tempR->begin_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->end_if();
- cl_i = temp->begin_if();
-
- // get the lfields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
-
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
-
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
-
- //get global dimensions of domain and local dimensions of subdomain
- int NR_l[Dim], NC_l[Dim], NR_g[Dim], NC_g[Dim];
- for (unsigned d = 0; d < Dim; d++) {
- NR_l[d] = (int)rldf->size(d);
- NC_l[d] = (int)cldf->size(d);
- NR_g[d] = (int)tempR->getDomain()[d].length();
- NC_g[d] = (int)temp->getDomain()[d].length();
- }
-
- //get global and local domain sizes
- int sizereal = NR_l[0]*NR_l[1]*NR_l[2];
- int sizecomp = NC_l[0]*NC_l[1]*NC_l[2];
-
- int totalreal = tempR->getDomain().size();
- int totalcomp = temp->getDomain().size();
-
-
- //local vnodes get starting position for real field subdomains
- int *idx = new int[Ippl::getNodes()];
- int *idy = new int[Ippl::getNodes()];
- int *idz = new int[Ippl::getNodes()];
- for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position for real field subdomains
- for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //local vnodes get starting position
- int *cidx = new int[Ippl::getNodes()];
- int *cidy = new int[Ippl::getNodes()];
- int *cidz = new int[Ippl::getNodes()];
- for (typename Layout_t::const_iterator_iv i_s = temp->getLayout().begin_iv(); i_s != temp->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- cidx[node] = tmp[0].min();
- cidy[node] = tmp[1].min();
- cidz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position
- for (typename Layout_t::iterator_dv remote = temp->getLayout().begin_rdv(); remote != temp->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- cidx[node] = tmp[0].min();
- cidy[node] = tmp[1].min();
- cidz[node] = tmp[2].min();
- }
-
- //int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
- //int cid[3] = {cidx[Ippl::myNode()], cidy[Ippl::myNode()], cidz[Ippl::myNode()]};
-
- int ierr;
- void *real_ptr, *comp_ptr;
-
- if (Ippl::myNode() == 0) {
-
- //allocate memory on device for real and complex arrays
- real_ptr = base.allocateMemory<T>(totalreal, ierr);
- comp_ptr = base.allocateMemory<Complex_t>(totalcomp, ierr);
-
- //if only one node is working do dksbase write otherwise use cuda aware mpi
- if (Ippl::getNodes() > 1) {
- //gather data from different mpi processes directly into gpu buffer
- base.gather3DData( real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ // transpose AND PERMUTE TO REAL FIELD WITH TRANSFORM DIM FIRST
+ FFTDBG(tmsg << "doing transpose of real field into temporary ");
+ FFTDBG(tmsg << "with layout = " << tempR->getLayout() << std::endl);
+ (*tempR)[tempR->getDomain()] = f[in_dom];
- } else {
- //write real data to device
- ierr = base.writeData<T>(real_ptr, localreal, totalreal);
}
- //call real to complex fft
- base.callR2CFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g);
+ // field* for temp field management:
+ ComplexField_t* temp = tempFields_m[0];
- //if only one node is working do dksbase read otherwise use cuda aware mpi
- if (Ippl::getNodes() > 1) {
- //scatter data to different mpi processes directly from gpu buffer
- MPI_Barrier( Ippl::getComm() );
+ // see if we can put final result directly into g. this is useful if
+ // we're doing just a 1d fft of one dimension of a multi-dimensional field.
+ if (nTransformDims == 1) { // only a single rc transform
+ bool skipTemp = true;
- base.scatter3DData(comp_ptr, localcomp, sizecomp, MPI_DOUBLE_COMPLEX, NC_g, NC_l, cidx, cidy,
- cidz, Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ // more rigorous match required here; check that layouts are identical
+ if (!(out_layout == *tempLayouts_m[0])) {
+ skipTemp = false;
+ } else {
+ for (d=0; d<Dim; ++d)
+ if (out_layout.getDistribution(d) !=
+ tempLayouts_m[0]->getDistribution(d))
+ skipTemp = false;
- } else {
- //read complex data from device
- base.readData<Complex_t>(comp_ptr, localcomp, totalcomp);
+ if ( out_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
+ skipTemp = false;
+
+ // also make sure there are no guard cells
+ if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
+ skipTemp = false;
+
+ // if we can skip using the temporary, set the pointer to the output
+ // field for the first fft to the second provided field (g)
+ if (skipTemp)
+ temp = &g;
+ }
}
- //free device memory
- base.freeMemory<T>(real_ptr, totalreal);
- base.freeMemory<Complex_t>(comp_ptr, totalcomp);
+ FFTDBG(tmsg << "doing real->complex fft of first dimension ..." << std::endl);
- } else {
+ // loop over all the vnodes, working on the lfield in each.
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
+ // get the lfields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
- //send data via gatherv to gpu controled by root process
- base.gather3DData( NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
- //receive data from gpu controled by root process
- MPI_Barrier(Ippl::getComm());
- base.scatter3DData(NULL, localcomp, sizecomp, MPI_DOUBLE_COMPLEX, NC_g, NC_l, cidx, cidy, cidz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
- }
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
- //assign complex field
- if (temp != &g) {
+ // number of strips should be the same for real and complex lfields!
+ int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
+ for (d=1; d<Dim; ++d)
+ nstrips *= rldf->size(d);
- (*tempFields_m[1])[tempLayouts_m[1]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
- temp = tempFields_m[1]; // field* management aid
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // move the data into the complex strip, which is two reals longer
+ for (int ilen=0; ilen<lengthreal; ilen+=2) {
+ localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
+ }
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- if (this->compressTemps()) *temp = 0;
- }
+ // do the 1d real-to-complex fft:
+ // note that real-to-complex fft direction is always +1
+ this->getEngine().callFFT(idim, +1, localcomp);
- if (tempR != &f) {
- if (this->compressTemps()) *tempR = 0;
- }
+ // advance the data pointers
+ localreal += lengthreal;
+ localcomp += lengthcomp;
+ } // loop over 1d strips
- // normalize:
- //if (direction == +1)
- // g = g * this->getNormFact();
+ } // loop over all the lfields
- // finish timing the whole mess
- IpplTimings::stopTimer(tottimer);
+ // compress temporary storage
+ if (this->compressTemps() && tempR != &f)
+ *tempR = 0;
-}
-#else
-*/
-template <size_t Dim, class T>
-void
-FFT<RCTransform,Dim,T>::transform(
- int direction,
- typename FFT<RCTransform,Dim,T>::RealField_t& f,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
-{
- FFTDBG(Inform tmsg("FFT-RC-forward"));
-
- // time the whole mess
-
- // indicate we're doing another fft
- // incipplstat(incffts);
+ // now proceed with the other complex-to-complex transforms
- // check domain of incoming fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
+ // local work array passed to fft:
+ Complex_t* localdata;
+ // loop over the remaining dimensions to be transformed:
+ for (idim = 1; idim < nTransformDims; ++idim) {
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(complexDomain_m,out_dom), true);
+ bool skipTranspose = false;
- // common loop iterate and other vars:
- size_t d;
- size_t idim; // idim loops over the number of transform dims.
- size_t nTransformDims = this->numTransformDims();
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into g
+ if (idim == nTransformDims-1) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
- // handle first rc transform separately
- idim = 0;
+ // make sure there are no guard cells, and that the first
+ // axis matches what we expect and is serial. only need to
+ // check first axis since we're just fft'ing that one dimension.
+ skipTranspose = (g.getGC() == FFT<RCTransform,Dim,T>::nullGC &&
+ out_dom[0].sameBase(last_dom[0]) &&
+ out_dom[0].length() == last_dom[0].length() &&
+ out_layout.getDistribution(0) == SERIAL);
+ }
- RealField_t* tempR = tempRField_m; // field* management aid
- if (!constInput) {
- // see if we can use input field f as a temporary
- bool skipTemp = true;
-
- // more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempRLayout_m) ) {
- skipTemp = false;
- } else {
- // make sure distributions match
- for (d=0; d<Dim; ++d)
- if (in_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- // make sure vnode counts match
- if (in_layout.numVnodes() != tempRLayout_m->numVnodes())
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
-
- // if we can skip using this temporary, set the tempr pointer to the
- // original incoming field. otherwise, it will stay pointing at the
- // temporary real field, and we'll need to do a transpose of the data
- // from the original into the temporary.
- if (skipTemp)
- tempR = &f;
- }
+ if (!skipTranspose) {
+ // transpose and permute to field with transform dim first
+ FFTDBG(tmsg << "doing complex->complex transpose into field ");
+ FFTDBG(tmsg << "with layout = " << tempFields_m[idim]->getLayout());
+ FFTDBG(tmsg << std::endl);
- // if we're not using input as a temporary ...
- if (tempR != &f) {
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+ // compress out previous iterate's storage:
+ if (this->compressTemps())
+ *temp = 0;
+ temp = tempFields_m[idim]; // field* management aid
- // transpose AND PERMUTE TO REAL FIELD WITH TRANSFORM DIM FIRST
- FFTDBG(tmsg << "doing transpose of real field into temporary ");
- FFTDBG(tmsg << "with layout = " << tempR->getLayout() << std::endl);
- (*tempR)[tempR->getDomain()] = f[in_dom];
+ } else if (idim == nTransformDims-1) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using g instead
- }
+ // transpose and permute to field with transform dim first
+ FFTDBG(tmsg << "doing final complex->complex transpose ");
+ FFTDBG(tmsg << "into return ");
+ FFTDBG(tmsg << "with layout = " << g.getLayout());
+ FFTDBG(tmsg << std::endl);
- // field* for temp field management:
- ComplexField_t* temp = tempFields_m[0];
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- // see if we can put final result directly into g. this is useful if
- // we're doing just a 1d fft of one dimension of a multi-dimensional field.
- if (nTransformDims == 1) { // only a single rc transform
- bool skipTemp = true;
+ // compress out previous iterate's storage:
+ if (this->compressTemps())
+ *temp = 0;
+ temp = &g; // field* management aid
- // more rigorous match required here; check that layouts are identical
- if (!(out_layout == *tempLayouts_m[0])) {
- skipTemp = false;
- } else {
- for (d=0; d<Dim; ++d)
- if (out_layout.getDistribution(d) !=
- tempLayouts_m[0]->getDistribution(d))
- skipTemp = false;
-
- if ( out_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
-
- // if we can skip using the temporary, set the pointer to the output
- // field for the first fft to the second provided field (g)
- if (skipTemp)
- temp = &g;
- }
- }
+ }
- FFTDBG(tmsg << "doing real->complex fft of first dimension ..." << std::endl);
- // loop over all the vnodes, working on the lfield in each.
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
- // get the lfields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
+ FFTDBG(tmsg << "doing complex->complex fft of other dimension .." << std::endl);
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
+ // loop over all the vnodes, working on the lfield in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+ // get the lfield
+ ComplexLField_t* ldf = (*l_i).second.get();
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
+ // make sure we are uncompressed
+ ldf->Uncompress();
- // number of strips should be the same for real and complex lfields!
- int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
- for (d=1; d<Dim; ++d)
- nstrips *= rldf->size(d);
+ // get the raw data pointer
+ localdata = ldf->getP();
+ // do 1d complex-to-complex fft's on all the strips in the lfield:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d)
+ nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // move the data into the complex strip, which is two reals longer
- for (int ilen=0; ilen<lengthreal; ilen+=2) {
- localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
- }
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // do the 1D FFT:
+ //this->getEngine().callFFT(idim, direction, localdata);
+ this->getEngine().callFFT(idim, +1, localdata);
- // do the 1d real-to-complex fft:
- // note that real-to-complex fft direction is always +1
- this->getEngine().callFFT(idim, +1, localcomp);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
- // advance the data pointers
- localreal += lengthreal;
- localcomp += lengthcomp;
- } // loop over 1d strips
+ } // loop over all transformed dimensions
- } // loop over all the lfields
- // compress temporary storage
- if (this->compressTemps() && tempR != &f)
- *tempR = 0;
+ // skip final assignment and compress if we used g as final temporary
+ if (temp != &g) {
- // now proceed with the other complex-to-complex transforms
- // local work array passed to fft:
- Complex_t* localdata;
+ // Now assign into output Field, and compress last temp's storage:
+ FFTDBG(tmsg << "Doing cleanup complex->complex transpose ");
+ FFTDBG(tmsg << "into return ");
+ FFTDBG(tmsg << "with layout = " << g.getLayout());
+ FFTDBG(tmsg << std::endl);
- // loop over the remaining dimensions to be transformed:
- for (idim = 1; idim < nTransformDims; ++idim) {
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- bool skipTranspose = false;
+ if (this->compressTemps()) *temp = 0;
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into g
- if (idim == nTransformDims-1) {
- // get the domain for comparison
- const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
-
- // make sure there are no guard cells, and that the first
- // axis matches what we expect and is serial. only need to
- // check first axis since we're just fft'ing that one dimension.
- skipTranspose = (g.getGC() == FFT<RCTransform,Dim,T>::nullGC &&
- out_dom[0].sameBase(last_dom[0]) &&
- out_dom[0].length() == last_dom[0].length() &&
- out_layout.getDistribution(0) == SERIAL);
}
- if (!skipTranspose) {
- // transpose and permute to field with transform dim first
- FFTDBG(tmsg << "doing complex->complex transpose into field ");
- FFTDBG(tmsg << "with layout = " << tempFields_m[idim]->getLayout());
- FFTDBG(tmsg << std::endl);
-
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // compress out previous iterate's storage:
- if (this->compressTemps())
- *temp = 0;
- temp = tempFields_m[idim]; // field* management aid
-
- } else if (idim == nTransformDims-1) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using g instead
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
- // transpose and permute to field with transform dim first
- FFTDBG(tmsg << "doing final complex->complex transpose ");
- FFTDBG(tmsg << "into return ");
- FFTDBG(tmsg << "with layout = " << g.getLayout());
- FFTDBG(tmsg << std::endl);
-
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
-
- // compress out previous iterate's storage:
- if (this->compressTemps())
- *temp = 0;
- temp = &g; // field* management aid
-
- }
-
-
- FFTDBG(tmsg << "doing complex->complex fft of other dimension .." << std::endl);
-
- // loop over all the vnodes, working on the lfield in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
- // get the lfield
- ComplexLField_t* ldf = (*l_i).second.get();
-
- // make sure we are uncompressed
- ldf->Uncompress();
-
- // get the raw data pointer
- localdata = ldf->getP();
-
- // do 1d complex-to-complex fft's on all the strips in the lfield:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d)
- nstrips *= ldf->size(d);
-
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // do the 1D FFT:
- //this->getEngine().callFFT(idim, direction, localdata);
- this->getEngine().callFFT(idim, +1, localdata);
-
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
- } // loop over all transformed dimensions
-
-
- // skip final assignment and compress if we used g as final temporary
- if (temp != &g) {
-
-
- // Now assign into output Field, and compress last temp's storage:
- FFTDBG(tmsg << "Doing cleanup complex->complex transpose ");
- FFTDBG(tmsg << "into return ");
- FFTDBG(tmsg << "with layout = " << g.getLayout());
- FFTDBG(tmsg << std::endl);
-
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
-
- if (this->compressTemps()) *temp = 0;
-
- }
-
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
-
- // finish timing the whole mess
+ // finish timing the whole mess
}
//#endif
@@ -1833,620 +1352,361 @@ FFT<RCTransform,Dim,T>::transform(
template <size_t Dim, class T>
void
FFT<RCTransform,Dim,T>::transformDKSCR(
- int direction,
- RealField_t& g,
- void* real_ptr,
- void* comp_ptr,
- DKSOPAL &dksbase,
- int streamId,
- const bool& constInput)
+ int direction,
+ RealField_t& g,
+ void* real_ptr,
+ void* comp_ptr,
+ DKSOPAL &dksbase,
+ int streamId,
+ const bool& constInput)
{
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
- //if (Ippl::myNode() == 0)
- // std::cout << "DEBUG INVERSE g: " << g.getLayout() << std::endl;
+ //if (Ippl::myNode() == 0)
+ // std::cout << "DEBUG INVERSE g: " << g.getLayout() << std::endl;
- PAssert_EQ( this->checkDomain(this->getDomain(),out_dom), true);
+ PAssert_EQ( this->checkDomain(this->getDomain(),out_dom), true);
- size_t nTransformDims = this->numTransformDims();
+ size_t nTransformDims = this->numTransformDims();
- // see if we can put final result directly into g
- RealField_t* tempR;
-
- /*** using tempRField_m and transposing to g field in the end ***/
- /*
- bool skipTemp = true;
-
- // more rigorous match required here; check that layouts are identical
- if (!(out_layout == *tempRLayout_m)) {
- skipTemp = false;
- } else {
- for (unsigned d=0; d<Dim; ++d)
- if (out_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
- skipTemp = false;
+ // see if we can put final result directly into g
+ RealField_t* tempR;
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
-
- if (skipTemp)
+ //***Use g as is and keep decomposition as defined in input file***/
tempR = &g;
- else
- tempR = tempRField_m;
- */
- //***Use g as is and keep decomposition as defined in input file***/
- tempR = &g;
- //if (Ippl::myNode() == 0)
- // std::cout << "DEBUG INVERSE tempR: " << tempR->getLayout() << std::endl;
-
-
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- rl_i = tempR->begin_if();
-
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
-
- // get the raw data pointers
- T* localreal = rldf->getP();
-
- //get sizes of global domains and local subdomains
- int NR_l[Dim], NR_g[Dim], NC_g[Dim];
- for (size_t d=0; d<Dim; d++) {
- NR_l[d] = (int)rldf->size(d);
- NR_g[d] = (int)tempR->getDomain()[d].length();
- NC_g[d] = NR_g[d];
- }
- NC_g[0] = (NC_g[0] / 2) + 1;
-
- //get sizes of global and local domains
- int totalreal = tempR->getDomain().size();
-
- //local vnodes get starting position for real field subdomains
- int *idx = new int[Ippl::getNodes()];
- int *idy = new int[Ippl::getNodes()];
- int *idz = new int[Ippl::getNodes()];
- for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position for real field subdomains
- for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
-
- /* DKS part */
- if (Ippl::myNode() == 0) {
-
- //call real to complex fft
- dksbase.callC2RFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g, streamId);
-
- //normalize
- if (direction == +1)
- dksbase.callNormalizeC2RFFT(real_ptr, nTransformDims, (int*)NR_g, streamId);
-
- if (Ippl::getNodes() > 1) {
- dksbase.syncDevice();
- MPI_Barrier(Ippl::getComm());
- /*
- dksbase.scatter3DData(real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l,
- idx, idy, idz, Ippl::getNodes(), Ippl::myNode(),
- 0, Ippl::getComm() );
- */
- dksbase.scatter3DDataAsync<T>(real_ptr, localreal, NR_g, NR_l, id);
- MPI_Barrier(Ippl::getComm());
- dksbase.syncDevice();
- MPI_Barrier(Ippl::getComm());
- } else {
- //read real data from device
- dksbase.readDataAsync<T>(real_ptr, localreal, totalreal, streamId);
- dksbase.syncDevice();
- //dksbase.readData<T>(real_ptr, localreal, totalreal);
- }
- } else {
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ rl_i = tempR->begin_if();
- //receive data from GPU controled by root process
- MPI_Barrier(Ippl::getComm());
- /*
- dksbase.scatter3DData(NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
- */
- dksbase.scatter3DDataAsync<T>(real_ptr, localreal, NR_g, NR_l, id);
- MPI_Barrier(Ippl::getComm());
- dksbase.syncDevice();
- MPI_Barrier(Ippl::getComm());
- }
- /* end dks part */
-
- // Now assign into output Field, and compress last temp's storage:
- if (tempR != &g) {
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
- if (this->compressTemps()) *tempR = 0;
- }
-
- // finish timing the whole mess
-}
-#endif
-
-
-/*
-#ifdef IPPL_DKS
-template <unsigned Dim, class T>
-void
-FFT<RCTransform,Dim,T>::transform(
- int direction,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& f,
- typename FFT<RCTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
-{
-
- // time the whole mess
- static IpplTimings::TimerRef tottimer=IpplTimings::getTimer("RC-total-gpu");
- IpplTimings::startTimer(tottimer);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
-
- PAssert( this->checkDomain(complexDomain_m,in_dom) &&
- this->checkDomain(this->getDomain(),out_dom) );
-
-
-
- unsigned nTransformDims = this->numTransformDims();
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
-
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[0]->getDomain();
-
- bool skipTranspose = true;
-
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<RCTransform,Dim,T>::nullGC) );
-
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[0])[tempLayouts_m[0]->getDomain()] = (*temp)[temp->getLayout().getDomain()];
-
- //Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f)
- *temp = 0;
- temp = tempFields_m[0]; // Field* management aid
- }
-
-
- // see if we can put final result directly into g
- RealField_t* tempR;
- bool skipTemp = true;
-
- // more rigorous match required here; check that layouts are identical
- if (!(out_layout == *tempRLayout_m)) {
- skipTemp = false;
- } else {
- for (unsigned d=0; d<Dim; ++d)
- if (out_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
- skipTemp = false;
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
+ // get the raw data pointers
+ T* localreal = rldf->getP();
- if (skipTemp)
- tempR = &g;
- else
- tempR = tempRField_m;
-
-
- //unsigned nTransformDims = this->numTransformDims();
- //RealField_t* tempR = &g;
- //ComplexField_t* temp = &f;
-
-
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- rl_i = tempR->begin_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->end_if();
- cl_i = temp->begin_if();
-
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
-
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
-
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
-
- //get sizes of global domains and local subdomains
- int NR_l[Dim], NC_l[Dim], NR_g[Dim], NC_g[Dim];
- for (unsigned d=0; d<Dim; d++) {
- NR_l[d] = (int)rldf->size(d);
- NC_l[d] = (int)cldf->size(d);
- NR_g[d] = (int)tempR->getDomain()[d].length();
- NC_g[d] = (int)temp->getDomain()[d].length();
- }
-
- //get sizes of global and local domains
- int sizereal = NR_l[0]*NR_l[1]*NR_l[2];
- int sizecomp = NC_l[0]*NC_l[1]*NC_l[2];
- int totalreal = tempR->getDomain().size();
- int totalcomp = temp->getDomain().size();
-
- //local vnodes get starting position for complex field subdomains
- int *cidx = new int[Ippl::getNodes()];
- int *cidy = new int[Ippl::getNodes()];
- int *cidz = new int[Ippl::getNodes()];
-
- for (typename Layout_t::const_iterator_iv i_s = temp->getLayout().begin_iv(); i_s != temp->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- cidx[node] = tmp[0].min();
- cidy[node] = tmp[1].min();
- cidz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position for complex field subdomains
- for (typename Layout_t::iterator_dv remote = temp->getLayout().begin_rdv(); remote != temp->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- cidx[node] = tmp[0].min();
- cidy[node] = tmp[1].min();
- cidz[node] = tmp[2].min();
- }
-
- //local vnodes get starting position for real field subdomains
- int *idx = new int[Ippl::getNodes()];
- int *idy = new int[Ippl::getNodes()];
- int *idz = new int[Ippl::getNodes()];
-
- for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
- Domain_t tmp = (*i_s).second->getDomain();
- int node = (*i_s).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //remote vnodes get starting position for real field subdomains
- for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
- Domain_t tmp = (*remote).second->getDomain();
- int node = (*remote).second->getNode();
- idx[node] = tmp[0].min();
- idy[node] = tmp[1].min();
- idz[node] = tmp[2].min();
- }
-
- //int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
- //int cid[3] = {cidx[Ippl::myNode()], cidy[Ippl::myNode()], cidz[Ippl::myNode()]};
-
- //do the FFT on GPU
- int ierr;
- void *real_ptr, *comp_ptr;
-
- //DKS part
- if (Ippl::myNode() == 0) {
-
- //allocate memory on device for real and complex arrays
- real_ptr = base.allocateMemory<T>(totalreal, ierr);
- comp_ptr = base.allocateMemory<Complex_t>(totalcomp, ierr);
-
- if (Ippl::getNodes() > 1) {
- base.gather3DData( comp_ptr, localcomp, sizecomp, MPI_DOUBLE_COMPLEX, NC_g, NC_l,
- cidx, cidy, cidz, Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ //get sizes of global domains and local subdomains
+ int NR_l[Dim], NR_g[Dim], NC_g[Dim];
+ for (size_t d=0; d<Dim; d++) {
+ NR_l[d] = (int)rldf->size(d);
+ NR_g[d] = (int)tempR->getDomain()[d].length();
+ NC_g[d] = NR_g[d];
+ }
+ NC_g[0] = (NC_g[0] / 2) + 1;
+
+ //get sizes of global and local domains
+ int totalreal = tempR->getDomain().size();
+
+ //local vnodes get starting position for real field subdomains
+ int *idx = new int[Ippl::getNodes()];
+ int *idy = new int[Ippl::getNodes()];
+ int *idz = new int[Ippl::getNodes()];
+ for (typename Layout_t::const_iterator_iv i_s = tempR->getLayout().begin_iv(); i_s != tempR->getLayout().end_iv(); ++i_s) {
+ Domain_t tmp = (*i_s).second->getDomain();
+ int node = (*i_s).second->getNode();
+ idx[node] = tmp[0].min();
+ idy[node] = tmp[1].min();
+ idz[node] = tmp[2].min();
+ }
+
+ //remote vnodes get starting position for real field subdomains
+ for (typename Layout_t::iterator_dv remote = tempR->getLayout().begin_rdv(); remote != tempR->getLayout().end_rdv(); ++remote) {
+ Domain_t tmp = (*remote).second->getDomain();
+ int node = (*remote).second->getNode();
+ idx[node] = tmp[0].min();
+ idy[node] = tmp[1].min();
+ idz[node] = tmp[2].min();
+ }
+
+ int id[3] = {idx[Ippl::myNode()], idy[Ippl::myNode()], idz[Ippl::myNode()]};
+
+ /* DKS part */
+ if (Ippl::myNode() == 0) {
+
+ //call real to complex fft
+ dksbase.callC2RFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g, streamId);
+
+ //normalize
+ if (direction == +1)
+ dksbase.callNormalizeC2RFFT(real_ptr, nTransformDims, (int*)NR_g, streamId);
+
+ if (Ippl::getNodes() > 1) {
+ dksbase.syncDevice();
+ MPI_Barrier(Ippl::getComm());
+ /*
+ dksbase.scatter3DData(real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l,
+ idx, idy, idz, Ippl::getNodes(), Ippl::myNode(),
+ 0, Ippl::getComm() );
+ */
+ dksbase.scatter3DDataAsync<T>(real_ptr, localreal, NR_g, NR_l, id);
+ MPI_Barrier(Ippl::getComm());
+ dksbase.syncDevice();
+ MPI_Barrier(Ippl::getComm());
+ } else {
+ //read real data from device
+ dksbase.readDataAsync<T>(real_ptr, localreal, totalreal, streamId);
+ dksbase.syncDevice();
+ //dksbase.readData<T>(real_ptr, localreal, totalreal);
+ }
} else {
- //write real data to device
- ierr = base.writeData<Complex_t>(comp_ptr, localcomp, totalcomp);
- }
-
- //call real to complex fft
- base.callC2RFFT(real_ptr, comp_ptr, nTransformDims, (int*)NR_g);
- if (direction == -1)
- base.callNormalizeC2RFFT(real_ptr, nTransformDims, (int*)NR_g);
- if (Ippl::getNodes() > 1) {
- MPI_Barrier(Ippl::getComm());
- base.scatter3DData(real_ptr, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ //receive data from GPU controled by root process
+ MPI_Barrier(Ippl::getComm());
+ /*
+ dksbase.scatter3DData(NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
+ Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
+ */
+ dksbase.scatter3DDataAsync<T>(real_ptr, localreal, NR_g, NR_l, id);
+ MPI_Barrier(Ippl::getComm());
+ dksbase.syncDevice();
+ MPI_Barrier(Ippl::getComm());
+ }
+ /* end dks part */
- } else {
- //read real data from device
- base.readData<T>(real_ptr, localreal, totalreal);
+ // Now assign into output Field, and compress last temp's storage:
+ if (tempR != &g) {
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+ if (this->compressTemps()) *tempR = 0;
}
- //free device memory
- base.freeMemory<T>(real_ptr, totalreal);
- base.freeMemory<Complex_t>(comp_ptr, totalcomp);
- } else {
-
- //send data via gatherv to gpu controled by root process
- base.gather3DData( NULL, localcomp, sizecomp, MPI_DOUBLE_COMPLEX, NC_g, NC_l, cidx, cidy, cidz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
- //receive data from GPU controled by root process
- MPI_Barrier(Ippl::getComm());
- base.scatter3DData(NULL, localreal, sizereal, MPI_DOUBLE, NR_g, NR_l, idx, idy, idz,
- Ippl::getNodes(), Ippl::myNode(), 0, Ippl::getComm() );
-
- }
- //end dks part
-
- // Now assign into output Field, and compress last temp's storage:
- if (tempR != &g) {
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
- if (this->compressTemps()) *tempR = 0;
- if (this->compressTemps() && temp != &f) *temp = 0;
- }
+ // finish timing the whole mess
+}
+#endif
- // Normalize:
- //if (direction == +1) g = g * this->getNormFact();
- // finish timing the whole mess
- IpplTimings::stopTimer(tottimer);
-}
-#else
-*/
template <size_t Dim, class T>
void
FFT<RCTransform,Dim,T>::transform(
- int direction,
- typename FFT<RCTransform,Dim,T>::ComplexField_t& f,
- typename FFT<RCTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<RCTransform,Dim,T>::ComplexField_t& f,
+ typename FFT<RCTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- FFTDBG(Inform tmsg("FFT-RC-reverse"));
+ FFTDBG(Inform tmsg("FFT-RC-reverse"));
+
+ // indicate we're doing another FFT
+ // INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(complexDomain_m,in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true);
+
+ // Common loop iterate and other vars:
+ size_t d;
+ size_t idim; // idim loops over the number of transform dims.
+ size_t nTransformDims = this->numTransformDims();
+
+ // proceed with the complex-to-complex transforms
+
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+
+ // Local work array passed to FFT:
+ Complex_t* localdata;
+
+ // Loop over all dimensions to be transformed except last one:
+ for (idim = nTransformDims-1; idim != 0; --idim) {
+
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == nTransformDims-1 && !constInput) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<RCTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ FFTDBG(tmsg << "Doing complex->complex transpose into field ");
+ FFTDBG(tmsg << "with layout = "<<tempFields_m[idim]->getLayout()<<std::endl);
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f)
+ *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+
+ FFTDBG(tmsg << "Doing complex->complex fft of other dimension .." << std::endl);
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+
+ // make sure we are uncompressed
+ ldf->Uncompress();
+
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d)
+ nstrips *= ldf->size(d);
+
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ //this->getEngine().callFFT(idim, direction, localdata);
+ this->getEngine().callFFT(idim, -1, localdata);
+
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // handle last CR transform separately
+ idim = 0;
+
+ // see if we can put final result directly into g
+ RealField_t* tempR;
+ bool skipTemp = true;
- // indicate we're doing another FFT
- // INCIPPLSTAT(incFFTs);
+ // more rigorous match required here; check that layouts are identical
+ if (!(out_layout == *tempRLayout_m)) {
+ skipTemp = false;
+ } else {
+ for (d=0; d<Dim; ++d)
+ if (out_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
+ skipTemp = false;
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(complexDomain_m,in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true);
+ if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
+ skipTemp = false;
- // Common loop iterate and other vars:
- size_t d;
- size_t idim; // idim loops over the number of transform dims.
- size_t nTransformDims = this->numTransformDims();
+ // also make sure there are no guard cells
+ if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
+ skipTemp = false;
+ }
- // proceed with the complex-to-complex transforms
+ if (skipTemp)
+ tempR = &g;
+ else
+ tempR = tempRField_m;
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
+ skipTemp = true;
+ if (nTransformDims == 1 && !constInput) {
+ // only one CR transform
+ // see if we really need to transpose input data
+ // more rigorous match required here; check that layouts are identical
+ if (!(in_layout == *tempLayouts_m[0])) {
+ skipTemp = false;
+ } else {
+ for (d=0; d<Dim; ++d)
+ if (in_layout.getDistribution(d) !=
+ tempLayouts_m[0]->getDistribution(d))
+ skipTemp = false;
- // Local work array passed to FFT:
- Complex_t* localdata;
+ if ( in_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
+ skipTemp = false;
- // Loop over all dimensions to be transformed except last one:
- for (idim = nTransformDims-1; idim != 0; --idim) {
+ // also make sure there are no guard cells
+ if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
+ skipTemp = false;
+ }
+ } else { // cannot skip transpose
+ skipTemp = false;
+ }
- // Now do the serial transforms along this dimension:
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == nTransformDims-1 && !constInput) {
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<RCTransform,Dim,T>::nullGC) );
- }
+ if (!skipTemp) {
+ // transpose and permute to complex Field with transform dim first
+ FFTDBG(tmsg << "Doing final complex->complex transpose into field ");
+ FFTDBG(tmsg << "with layout = "<<tempFields_m[0]->getLayout()<<std::endl);
+ (*tempFields_m[0])[tempLayouts_m[0]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- FFTDBG(tmsg << "Doing complex->complex transpose into field ");
- FFTDBG(tmsg << "with layout = "<<tempFields_m[idim]->getLayout()<<std::endl);
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f)
- *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
+ // compress previous iterates storage
+ if (this->compressTemps() && temp != &f)
+ *temp = 0;
+ temp = tempFields_m[0];
}
- FFTDBG(tmsg << "Doing complex->complex fft of other dimension .." << std::endl);
+ FFTDBG(tmsg << "Doing complex->real fft of final dimension ..." << std::endl);
// Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
-
- // make sure we are uncompressed
- ldf->Uncompress();
-
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d)
- nstrips *= ldf->size(d);
-
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- //this->getEngine().callFFT(idim, direction, localdata);
- this->getEngine().callFFT(idim, -1, localdata);
-
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
+
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
+
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
+
+ // number of strips should be the same for real and complex LFields!
+ int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
+ for (d=1; d<Dim; ++d)
+ nstrips *= rldf->size(d);
+
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D complex-to-real FFT:
+ // note that complex-to-real FFT direction is always -1
+ this->getEngine().callFFT(idim, -1, localcomp);
+
+ // move the data into the real strip, which is two reals shorter
+ for (int ilen=0; ilen<lengthreal; ilen+=2) {
+ localreal[ilen] = real(localcomp[ilen/2]);
+ localreal[ilen+1] = imag(localcomp[ilen/2]);
+ }
+
+ // advance the data pointers
+ localreal += lengthreal;
+ localcomp += lengthcomp;
+ } // loop over 1D strips
} // loop over all the LFields
- } // loop over all transformed dimensions
-
- // handle last CR transform separately
- idim = 0;
-
- // see if we can put final result directly into g
- RealField_t* tempR;
- bool skipTemp = true;
-
- // more rigorous match required here; check that layouts are identical
- if (!(out_layout == *tempRLayout_m)) {
- skipTemp = false;
- } else {
- for (d=0; d<Dim; ++d)
- if (out_layout.getDistribution(d) != tempRLayout_m->getDistribution(d))
- skipTemp = false;
-
- if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
-
- if (skipTemp)
- tempR = &g;
- else
- tempR = tempRField_m;
-
- skipTemp = true;
- if (nTransformDims == 1 && !constInput) {
- // only one CR transform
- // see if we really need to transpose input data
- // more rigorous match required here; check that layouts are identical
- if (!(in_layout == *tempLayouts_m[0])) {
- skipTemp = false;
- } else {
- for (d=0; d<Dim; ++d)
- if (in_layout.getDistribution(d) !=
- tempLayouts_m[0]->getDistribution(d))
- skipTemp = false;
-
- if ( in_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
- skipTemp = false;
-
- // also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,Dim,T>::nullGC))
- skipTemp = false;
- }
- } else { // cannot skip transpose
- skipTemp = false;
- }
-
-
- if (!skipTemp) {
- // transpose and permute to complex Field with transform dim first
- FFTDBG(tmsg << "Doing final complex->complex transpose into field ");
- FFTDBG(tmsg << "with layout = "<<tempFields_m[0]->getLayout()<<std::endl);
- (*tempFields_m[0])[tempLayouts_m[0]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
// compress previous iterates storage
if (this->compressTemps() && temp != &f)
- *temp = 0;
- temp = tempFields_m[0];
- }
-
-
- FFTDBG(tmsg << "Doing complex->real fft of final dimension ..." << std::endl);
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
-
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
+ *temp = 0;
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &g) {
- // number of strips should be the same for real and complex LFields!
- int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
- for (d=1; d<Dim; ++d)
- nstrips *= rldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D complex-to-real FFT:
- // note that complex-to-real FFT direction is always -1
- this->getEngine().callFFT(idim, -1, localcomp);
+ // Now assign into output Field, and compress last temp's storage:
+ FFTDBG(tmsg << "Doing cleanup real->real transpose into return ");
+ FFTDBG(tmsg << "with layout = " << g.getLayout() << std::endl);
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
- // move the data into the real strip, which is two reals shorter
- for (int ilen=0; ilen<lengthreal; ilen+=2) {
- localreal[ilen] = real(localcomp[ilen/2]);
- localreal[ilen+1] = imag(localcomp[ilen/2]);
- }
+ if (this->compressTemps())
+ *tempR = 0;
- // advance the data pointers
- localreal += lengthreal;
- localcomp += lengthcomp;
- } // loop over 1D strips
- } // loop over all the LFields
-
- // compress previous iterates storage
- if (this->compressTemps() && temp != &f)
- *temp = 0;
-
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &g) {
-
-
- // Now assign into output Field, and compress last temp's storage:
- FFTDBG(tmsg << "Doing cleanup real->real transpose into return ");
- FFTDBG(tmsg << "with layout = " << g.getLayout() << std::endl);
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
-
- if (this->compressTemps())
- *tempR = 0;
-
- }
+ }
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
- // finish up timing the whole mess
+ // finish up timing the whole mess
}
-//#endif
//=============================================================================
// 1D FFT RCTransform Constructors
@@ -2461,28 +1721,28 @@ FFT<RCTransform,Dim,T>::transform(
template <class T>
FFT<RCTransform,1U,T>::FFT(
- const typename FFT<RCTransform,1U,T>::Domain_t& rdomain,
- const typename FFT<RCTransform,1U,T>::Domain_t& cdomain,
- const bool transformTheseDims[1U], const bool& compressTemps)
- : FFTBase<1U,T>(FFT<RCTransform,1U,T>::rcFFT, rdomain,
- transformTheseDims, compressTemps),
+ const typename FFT<RCTransform,1U,T>::Domain_t& rdomain,
+ const typename FFT<RCTransform,1U,T>::Domain_t& cdomain,
+ const bool transformTheseDims[1U], const bool& compressTemps)
+: FFTBase<1U,T>(FFT<RCTransform,1U,T>::rcFFT, rdomain,
+ transformTheseDims, compressTemps),
complexDomain_m(cdomain)
{
- size_t nTransformDims = 1U;
- // get axis length
- int length;
- length = rdomain[0].length();
-
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::rcFFT; // first transform is real-to-complex
- T& normFact = this->getNormFact();
- normFact = 1.0 / length;
-
- // set up FFT Engine
- this->getEngine().setup(nTransformDims, &transformType, &length);
- // set up the temporary fields
- setup();
+ size_t nTransformDims = 1U;
+ // get axis length
+ int length;
+ length = rdomain[0].length();
+
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::rcFFT; // first transform is real-to-complex
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / length;
+
+ // set up FFT Engine
+ this->getEngine().setup(nTransformDims, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -2492,31 +1752,31 @@ FFT<RCTransform,1U,T>::FFT(
template <class T>
FFT<RCTransform,1U,T>::FFT(
- const typename FFT<RCTransform,1U,T>::Domain_t& rdomain,
- const typename FFT<RCTransform,1U,T>::Domain_t& cdomain,
- const bool& compressTemps)
- : FFTBase<1U,T>(FFT<RCTransform,1U,T>::rcFFT, rdomain, compressTemps),
+ const typename FFT<RCTransform,1U,T>::Domain_t& rdomain,
+ const typename FFT<RCTransform,1U,T>::Domain_t& cdomain,
+ const bool& compressTemps)
+: FFTBase<1U,T>(FFT<RCTransform,1U,T>::rcFFT, rdomain, compressTemps),
complexDomain_m(cdomain)
{
- // Tau profiling
+ // Tau profiling
- // get axis length
- int length;
- length = rdomain[0].length();
+ // get axis length
+ int length;
+ length = rdomain[0].length();
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::rcFFT; // first transform is real-to-complex
- T& normFact = this->getNormFact();
- normFact = 1.0 / length;
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::rcFFT; // first transform is real-to-complex
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / length;
- // set up FFT Engine
- this->getEngine().setup(1U, &transformType, &length);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(1U, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -2528,35 +1788,35 @@ template <class T>
void
FFT<RCTransform,1U,T>::setup(void) {
- // Tau profiling
+ // Tau profiling
- // check that domain lengths agree between real and complex domains
- const Domain_t& domain = this->getDomain();
- bool match = true;
- // real array length n, complex array length n/2+1
- if ( complexDomain_m[0].length() !=
- (domain[0].length()/2 + 1) ) match = false;
- PInsist(match,
- "Domains provided for real and complex Fields are incompatible!");
+ // check that domain lengths agree between real and complex domains
+ const Domain_t& domain = this->getDomain();
+ bool match = true;
+ // real array length n, complex array length n/2+1
+ if ( complexDomain_m[0].length() !=
+ (domain[0].length()/2 + 1) ) match = false;
+ PInsist(match,
+ "Domains provided for real and complex Fields are incompatible!");
- // generate layout for temporary real Field
- tempRLayout_m = new Layout_t(domain[0], PARALLEL, 1);
- // create temporary real Field
- this->tempRField_m = new RealField_t(*tempRLayout_m);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) this->tempRField_m->Uncompress();
+ // generate layout for temporary real Field
+ tempRLayout_m = new Layout_t(domain[0], PARALLEL, 1);
+ // create temporary real Field
+ this->tempRField_m = new RealField_t(*tempRLayout_m);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) this->tempRField_m->Uncompress();
- // generate temporary field layout
- tempLayouts_m = new Layout_t(complexDomain_m[0], PARALLEL, 1);
- // create temporary complex Field
- tempFields_m = new ComplexField_t(*tempLayouts_m);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) this->tempFields_m->Uncompress();
+ // generate temporary field layout
+ tempLayouts_m = new Layout_t(complexDomain_m[0], PARALLEL, 1);
+ // create temporary complex Field
+ tempFields_m = new ComplexField_t(*tempLayouts_m);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) this->tempFields_m->Uncompress();
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -2566,16 +1826,16 @@ FFT<RCTransform,1U,T>::setup(void) {
template <class T>
FFT<RCTransform,1U,T>::~FFT(void) {
- // Tau profiling
+ // Tau profiling
- // delete temporary fields and layouts
- delete tempFields_m;
- delete tempLayouts_m;
- delete this->tempRField_m;
- delete tempRLayout_m;
+ // delete temporary fields and layouts
+ delete tempFields_m;
+ delete tempLayouts_m;
+ delete this->tempRField_m;
+ delete tempRLayout_m;
}
//-----------------------------------------------------------------------------
@@ -2585,103 +1845,103 @@ FFT<RCTransform,1U,T>::~FFT(void) {
template <class T>
void
FFT<RCTransform,1U,T>::transform(
- int direction,
- typename FFT<RCTransform,1U,T>::RealField_t& f,
- typename FFT<RCTransform,1U,T>::ComplexField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<RCTransform,1U,T>::RealField_t& f,
+ typename FFT<RCTransform,1U,T>::ComplexField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(complexDomain_m,out_dom), true);
-
- // Common loop iterate and other vars:
- RealField_t* tempR = this->tempRField_m; // Field* management aid
- if (!constInput) {
- // see if we can use input field f as a temporary
- bool skipTemp = true;
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(complexDomain_m,out_dom), true);
+
+ // Common loop iterate and other vars:
+ RealField_t* tempR = this->tempRField_m; // Field* management aid
+ if (!constInput) {
+ // see if we can use input field f as a temporary
+ bool skipTemp = true;
+ // more rigorous match required here; check that layouts are identical
+ if ( !(in_layout == *tempRLayout_m) ) skipTemp = false;
+ if ( in_layout.getDistribution(0) !=
+ tempRLayout_m->getDistribution(0) ) skipTemp = false;
+ if ( in_layout.numVnodes() != tempRLayout_m->numVnodes() )
+ skipTemp = false;
+ // also make sure there are no guard cells
+ if (!(f.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipTemp = false;
+ if (skipTemp) tempR = &f;
+ }
+
+ if (tempR != &f) { // not using input as a temporary
+
+ // assign to real Field with proper layout
+ (*tempR) = f;
+
+ }
+
+ // Field* for temp Field management:
+ ComplexField_t* temp = tempFields_m;
+ // see if we can put final result directly into g
+ bool skipFinal = true;
// more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempRLayout_m) ) skipTemp = false;
- if ( in_layout.getDistribution(0) !=
- tempRLayout_m->getDistribution(0) ) skipTemp = false;
- if ( in_layout.numVnodes() != tempRLayout_m->numVnodes() )
- skipTemp = false;
+ if ( !(out_layout == *tempLayouts_m) ) skipFinal = false;
+ if ( out_layout.getDistribution(0) !=
+ tempLayouts_m->getDistribution(0) ) skipFinal = false;
+ if ( out_layout.numVnodes() != tempLayouts_m->numVnodes() )
+ skipFinal = false;
// also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipTemp = false;
- if (skipTemp) tempR = &f;
- }
-
- if (tempR != &f) { // not using input as a temporary
-
- // assign to real Field with proper layout
- (*tempR) = f;
-
- }
-
- // Field* for temp Field management:
- ComplexField_t* temp = tempFields_m;
- // see if we can put final result directly into g
- bool skipFinal = true;
- // more rigorous match required here; check that layouts are identical
- if ( !(out_layout == *tempLayouts_m) ) skipFinal = false;
- if ( out_layout.getDistribution(0) !=
- tempLayouts_m->getDistribution(0) ) skipFinal = false;
- if ( out_layout.numVnodes() != tempLayouts_m->numVnodes() )
- skipFinal = false;
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipFinal = false;
- if (skipFinal) temp = &g;
-
-
- // There should be just one vnode!
- typename RealField_t::const_iterator_if rl_i = tempR->begin_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- if (rl_i != tempR->end_if() && cl_i != temp->end_if()) {
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
+ if (!(g.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipFinal = false;
+ if (skipFinal) temp = &g;
- int lengthreal = rldf->size(0);
- // move the data into the complex strip, which is two reals longer
- for (int ilen=0; ilen<lengthreal; ilen+=2)
- localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
- // Do the 1D real-to-complex FFT:
- // note that real-to-complex FFT direction is always +1
- this->getEngine().callFFT(0, +1, localcomp);
- }
+ // There should be just one vnode!
+ typename RealField_t::const_iterator_if rl_i = tempR->begin_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ if (rl_i != tempR->end_if() && cl_i != temp->end_if()) {
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
+ int lengthreal = rldf->size(0);
+ // move the data into the complex strip, which is two reals longer
+ for (int ilen=0; ilen<lengthreal; ilen+=2)
+ localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
+ // Do the 1D real-to-complex FFT:
+ // note that real-to-complex FFT direction is always +1
+ this->getEngine().callFFT(0, +1, localcomp);
+ }
- // compress temporary storage
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- // skip final assignment and compress if we used g as final temporary
- if (temp != &g) {
+ // compress temporary storage
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
- // Now assign into output Field, and compress last temp's storage:
- g = (*temp);
- if (this->compressTemps()) *temp = 0;
- }
+ // skip final assignment and compress if we used g as final temporary
+ if (temp != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g = (*temp);
+ if (this->compressTemps()) *temp = 0;
+
+ }
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -2691,114 +1951,114 @@ FFT<RCTransform,1U,T>::transform(
template <class T>
void
FFT<RCTransform,1U,T>::transform(
- int direction,
- typename FFT<RCTransform,1U,T>::ComplexField_t& f,
- typename FFT<RCTransform,1U,T>::RealField_t& g,
- const bool& constInput)
+ int direction,
+ typename FFT<RCTransform,1U,T>::ComplexField_t& f,
+ typename FFT<RCTransform,1U,T>::RealField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(complexDomain_m,in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true);
-
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
-
- // see if we can put final result directly into g
- RealField_t* tempR;
- bool skipFinal = true;
- // more rigorous match required here; check that layouts are identical
- if ( !(out_layout == *tempRLayout_m) ) skipFinal = false;
- if ( out_layout.getDistribution(0) !=
- tempRLayout_m->getDistribution(0) ) skipFinal = false;
- if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
- skipFinal = false;
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipFinal = false;
- if (skipFinal)
- tempR = &g;
- else
- tempR = this->tempRField_m;
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(complexDomain_m,in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true);
+
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
- bool skipTemp = true;
- if (!constInput) {
- // only one CR transform
- // see if we really need to transpose input data
+ // see if we can put final result directly into g
+ RealField_t* tempR;
+ bool skipFinal = true;
// more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempLayouts_m) ) skipTemp = false;
- if ( in_layout.getDistribution(0) !=
- tempLayouts_m->getDistribution(0) ) skipTemp = false;
- if ( in_layout.numVnodes() != tempLayouts_m->numVnodes() )
- skipTemp = false;
+ if ( !(out_layout == *tempRLayout_m) ) skipFinal = false;
+ if ( out_layout.getDistribution(0) !=
+ tempRLayout_m->getDistribution(0) ) skipFinal = false;
+ if ( out_layout.numVnodes() != tempRLayout_m->numVnodes() )
+ skipFinal = false;
// also make sure there are no guard cells
- if (!(f.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipTemp = false;
- }
- else { // cannot skip transpose
- skipTemp = false;
- }
+ if (!(g.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipFinal = false;
+ if (skipFinal)
+ tempR = &g;
+ else
+ tempR = this->tempRField_m;
+ bool skipTemp = true;
+ if (!constInput) {
+ // only one CR transform
+ // see if we really need to transpose input data
+ // more rigorous match required here; check that layouts are identical
+ if ( !(in_layout == *tempLayouts_m) ) skipTemp = false;
+ if ( in_layout.getDistribution(0) !=
+ tempLayouts_m->getDistribution(0) ) skipTemp = false;
+ if ( in_layout.numVnodes() != tempLayouts_m->numVnodes() )
+ skipTemp = false;
+ // also make sure there are no guard cells
+ if (!(f.getGC() == FFT<RCTransform,1U,T>::nullGC)) skipTemp = false;
+ }
+ else { // cannot skip transpose
+ skipTemp = false;
+ }
- if (!skipTemp) {
- // assign to complex Field with proper layout
- (*tempFields_m) = (*temp);
- // compress previous iterates storage
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m;
- }
+ if (!skipTemp) {
+ // assign to complex Field with proper layout
+ (*tempFields_m) = (*temp);
+ // compress previous iterates storage
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m;
+ }
- // There should be just one vnode!
- typename RealField_t::const_iterator_if rl_i = tempR->begin_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- if (rl_i != tempR->end_if() && cl_i != temp->end_if()) {
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
-
- int lengthreal = rldf->size(0);
- // Do the 1D complex-to-real FFT:
- // note that complex-to-real FFT direction is always -1
- this->getEngine().callFFT(0, -1, localcomp);
- // move the data into the real strip, which is two reals shorter
- for (int ilen=0; ilen<lengthreal; ilen+=2) {
- localreal[ilen] = real(localcomp[ilen/2]);
- localreal[ilen+1] = imag(localcomp[ilen/2]);
+
+ // There should be just one vnode!
+ typename RealField_t::const_iterator_if rl_i = tempR->begin_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ if (rl_i != tempR->end_if() && cl_i != temp->end_if()) {
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
+
+ int lengthreal = rldf->size(0);
+ // Do the 1D complex-to-real FFT:
+ // note that complex-to-real FFT direction is always -1
+ this->getEngine().callFFT(0, -1, localcomp);
+ // move the data into the real strip, which is two reals shorter
+ for (int ilen=0; ilen<lengthreal; ilen+=2) {
+ localreal[ilen] = real(localcomp[ilen/2]);
+ localreal[ilen+1] = imag(localcomp[ilen/2]);
+ }
}
- }
- // compress previous iterates storage
- if (this->compressTemps() && temp != &f) *temp = 0;
+ // compress previous iterates storage
+ if (this->compressTemps() && temp != &f) *temp = 0;
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &g) {
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &g) {
- // Now assign into output Field, and compress last temp's storage:
- g = (*tempR);
- if (this->compressTemps()) *tempR = 0;
+ // Now assign into output Field, and compress last temp's storage:
+ g = (*tempR);
+ if (this->compressTemps()) *tempR = 0;
- }
+ }
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
- return;
+ return;
}
@@ -2816,59 +2076,59 @@ FFT<RCTransform,1U,T>::transform(
template <size_t Dim, class T>
FFT<SineTransform,Dim,T>::FFT(
- const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
- const typename FFT<SineTransform,Dim,T>::Domain_t& cdomain,
- const bool transformTheseDims[Dim], const bool sineTransformDims[Dim],
- const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain,
- transformTheseDims, compressTemps),
+ const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
+ const typename FFT<SineTransform,Dim,T>::Domain_t& cdomain,
+ const bool transformTheseDims[Dim], const bool sineTransformDims[Dim],
+ const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain,
+ transformTheseDims, compressTemps),
complexDomain_m(&cdomain)
{
- size_t d;
- // store which dimensions get sine transforms and count how many
- numSineTransforms_m = 0;
- for (d=0; d<Dim; ++d) {
- sineTransformDims_m[d] = sineTransformDims[d];
- if (sineTransformDims[d]) {
- PAssert_EQ(transformTheseDims[d], true); // should be marked as a transform dim
- ++numSineTransforms_m;
- }
- }
-
- // construct array of axis lengths for all transform dims
- size_t nTransformDims = this->numTransformDims();
- int* lengths = new int[nTransformDims];
- for (d=0; d<nTransformDims; ++d)
- lengths[d] = rdomain[this->activeDimension(d)].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int* transformTypes = new int[nTransformDims];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- bool foundRC = false;
- for (d=0; d<nTransformDims; ++d) {
- if (sineTransformDims_m[this->activeDimension(d)]) {
- transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
- normFact /= (2.0 * (lengths[d] + 1));
- }
- else if (!foundRC) {
- transformTypes[d] = FFTBase<Dim,T>::rcFFT; // real-to-complex FFT
- normFact /= lengths[d];
- foundRC = true;
- }
- else {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // complex-to-complex FFT
- normFact /= lengths[d];
- }
- }
-
- // set up FFT Engine
- this->getEngine().setup(nTransformDims, transformTypes, lengths);
- delete [] transformTypes;
- delete [] lengths;
- // set up the temporary fields
- setup();
+ size_t d;
+ // store which dimensions get sine transforms and count how many
+ numSineTransforms_m = 0;
+ for (d=0; d<Dim; ++d) {
+ sineTransformDims_m[d] = sineTransformDims[d];
+ if (sineTransformDims[d]) {
+ PAssert_EQ(transformTheseDims[d], true); // should be marked as a transform dim
+ ++numSineTransforms_m;
+ }
+ }
+
+ // construct array of axis lengths for all transform dims
+ size_t nTransformDims = this->numTransformDims();
+ int* lengths = new int[nTransformDims];
+ for (d=0; d<nTransformDims; ++d)
+ lengths[d] = rdomain[this->activeDimension(d)].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int* transformTypes = new int[nTransformDims];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ bool foundRC = false;
+ for (d=0; d<nTransformDims; ++d) {
+ if (sineTransformDims_m[this->activeDimension(d)]) {
+ transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
+ normFact /= (2.0 * (lengths[d] + 1));
+ }
+ else if (!foundRC) {
+ transformTypes[d] = FFTBase<Dim,T>::rcFFT; // real-to-complex FFT
+ normFact /= lengths[d];
+ foundRC = true;
+ }
+ else {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // complex-to-complex FFT
+ normFact /= lengths[d];
+ }
+ }
+
+ // set up FFT Engine
+ this->getEngine().setup(nTransformDims, transformTypes, lengths);
+ delete [] transformTypes;
+ delete [] lengths;
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -2878,51 +2138,51 @@ FFT<SineTransform,Dim,T>::FFT(
template <size_t Dim, class T>
FFT<SineTransform,Dim,T>::FFT(
- const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
- const typename FFT<SineTransform,Dim,T>::Domain_t& cdomain,
- const bool sineTransformDims[Dim], const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain, compressTemps),
+ const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
+ const typename FFT<SineTransform,Dim,T>::Domain_t& cdomain,
+ const bool sineTransformDims[Dim], const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain, compressTemps),
complexDomain_m(&cdomain)
{
- size_t d;
- // store which dimensions get sine transforms and count how many
- numSineTransforms_m = 0;
- for (d=0; d<Dim; ++d) {
- sineTransformDims_m[d] = sineTransformDims[d];
- if (sineTransformDims[d]) ++numSineTransforms_m;
- }
-
- // construct array of axis lengths for all transform dims
- int lengths[Dim];
- for (d=0; d<Dim; ++d)
- lengths[d] = rdomain[d].length();
-
- // construct array of transform types for FFT Engine, compute normalization
- int transformTypes[Dim];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- bool foundRC = false;
- for (d=0; d<Dim; ++d) {
- if (sineTransformDims_m[d]) {
- transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
- normFact /= (2.0 * (lengths[d] + 1));
- }
- else if (!foundRC) {
- transformTypes[d] = FFTBase<Dim,T>::rcFFT; // real-to-complex FFT
- normFact /= lengths[d];
- foundRC = true;
- }
- else {
- transformTypes[d] = FFTBase<Dim,T>::ccFFT; // complex-to-complex FFT
- normFact /= lengths[d];
+ size_t d;
+ // store which dimensions get sine transforms and count how many
+ numSineTransforms_m = 0;
+ for (d=0; d<Dim; ++d) {
+ sineTransformDims_m[d] = sineTransformDims[d];
+ if (sineTransformDims[d]) ++numSineTransforms_m;
}
- }
- // set up FFT Engine
- this->getEngine().setup(Dim, transformTypes, lengths);
- // set up the temporary fields
- setup();
+ // construct array of axis lengths for all transform dims
+ int lengths[Dim];
+ for (d=0; d<Dim; ++d)
+ lengths[d] = rdomain[d].length();
+
+ // construct array of transform types for FFT Engine, compute normalization
+ int transformTypes[Dim];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ bool foundRC = false;
+ for (d=0; d<Dim; ++d) {
+ if (sineTransformDims_m[d]) {
+ transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
+ normFact /= (2.0 * (lengths[d] + 1));
+ }
+ else if (!foundRC) {
+ transformTypes[d] = FFTBase<Dim,T>::rcFFT; // real-to-complex FFT
+ normFact /= lengths[d];
+ foundRC = true;
+ }
+ else {
+ transformTypes[d] = FFTBase<Dim,T>::ccFFT; // complex-to-complex FFT
+ normFact /= lengths[d];
+ }
+ }
+
+ // set up FFT Engine
+ this->getEngine().setup(Dim, transformTypes, lengths);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -2933,42 +2193,42 @@ FFT<SineTransform,Dim,T>::FFT(
template <size_t Dim, class T>
FFT<SineTransform,Dim,T>::FFT(
- const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
- const bool sineTransformDims[Dim], const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain,
- sineTransformDims, compressTemps)
+ const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain,
+ const bool sineTransformDims[Dim], const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain,
+ sineTransformDims, compressTemps)
{
- // Tau profiling
+ // Tau profiling
- // store which dimensions get sine transforms and how many
- numSineTransforms_m = this->numTransformDims();
- size_t d;
- for (d=0; d<Dim; ++d)
- sineTransformDims_m[d] = sineTransformDims[d];
+ // store which dimensions get sine transforms and how many
+ numSineTransforms_m = this->numTransformDims();
+ size_t d;
+ for (d=0; d<Dim; ++d)
+ sineTransformDims_m[d] = sineTransformDims[d];
- // construct array of axis lengths
- int* lengths = new int[numSineTransforms_m];
- for (d=0; d<numSineTransforms_m; ++d)
- lengths[d] = rdomain[this->activeDimension(d)].length();
+ // construct array of axis lengths
+ int* lengths = new int[numSineTransforms_m];
+ for (d=0; d<numSineTransforms_m; ++d)
+ lengths[d] = rdomain[this->activeDimension(d)].length();
- // construct array of transform types for FFT Engine, compute normalization
- int* transformTypes = new int[numSineTransforms_m];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- for (d=0; d<numSineTransforms_m; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
- normFact /= (2.0 * (lengths[d] + 1));
- }
+ // construct array of transform types for FFT Engine, compute normalization
+ int* transformTypes = new int[numSineTransforms_m];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ for (d=0; d<numSineTransforms_m; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
+ normFact /= (2.0 * (lengths[d] + 1));
+ }
- // set up FFT Engine
- this->getEngine().setup(numSineTransforms_m, transformTypes, lengths);
- delete [] transformTypes;
- delete [] lengths;
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(numSineTransforms_m, transformTypes, lengths);
+ delete [] transformTypes;
+ delete [] lengths;
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -2978,38 +2238,38 @@ FFT<SineTransform,Dim,T>::FFT(
template <size_t Dim, class T>
FFT<SineTransform,Dim,T>::FFT(
- const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain, const bool& compressTemps)
- : FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain, compressTemps)
+ const typename FFT<SineTransform,Dim,T>::Domain_t& rdomain, const bool& compressTemps)
+: FFTBase<Dim,T>(FFT<SineTransform,Dim,T>::sineFFT, rdomain, compressTemps)
{
- // Tau profiling
+ // Tau profiling
- size_t d;
- // store which dimensions get sine transforms and how many
- numSineTransforms_m = this->numTransformDims();
- for (d=0; d<Dim; ++d)
- sineTransformDims_m[d] = true;
+ size_t d;
+ // store which dimensions get sine transforms and how many
+ numSineTransforms_m = this->numTransformDims();
+ for (d=0; d<Dim; ++d)
+ sineTransformDims_m[d] = true;
- // construct array of axis lengths
- int lengths[Dim];
- for (d=0; d<Dim; ++d)
- lengths[d] = rdomain[d].length();
+ // construct array of axis lengths
+ int lengths[Dim];
+ for (d=0; d<Dim; ++d)
+ lengths[d] = rdomain[d].length();
- // construct array of transform types for FFT Engine, compute normalization
- int transformTypes[Dim];
- T& normFact = this->getNormFact();
- normFact = 1.0;
- for (d=0; d<Dim; ++d) {
- transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
- normFact /= (2.0 * (lengths[d] + 1));
- }
+ // construct array of transform types for FFT Engine, compute normalization
+ int transformTypes[Dim];
+ T& normFact = this->getNormFact();
+ normFact = 1.0;
+ for (d=0; d<Dim; ++d) {
+ transformTypes[d] = FFTBase<Dim,T>::sineFFT; // sine transform
+ normFact /= (2.0 * (lengths[d] + 1));
+ }
- // set up FFT Engine
- this->getEngine().setup(Dim, transformTypes, lengths);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(Dim, transformTypes, lengths);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -3021,151 +2281,151 @@ template <size_t Dim, class T>
void
FFT<SineTransform,Dim,T>::setup(void) {
- // Tau profiling
-
-
-
-
- size_t d, dim, activeDim = 0;
- size_t icount;
- Domain_t ndip;
- size_t nTransformDims = this->numTransformDims(); // total number of transforms
- const Domain_t& domain = this->getDomain(); // get real Field domain
-
- // Set up the arrays of temporary Fields and FieldLayouts:
- e_dim_tag serialParallel[Dim]; // Specifies SERIAL, PARALLEL dims in temp
- // make zeroth dimension always SERIAL
- serialParallel[0] = SERIAL;
- // all other dimensions parallel
- for (d=1; d<Dim; ++d)
- serialParallel[d] = PARALLEL;
-
- // do we have a real-to-complex transform to do or not?
- if (nTransformDims > numSineTransforms_m) { // have RC transform
-
- PAssert(complexDomain_m); // This pointer should be initialized!
- // find first non-sine transform dimension; this is rc transform
- bool match = false;
- d=0;
- while (d<Dim && !match) {
- if (this->transformDim(d) && !sineTransformDims_m[d]) {
- activeDim = this->activeDimension(d);
- match = true;
- }
- ++d;
- }
- PAssert_EQ(match, true); // check that we found rc transform dimension
- // compare lengths of real and complex Field domains
- for (d=0; d<Dim; ++d) {
- if (d == activeDim) {
- // real array length n, complex array length n/2+1
- if ( (*complexDomain_m)[d].length() !=
- (domain[d].length()/2 + 1) ) match = false;
- }
- else {
- // real and complex arrays should be same length for all other dims
- if ( (*complexDomain_m)[d].length() !=
- domain[d].length() ) match = false;
- }
- }
- PInsist(match,
- "Domains provided for real and complex Fields are incompatible!");
-
- // set up the real Fields first
- // we will have one for each sine transform, plus one for the rc transform
- tempRLayouts_m = new Layout_t*[numSineTransforms_m+1];
- tempRFields_m = new RealField_t*[numSineTransforms_m+1];
- // loop over the sine transform dimensions
- icount=0;
- for (dim=0; dim<numSineTransforms_m; ++dim, ++icount) {
- // get next dimension to be sine transformed
- while (!sineTransformDims_m[icount]) ++icount;
- PAssert_LT(icount, Dim); // check that icount is valid dimension
- // make new domain with permuted Indexes, icount first
- ndip[0] = domain[icount];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = icount + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = domain[nextDim];
- }
- // generate temporary field layout
- tempRLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
- // create temporary real Field
- tempRFields_m[dim] = new RealField_t(*tempRLayouts_m[dim]);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) (*tempRFields_m[dim]).Uncompress();
- }
+ // Tau profiling
- // build final real Field for rc transform along activeDim
- ndip[0] = domain[activeDim];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = activeDim + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = domain[nextDim];
- }
- // generate temporary field layout
- tempRLayouts_m[numSineTransforms_m] = new Layout_t(ndip, serialParallel,
- this->transVnodes());
- // create temporary real Field
- tempRFields_m[numSineTransforms_m] =
- new RealField_t(*tempRLayouts_m[numSineTransforms_m]);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) (*tempRFields_m[numSineTransforms_m]).Uncompress();
-
- // now create the temporary complex Fields
- size_t numComplex = nTransformDims - numSineTransforms_m;
- // allocate arrays of temp fields and layouts
- tempLayouts_m = new Layout_t*[numComplex];
- tempFields_m = new ComplexField_t*[numComplex];
- icount=0; // reset counter
- for (dim=0; dim<numComplex; ++dim, ++icount) {
- // get next non-sine transform dimension
- while (!this->transformDim(icount) || sineTransformDims_m[icount]) ++icount;
- PAssert_LT(icount, Dim); // check that this is a valid dimension
- // make new domain with permuted Indexes, icount first
- ndip[0] = (*complexDomain_m)[icount];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = icount + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = (*complexDomain_m)[nextDim];
- }
- // generate temporary field layout
- tempLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
- // create temporary complex Field
- tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) (*tempFields_m[dim]).Uncompress();
- }
- }
- else { // sine transforms only
-
- // set up the real Fields
- // we will have one for each sine transform
- tempRLayouts_m = new Layout_t*[numSineTransforms_m];
- tempRFields_m = new RealField_t*[numSineTransforms_m];
- // loop over the sine transform dimensions
- for (dim=0; dim<numSineTransforms_m; ++dim) {
- // get next dimension to be sine transformed
- activeDim = this->activeDimension(dim);
- // make new domain with permuted Indexes, activeDim first
- ndip[0] = domain[activeDim];
- for (d=1; d<Dim; ++d) {
- size_t nextDim = activeDim + d;
- if (nextDim >= Dim) nextDim -= Dim;
- ndip[d] = domain[nextDim];
- }
- // generate temporary field layout
- tempRLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
- // create temporary real Field
- tempRFields_m[dim] = new RealField_t(*tempRLayouts_m[dim]);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) (*tempRFields_m[dim]).Uncompress();
- }
- }
- return;
+ size_t d, dim, activeDim = 0;
+ size_t icount;
+ Domain_t ndip;
+ size_t nTransformDims = this->numTransformDims(); // total number of transforms
+ const Domain_t& domain = this->getDomain(); // get real Field domain
+
+ // Set up the arrays of temporary Fields and FieldLayouts:
+ e_dim_tag serialParallel[Dim]; // Specifies SERIAL, PARALLEL dims in temp
+ // make zeroth dimension always SERIAL
+ serialParallel[0] = SERIAL;
+ // all other dimensions parallel
+ for (d=1; d<Dim; ++d)
+ serialParallel[d] = PARALLEL;
+
+ // do we have a real-to-complex transform to do or not?
+ if (nTransformDims > numSineTransforms_m) { // have RC transform
+
+ PAssert(complexDomain_m); // This pointer should be initialized!
+ // find first non-sine transform dimension; this is rc transform
+ bool match = false;
+ d=0;
+ while (d<Dim && !match) {
+ if (this->transformDim(d) && !sineTransformDims_m[d]) {
+ activeDim = this->activeDimension(d);
+ match = true;
+ }
+ ++d;
+ }
+ PAssert_EQ(match, true); // check that we found rc transform dimension
+ // compare lengths of real and complex Field domains
+ for (d=0; d<Dim; ++d) {
+ if (d == activeDim) {
+ // real array length n, complex array length n/2+1
+ if ( (*complexDomain_m)[d].length() !=
+ (domain[d].length()/2 + 1) ) match = false;
+ }
+ else {
+ // real and complex arrays should be same length for all other dims
+ if ( (*complexDomain_m)[d].length() !=
+ domain[d].length() ) match = false;
+ }
+ }
+ PInsist(match,
+ "Domains provided for real and complex Fields are incompatible!");
+
+ // set up the real Fields first
+ // we will have one for each sine transform, plus one for the rc transform
+ tempRLayouts_m = new Layout_t*[numSineTransforms_m+1];
+ tempRFields_m = new RealField_t*[numSineTransforms_m+1];
+ // loop over the sine transform dimensions
+ icount=0;
+ for (dim=0; dim<numSineTransforms_m; ++dim, ++icount) {
+ // get next dimension to be sine transformed
+ while (!sineTransformDims_m[icount]) ++icount;
+ PAssert_LT(icount, Dim); // check that icount is valid dimension
+ // make new domain with permuted Indexes, icount first
+ ndip[0] = domain[icount];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = icount + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = domain[nextDim];
+ }
+ // generate temporary field layout
+ tempRLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
+ // create temporary real Field
+ tempRFields_m[dim] = new RealField_t(*tempRLayouts_m[dim]);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) (*tempRFields_m[dim]).Uncompress();
+ }
+
+ // build final real Field for rc transform along activeDim
+ ndip[0] = domain[activeDim];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = activeDim + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = domain[nextDim];
+ }
+ // generate temporary field layout
+ tempRLayouts_m[numSineTransforms_m] = new Layout_t(ndip, serialParallel,
+ this->transVnodes());
+ // create temporary real Field
+ tempRFields_m[numSineTransforms_m] =
+ new RealField_t(*tempRLayouts_m[numSineTransforms_m]);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) (*tempRFields_m[numSineTransforms_m]).Uncompress();
+
+ // now create the temporary complex Fields
+ size_t numComplex = nTransformDims - numSineTransforms_m;
+ // allocate arrays of temp fields and layouts
+ tempLayouts_m = new Layout_t*[numComplex];
+ tempFields_m = new ComplexField_t*[numComplex];
+ icount=0; // reset counter
+ for (dim=0; dim<numComplex; ++dim, ++icount) {
+ // get next non-sine transform dimension
+ while (!this->transformDim(icount) || sineTransformDims_m[icount]) ++icount;
+ PAssert_LT(icount, Dim); // check that this is a valid dimension
+ // make new domain with permuted Indexes, icount first
+ ndip[0] = (*complexDomain_m)[icount];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = icount + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = (*complexDomain_m)[nextDim];
+ }
+ // generate temporary field layout
+ tempLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
+ // create temporary complex Field
+ tempFields_m[dim] = new ComplexField_t(*tempLayouts_m[dim]);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) (*tempFields_m[dim]).Uncompress();
+ }
+
+ }
+ else { // sine transforms only
+
+ // set up the real Fields
+ // we will have one for each sine transform
+ tempRLayouts_m = new Layout_t*[numSineTransforms_m];
+ tempRFields_m = new RealField_t*[numSineTransforms_m];
+ // loop over the sine transform dimensions
+ for (dim=0; dim<numSineTransforms_m; ++dim) {
+ // get next dimension to be sine transformed
+ activeDim = this->activeDimension(dim);
+ // make new domain with permuted Indexes, activeDim first
+ ndip[0] = domain[activeDim];
+ for (d=1; d<Dim; ++d) {
+ size_t nextDim = activeDim + d;
+ if (nextDim >= Dim) nextDim -= Dim;
+ ndip[d] = domain[nextDim];
+ }
+ // generate temporary field layout
+ tempRLayouts_m[dim] = new Layout_t(ndip, serialParallel, this->transVnodes());
+ // create temporary real Field
+ tempRFields_m[dim] = new RealField_t(*tempRLayouts_m[dim]);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) (*tempRFields_m[dim]).Uncompress();
+ }
+
+ }
+
+ return;
}
//-----------------------------------------------------------------------------
@@ -3175,35 +2435,35 @@ FFT<SineTransform,Dim,T>::setup(void) {
template <size_t Dim, class T>
FFT<SineTransform,Dim,T>::~FFT(void) {
- // Tau profiling
+ // Tau profiling
- // delete temporary fields and layouts
- size_t d;
- size_t nTransformDims = this->numTransformDims();
- if (nTransformDims > numSineTransforms_m) {
- for (d=0; d<numSineTransforms_m+1; ++d) {
- delete tempRFields_m[d];
- delete tempRLayouts_m[d];
- }
- size_t numComplex = nTransformDims - numSineTransforms_m;
- for (d=0; d<numComplex; ++d) {
- delete tempFields_m[d];
- delete tempLayouts_m[d];
+ // delete temporary fields and layouts
+ size_t d;
+ size_t nTransformDims = this->numTransformDims();
+ if (nTransformDims > numSineTransforms_m) {
+ for (d=0; d<numSineTransforms_m+1; ++d) {
+ delete tempRFields_m[d];
+ delete tempRLayouts_m[d];
+ }
+ size_t numComplex = nTransformDims - numSineTransforms_m;
+ for (d=0; d<numComplex; ++d) {
+ delete tempFields_m[d];
+ delete tempLayouts_m[d];
+ }
+ delete [] tempFields_m;
+ delete [] tempLayouts_m;
}
- delete [] tempFields_m;
- delete [] tempLayouts_m;
- }
- else {
- for (d=0; d<numSineTransforms_m; ++d) {
- delete tempRFields_m[d];
- delete tempRLayouts_m[d];
+ else {
+ for (d=0; d<numSineTransforms_m; ++d) {
+ delete tempRFields_m[d];
+ delete tempRLayouts_m[d];
+ }
}
- }
- delete [] tempRFields_m;
- delete [] tempRLayouts_m;
+ delete [] tempRFields_m;
+ delete [] tempRLayouts_m;
}
//-----------------------------------------------------------------------------
@@ -3213,268 +2473,268 @@ FFT<SineTransform,Dim,T>::~FFT(void) {
template <size_t Dim, class T>
void
FFT<SineTransform,Dim,T>::transform(
- int direction,
- FFT<SineTransform,Dim,T>::RealField_t& f,
- FFT<SineTransform,Dim,T>::ComplexField_t& g,
- const bool& constInput)
+ int direction,
+ FFT<SineTransform,Dim,T>::RealField_t& f,
+ FFT<SineTransform,Dim,T>::ComplexField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(*complexDomain_m,out_dom), true );
-
- // Common loop iterate and other vars:
- size_t d;
- int icount, activeDim;
- int idim; // idim loops over the number of transform dims.
- size_t nTransformDims = this->numTransformDims();
- // check that there is a real-to-complex transform to do
- PInsist(nTransformDims>numSineTransforms_m,
- "Wrong output Field type for real-to-real transform!!");
-
- // first do all the sine transforms
-
- // Field* management aid
- RealField_t* tempR = &f;
- // Local work array passed to FFT:
- T* localdataR;
-
- // Loop over the dimensions to be sine transformed:
- icount = 0;
- activeDim = 0;
- for (idim = 0; idim != numSineTransforms_m; ++idim, ++icount, ++activeDim) {
-
- // find next sine transform dim
- while (!sineTransformDims_m[icount]) {
- if (this->transformDim(icount)) ++activeDim;
- ++icount;
- }
- PAssert_LT(activeDim, Dim); // check that this is a valid dimension!
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == 0 && !constInput) {
- // get domain for comparison
- const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
- (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = tempRFields_m[idim]; // Field* management aid
- }
-
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
- for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
-
- // Do 1D real-to-real FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(activeDim, direction, localdataR);
- // advance the data pointer
- localdataR += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
- } // loop over all transformed dimensions
-
- // now handle the RC transform separately
-
- // find first non-sine transform dimension
- icount = 0;
- activeDim = 0;
- while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
- if (sineTransformDims_m[icount]) ++activeDim;
- ++icount;
- }
- PAssert_LT(activeDim, Dim); // check that this is a valid dimension!
-
-
- // transpose and permute to final real Field with transform dim first
- int last = numSineTransforms_m;
- (*tempRFields_m[last])[tempRLayouts_m[last]->getDomain()] =
- (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = tempRFields_m[last]; // Field* management aid
-
-
- // Field* for temp Field management:
- ComplexField_t* temp = tempFields_m[0];
- // see if we can put final result directly into g
- int numComplex = nTransformDims-numSineTransforms_m;
- if (numComplex == 1) { // only a single RC transform
- bool skipTemp = true;
- // more rigorous match required here; check that layouts are identical
- if ( !(out_layout == *tempLayouts_m[0]) ) skipTemp = false;
- for (d=0; d<Dim; ++d) {
- if ( out_layout.getDistribution(d) !=
- tempLayouts_m[0]->getDistribution(d) ) skipTemp = false;
- }
- if ( out_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
- skipTemp = false;
- // also make sure there are no guard cells
- if (!(g.getGC() == FFT<SineTransform,Dim,T>::nullGC)) skipTemp = false;
- if (skipTemp) temp = &g;
- }
-
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
-
- int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
- // number of strips should be the same for real and complex LFields!
- for (d=1; d<Dim; ++d) nstrips *= rldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // move the data into the complex strip, which is two reals longer
- for (int ilen=0; ilen<lengthreal; ilen+=2)
- localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
- // Do the 1D real-to-complex FFT:
- // note that real-to-complex FFT direction is always +1
- this->getEngine().callFFT(activeDim, +1, localcomp);
- // advance the data pointers
- localreal += lengthreal;
- localcomp += lengthcomp;
- } // loop over 1D strips
- } // loop over all the LFields
-
-
- // now proceed with the other complex-to-complex transforms
-
- // Local work array passed to FFT:
- Complex_t* localdata;
-
- // Loop over the remaining dimensions to be transformed:
- ++icount;
- ++activeDim;
- for (idim = 1; idim != numComplex; ++idim, ++icount, ++activeDim) {
-
- // find the next non-sine transform dimension
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(*complexDomain_m,out_dom), true );
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int icount, activeDim;
+ int idim; // idim loops over the number of transform dims.
+ size_t nTransformDims = this->numTransformDims();
+ // check that there is a real-to-complex transform to do
+ PInsist(nTransformDims>numSineTransforms_m,
+ "Wrong output Field type for real-to-real transform!!");
+
+ // first do all the sine transforms
+
+ // Field* management aid
+ RealField_t* tempR = &f;
+ // Local work array passed to FFT:
+ T* localdataR;
+
+ // Loop over the dimensions to be sine transformed:
+ icount = 0;
+ activeDim = 0;
+ for (idim = 0; idim != numSineTransforms_m; ++idim, ++icount, ++activeDim) {
+
+ // find next sine transform dim
+ while (!sineTransformDims_m[icount]) {
+ if (this->transformDim(icount)) ++activeDim;
+ ++icount;
+ }
+ PAssert_LT(activeDim, Dim); // check that this is a valid dimension!
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == 0 && !constInput) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
+ (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = tempRFields_m[idim]; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
+ for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
+
+ // Do 1D real-to-real FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(activeDim, direction, localdataR);
+ // advance the data pointer
+ localdataR += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // now handle the RC transform separately
+
+ // find first non-sine transform dimension
+ icount = 0;
+ activeDim = 0;
while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
- if (sineTransformDims_m[icount]) ++activeDim;
- ++icount;
+ if (sineTransformDims_m[icount]) ++activeDim;
+ ++icount;
}
PAssert_LT(activeDim, Dim); // check that this is a valid dimension!
- // Now do the serial transforms along this dimension:
- bool skipTranspose = false;
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into g
- if (idim == numComplex-1) {
- // get the domain for comparison
- const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
- (out_dom[0].length() == last_dom[0].length()) &&
- (out_layout.getDistribution(0) == SERIAL) &&
- (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
+ // transpose and permute to final real Field with transform dim first
+ int last = numSineTransforms_m;
+ (*tempRFields_m[last])[tempRLayouts_m[last]->getDomain()] =
+ (*tempR)[tempR->getLayout().getDomain()];
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
- }
- else if (idim == numComplex-1) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using g instead
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = tempRFields_m[last]; // Field* management aid
- // transpose and permute to Field with transform dim first
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *temp = 0;
- temp = &g; // Field* management aid
+ // Field* for temp Field management:
+ ComplexField_t* temp = tempFields_m[0];
+ // see if we can put final result directly into g
+ int numComplex = nTransformDims-numSineTransforms_m;
+ if (numComplex == 1) { // only a single RC transform
+ bool skipTemp = true;
+ // more rigorous match required here; check that layouts are identical
+ if ( !(out_layout == *tempLayouts_m[0]) ) skipTemp = false;
+ for (d=0; d<Dim; ++d) {
+ if ( out_layout.getDistribution(d) !=
+ tempLayouts_m[0]->getDistribution(d) ) skipTemp = false;
+ }
+ if ( out_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
+ skipTemp = false;
+ // also make sure there are no guard cells
+ if (!(g.getGC() == FFT<SineTransform,Dim,T>::nullGC)) skipTemp = false;
+ if (skipTemp) temp = &g;
}
// Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(activeDim, direction, localdata);
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
+
+ int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
+ // number of strips should be the same for real and complex LFields!
+ for (d=1; d<Dim; ++d) nstrips *= rldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // move the data into the complex strip, which is two reals longer
+ for (int ilen=0; ilen<lengthreal; ilen+=2)
+ localcomp[ilen/2] = Complex_t(localreal[ilen],localreal[ilen+1]);
+ // Do the 1D real-to-complex FFT:
+ // note that real-to-complex FFT direction is always +1
+ this->getEngine().callFFT(activeDim, +1, localcomp);
+ // advance the data pointers
+ localreal += lengthreal;
+ localcomp += lengthcomp;
+ } // loop over 1D strips
} // loop over all the LFields
- } // loop over all transformed dimensions
-
- // skip final assignment and compress if we used g as final temporary
- if (temp != &g) {
- // Now assign into output Field, and compress last temp's storage:
- g[out_dom] = (*temp)[temp->getLayout().getDomain()];
- if (this->compressTemps()) *temp = 0;
-
- }
+ // now proceed with the other complex-to-complex transforms
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
+ // Local work array passed to FFT:
+ Complex_t* localdata;
- return;
+ // Loop over the remaining dimensions to be transformed:
+ ++icount;
+ ++activeDim;
+ for (idim = 1; idim != numComplex; ++idim, ++icount, ++activeDim) {
+
+ // find the next non-sine transform dimension
+ while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
+ if (sineTransformDims_m[icount]) ++activeDim;
+ ++icount;
+ }
+ PAssert_LT(activeDim, Dim); // check that this is a valid dimension!
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into g
+ if (idim == numComplex-1) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
+ (out_dom[0].length() == last_dom[0].length()) &&
+ (out_layout.getDistribution(0) == SERIAL) &&
+ (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+ else if (idim == numComplex-1) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using g instead
+
+ // transpose and permute to Field with transform dim first
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *temp = 0;
+ temp = &g; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(activeDim, direction, localdata);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used g as final temporary
+ if (temp != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g[out_dom] = (*temp)[temp->getLayout().getDomain()];
+ if (this->compressTemps()) *temp = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
+
+ return;
}
//-----------------------------------------------------------------------------
@@ -3484,274 +2744,274 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
void
FFT<SineTransform,Dim,T>::transform(
- int direction,
- FFT<SineTransform,Dim,T>::ComplexField_t& f,
- FFT<SineTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ int direction,
+ FFT<SineTransform,Dim,T>::ComplexField_t& f,
+ FFT<SineTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- // INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(*complexDomain_m,in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true );
-
- // Common loop iterate and other vars:
- size_t d;
- int icount, activeDim;
- int idim; // idim loops over the number of transform dims.
- size_t nTransformDims = this->numTransformDims();
-
- // proceed with the complex-to-complex transforms
-
- // Field* for temp Field management:
- ComplexField_t* temp = &f;
- // Local work array passed to FFT:
- Complex_t* localdata;
-
- // Loop over all dimensions to be non-sine transformed except last one:
- int numComplex = nTransformDims - numSineTransforms_m;
- icount = Dim-1; // start with last dimension
- activeDim = nTransformDims-1;
- for (idim = numComplex-1; idim != 0; --idim, --icount, --activeDim) {
+ // indicate we're doing another FFT
+ // INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(*complexDomain_m,in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true );
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int icount, activeDim;
+ int idim; // idim loops over the number of transform dims.
+ size_t nTransformDims = this->numTransformDims();
+
+ // proceed with the complex-to-complex transforms
+
+ // Field* for temp Field management:
+ ComplexField_t* temp = &f;
+ // Local work array passed to FFT:
+ Complex_t* localdata;
+
+ // Loop over all dimensions to be non-sine transformed except last one:
+ int numComplex = nTransformDims - numSineTransforms_m;
+ icount = Dim-1; // start with last dimension
+ activeDim = nTransformDims-1;
+ for (idim = numComplex-1; idim != 0; --idim, --icount, --activeDim) {
+
+ // find next non-sine transform dim
+ while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
+ if (sineTransformDims_m[icount]) --activeDim;
+ --icount;
+ }
+ PAssert_GE(activeDim, 0); // check that this is a valid dimension!
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == numComplex-1 && !constInput) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m[idim]; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
+ for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ ComplexLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdata = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(activeDim, direction, localdata);
+ // advance the data pointer
+ localdata += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // handle the CR transform separately
// find next non-sine transform dim
while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
- if (sineTransformDims_m[icount]) --activeDim;
- --icount;
+ if (sineTransformDims_m[icount]) --activeDim;
+ --icount;
}
PAssert_GE(activeDim, 0); // check that this is a valid dimension!
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == numComplex-1 && !constInput) {
- // get domain for comparison
- const Domain_t& first_dom = tempLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempFields_m[idim])[tempLayouts_m[idim]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m[idim]; // Field* management aid
- }
-
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename ComplexField_t::const_iterator_if l_i, l_end = temp->end_if();
- for (l_i = temp->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- ComplexLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdata = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(activeDim, direction, localdata);
- // advance the data pointer
- localdata += length;
- } // loop over 1D strips
- } // loop over all the LFields
- } // loop over all transformed dimensions
+ // Temp Field* management aid
+ RealField_t* tempR = tempRFields_m[numSineTransforms_m];
- // handle the CR transform separately
-
- // find next non-sine transform dim
- while (!this->transformDim(icount) || sineTransformDims_m[icount]) {
- if (sineTransformDims_m[icount]) --activeDim;
- --icount;
- }
- PAssert_GE(activeDim, 0); // check that this is a valid dimension!
-
- // Temp Field* management aid
- RealField_t* tempR = tempRFields_m[numSineTransforms_m];
-
- bool skipTemp = true;
- if (numComplex == 1 && !constInput) {
- // only one CR transform
- // see if we really need to transpose input data
- // more rigorous match required here; check that layouts are identical
- if ( !(in_layout == *tempLayouts_m[0]) ) skipTemp = false;
- for (d=0; d<Dim; ++d) {
- if ( in_layout.getDistribution(d) !=
- tempLayouts_m[0]->getDistribution(d) ) skipTemp = false;
- }
- if ( in_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
- skipTemp = false;
- // also make sure there are no guard cells
- if (!(f.getGC() == FFT<SineTransform,Dim,T>::nullGC)) skipTemp = false;
- }
- else { // cannot skip transpose
- skipTemp = false;
- }
-
- if (!skipTemp) {
-
- // transpose and permute to complex Field with transform dim first
- (*tempFields_m[0])[tempLayouts_m[0]->getDomain()] =
- (*temp)[temp->getLayout().getDomain()];
- // compress previous iterates storage
- if (this->compressTemps() && temp != &f) *temp = 0;
- temp = tempFields_m[0];
-
- }
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
- typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
- for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
- // Get the LFields
- RealLField_t* rldf = (*rl_i).second.get();
- ComplexLField_t* cldf = (*cl_i).second.get();
- // make sure we are uncompressed
- rldf->Uncompress();
- cldf->Uncompress();
- // get the raw data pointers
- T* localreal = rldf->getP();
- Complex_t* localcomp = cldf->getP();
-
- int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
- // number of strips should be the same for real and complex LFields!
- for (d=1; d<Dim; ++d) nstrips *= rldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D complex-to-real FFT:
- // note that complex-to-real FFT direction is always -1
- this->getEngine().callFFT(activeDim, -1, localcomp);
- // move the data into the real strip, which is two reals shorter
- for (int ilen=0; ilen<lengthreal; ilen+=2) {
- localreal[ilen] = real(localcomp[ilen/2]);
- localreal[ilen+1] = imag(localcomp[ilen/2]);
- }
- // advance the data pointers
- localreal += lengthreal;
- localcomp += lengthcomp;
- } // loop over 1D strips
- } // loop over all the LFields
-
-
-
- // compress previous iterates storage
- if (this->compressTemps() && temp != &f) *temp = 0;
-
-
- // now do the real-to-real FFTs
-
- // Local work array passed to FFT:
- T* localdataR;
-
- // Loop over the remaining dimensions to be transformed:
- icount = Dim-1; // start with last dimension
- activeDim = nTransformDims - 1;
- for (idim = numSineTransforms_m-1; idim != -1;
- --idim, --icount, --activeDim) {
-
- // find next sine transform dim
- while (!sineTransformDims_m[icount]) {
- if (this->transformDim(icount)) --activeDim;
- --icount;
+ bool skipTemp = true;
+ if (numComplex == 1 && !constInput) {
+ // only one CR transform
+ // see if we really need to transpose input data
+ // more rigorous match required here; check that layouts are identical
+ if ( !(in_layout == *tempLayouts_m[0]) ) skipTemp = false;
+ for (d=0; d<Dim; ++d) {
+ if ( in_layout.getDistribution(d) !=
+ tempLayouts_m[0]->getDistribution(d) ) skipTemp = false;
+ }
+ if ( in_layout.numVnodes() != tempLayouts_m[0]->numVnodes() )
+ skipTemp = false;
+ // also make sure there are no guard cells
+ if (!(f.getGC() == FFT<SineTransform,Dim,T>::nullGC)) skipTemp = false;
}
- PAssert_GE(activeDim, 0); // check that this is a valid dimension!
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into g
- if (idim == 0) {
- // get the domain for comparison
- const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
- (out_dom[0].length() == last_dom[0].length()) &&
- (out_layout.getDistribution(0) == SERIAL) &&
- (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ else { // cannot skip transpose
+ skipTemp = false;
}
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
- (*tempR)[tempR->getLayout().getDomain()];
+ if (!skipTemp) {
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *tempR = 0;
- tempR = tempRFields_m[idim]; // Field* management aid
- }
- else if (idim == 0) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using g instead
-
- // transpose and permute to Field with transform dim first
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+ // transpose and permute to complex Field with transform dim first
+ (*tempFields_m[0])[tempLayouts_m[0]->getDomain()] =
+ (*temp)[temp->getLayout().getDomain()];
+ // compress previous iterates storage
+ if (this->compressTemps() && temp != &f) *temp = 0;
+ temp = tempFields_m[0];
- // Compress out previous iterate's storage:
- if (this->compressTemps()) *tempR = 0;
- tempR = &g; // Field* management aid
}
-
// Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
- for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
-
- // Do 1D complex-to-complex FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(activeDim, direction, localdataR);
- // advance the data pointer
- localdataR += length;
- } // loop over 1D strips
+ typename RealField_t::const_iterator_if rl_i, rl_end = tempR->end_if();
+ typename ComplexField_t::const_iterator_if cl_i = temp->begin_if();
+ for (rl_i = tempR->begin_if(); rl_i != rl_end; ++rl_i, ++cl_i) {
+ // Get the LFields
+ RealLField_t* rldf = (*rl_i).second.get();
+ ComplexLField_t* cldf = (*cl_i).second.get();
+ // make sure we are uncompressed
+ rldf->Uncompress();
+ cldf->Uncompress();
+ // get the raw data pointers
+ T* localreal = rldf->getP();
+ Complex_t* localcomp = cldf->getP();
+
+ int nstrips = 1, lengthreal = rldf->size(0), lengthcomp = cldf->size(0);
+ // number of strips should be the same for real and complex LFields!
+ for (d=1; d<Dim; ++d) nstrips *= rldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D complex-to-real FFT:
+ // note that complex-to-real FFT direction is always -1
+ this->getEngine().callFFT(activeDim, -1, localcomp);
+ // move the data into the real strip, which is two reals shorter
+ for (int ilen=0; ilen<lengthreal; ilen+=2) {
+ localreal[ilen] = real(localcomp[ilen/2]);
+ localreal[ilen+1] = imag(localcomp[ilen/2]);
+ }
+ // advance the data pointers
+ localreal += lengthreal;
+ localcomp += lengthcomp;
+ } // loop over 1D strips
} // loop over all the LFields
- } // loop over all transformed dimensions
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &g) {
- // Now assign into output Field, and compress last temp's storage:
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
- if (this->compressTemps()) *tempR = 0;
-
- }
+ // compress previous iterates storage
+ if (this->compressTemps() && temp != &f) *temp = 0;
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
- return;
+ // now do the real-to-real FFTs
+
+ // Local work array passed to FFT:
+ T* localdataR;
+
+ // Loop over the remaining dimensions to be transformed:
+ icount = Dim-1; // start with last dimension
+ activeDim = nTransformDims - 1;
+ for (idim = numSineTransforms_m-1; idim != -1;
+ --idim, --icount, --activeDim) {
+
+ // find next sine transform dim
+ while (!sineTransformDims_m[icount]) {
+ if (this->transformDim(icount)) --activeDim;
+ --icount;
+ }
+ PAssert_GE(activeDim, 0); // check that this is a valid dimension!
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into g
+ if (idim == 0) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
+ (out_dom[0].length() == last_dom[0].length()) &&
+ (out_layout.getDistribution(0) == SERIAL) &&
+ (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
+ (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *tempR = 0;
+ tempR = tempRFields_m[idim]; // Field* management aid
+ }
+ else if (idim == 0) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using g instead
+
+ // transpose and permute to Field with transform dim first
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps()) *tempR = 0;
+ tempR = &g; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
+ for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
+
+ // Do 1D complex-to-complex FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(activeDim, direction, localdataR);
+ // advance the data pointer
+ localdataR += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+ if (this->compressTemps()) *tempR = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
+
+ return;
}
//-----------------------------------------------------------------------------
@@ -3761,133 +3021,133 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
void
FFT<SineTransform,Dim,T>::transform(
- int direction,
- FFT<SineTransform,Dim,T>::RealField_t& f,
- FFT<SineTransform,Dim,T>::RealField_t& g,
- const bool& constInput)
+ int direction,
+ FFT<SineTransform,Dim,T>::RealField_t& f,
+ FFT<SineTransform,Dim,T>::RealField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true );
-
- // Common loop iterate and other vars:
- size_t d;
- int idim; // idim loops over the number of transform dims.
- int begdim, enddim;
- size_t nTransformDims = this->numTransformDims();
- // check that there is no real-to-complex transform to do
- PInsist(nTransformDims==numSineTransforms_m,
- "Wrong output Field type for real-to-complex transform!!");
-
- // do all the sine transforms
-
- // Field* management aid
- RealField_t* tempR = &f;
- // Local work array passed to FFT:
- T* localdataR;
-
- // Loop over the dimensions to be sine transformed:
- begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
- enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
- for (idim = begdim; idim != enddim; idim+=direction) {
-
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the first temporary and just use f
- if (idim == begdim && !constInput) {
- // get the domain for comparison
- const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into g
- if (idim == enddim-direction) {
- // get the domain for comparison
- const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
- (out_dom[0].length() == last_dom[0].length()) &&
- (out_layout.getDistribution(0) == SERIAL) &&
- (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
- (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = tempRFields_m[idim]; // Field* management aid
- }
- else if (idim == enddim-direction && tempR != &g) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using g instead
-
- // transpose and permute to Field with transform dim first
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = &g; // Field* management aid
- }
-
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
- for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
-
- // Do 1D real-to-real FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(idim, direction, localdataR);
- // advance the data pointer
- localdataR += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
- } // loop over all transformed dimensions
-
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &g) {
-
- // Now assign into output Field, and compress last temp's storage:
- g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
- if (this->compressTemps() && tempR != &f) *tempR = 0;
-
- }
-
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
-
- return;
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true );
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int idim; // idim loops over the number of transform dims.
+ int begdim, enddim;
+ size_t nTransformDims = this->numTransformDims();
+ // check that there is no real-to-complex transform to do
+ PInsist(nTransformDims==numSineTransforms_m,
+ "Wrong output Field type for real-to-complex transform!!");
+
+ // do all the sine transforms
+
+ // Field* management aid
+ RealField_t* tempR = &f;
+ // Local work array passed to FFT:
+ T* localdataR;
+
+ // Loop over the dimensions to be sine transformed:
+ begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
+ enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
+ for (idim = begdim; idim != enddim; idim+=direction) {
+
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the first temporary and just use f
+ if (idim == begdim && !constInput) {
+ // get the domain for comparison
+ const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into g
+ if (idim == enddim-direction) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (out_dom[0].sameBase(last_dom[0])) &&
+ (out_dom[0].length() == last_dom[0].length()) &&
+ (out_layout.getDistribution(0) == SERIAL) &&
+ (g.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
+ (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = tempRFields_m[idim]; // Field* management aid
+ }
+ else if (idim == enddim-direction && tempR != &g) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using g instead
+
+ // transpose and permute to Field with transform dim first
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = &g; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
+ for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
+
+ // Do 1D real-to-real FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(idim, direction, localdataR);
+ // advance the data pointer
+ localdataR += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g[out_dom] = (*tempR)[tempR->getLayout().getDomain()];
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
+
+ return;
}
//-----------------------------------------------------------------------------
@@ -3897,127 +3157,127 @@ FFT<SineTransform,Dim,T>::transform(
template <size_t Dim, class T>
void
FFT<SineTransform,Dim,T>::transform(
- int direction,
- FFT<SineTransform,Dim,T>::RealField_t& f)
+ int direction,
+ FFT<SineTransform,Dim,T>::RealField_t& f)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
-
- // Common loop iterate and other vars:
- size_t d;
- int idim; // idim loops over the number of transform dims.
- int begdim, enddim;
- size_t nTransformDims = this->numTransformDims();
- // check that there is no real-to-complex transform to do
- PInsist(nTransformDims==numSineTransforms_m,
- "Cannot perform real-to-complex transform in-place!!");
-
- // do all the sine transforms
-
- // Field* management aid
- RealField_t* tempR = &f;
- // Local work array passed to FFT:
- T* localdataR;
-
- // Loop over the dimensions to be sine transformed:
- begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
- enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
- for (idim = begdim; idim != enddim; idim+=direction) {
-
- // Now do the serial transforms along this dimension:
-
- bool skipTranspose = false;
- // if this is the first transform dimension, we might be able
- // to skip the transpose into the first temporary Field
- if (idim == begdim) {
- // get domain for comparison
- const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
- (in_dom[0].length() == first_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- // if this is the last transform dimension, we might be able
- // to skip the last temporary and transpose right into f
- if (idim == enddim-direction) {
- // get the domain for comparison
- const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
- (in_dom[0].length() == last_dom[0].length()) &&
- (in_layout.getDistribution(0) == SERIAL) &&
- (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
- }
-
- if (!skipTranspose) {
- // transpose and permute to Field with transform dim first
- (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
- (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = tempRFields_m[idim]; // Field* management aid
- }
- else if (idim == enddim-direction && tempR != &f) {
- // last transform and we can skip the last temporary field
- // so do the transpose here using f instead
-
- // transpose and permute to Field with transform dim first
- f[in_dom] = (*tempR)[tempR->getLayout().getDomain()];
-
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = &f; // Field* management aid
- }
-
-
-
- // Loop over all the Vnodes, working on the LField in each.
- typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
- for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
-
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
-
- // Do 1D real-to-real FFT's on all the strips in the LField:
- int nstrips = 1, length = ldf->size(0);
- for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
- for (int istrip=0; istrip<nstrips; ++istrip) {
- // Do the 1D FFT:
- this->getEngine().callFFT(idim, direction, localdataR);
- // advance the data pointer
- localdataR += length;
- } // loop over 1D strips
- } // loop over all the LFields
-
- } // loop over all transformed dimensions
-
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &f) {
-
- // Now assign into output Field, and compress last temp's storage:
- f[in_dom] = (*tempR)[tempR->getLayout().getDomain()];
- if (this->compressTemps()) *tempR = 0;
-
- }
-
- // Normalize:
- if (direction == +1) f = f * this->getNormFact();
-
- return;
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
+
+ // Common loop iterate and other vars:
+ size_t d;
+ int idim; // idim loops over the number of transform dims.
+ int begdim, enddim;
+ size_t nTransformDims = this->numTransformDims();
+ // check that there is no real-to-complex transform to do
+ PInsist(nTransformDims==numSineTransforms_m,
+ "Cannot perform real-to-complex transform in-place!!");
+
+ // do all the sine transforms
+
+ // Field* management aid
+ RealField_t* tempR = &f;
+ // Local work array passed to FFT:
+ T* localdataR;
+
+ // Loop over the dimensions to be sine transformed:
+ begdim = (direction == +1) ? 0 : static_cast<int>(nTransformDims-1);
+ enddim = (direction == +1) ? static_cast<int>(nTransformDims) : -1;
+ for (idim = begdim; idim != enddim; idim+=direction) {
+
+ // Now do the serial transforms along this dimension:
+
+ bool skipTranspose = false;
+ // if this is the first transform dimension, we might be able
+ // to skip the transpose into the first temporary Field
+ if (idim == begdim) {
+ // get domain for comparison
+ const Domain_t& first_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(first_dom[0])) &&
+ (in_dom[0].length() == first_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ // if this is the last transform dimension, we might be able
+ // to skip the last temporary and transpose right into f
+ if (idim == enddim-direction) {
+ // get the domain for comparison
+ const Domain_t& last_dom = tempRLayouts_m[idim]->getDomain();
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(last_dom[0])) &&
+ (in_dom[0].length() == last_dom[0].length()) &&
+ (in_layout.getDistribution(0) == SERIAL) &&
+ (f.getGC() == FFT<SineTransform,Dim,T>::nullGC) );
+ }
+
+ if (!skipTranspose) {
+ // transpose and permute to Field with transform dim first
+ (*tempRFields_m[idim])[tempRLayouts_m[idim]->getDomain()] =
+ (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = tempRFields_m[idim]; // Field* management aid
+ }
+ else if (idim == enddim-direction && tempR != &f) {
+ // last transform and we can skip the last temporary field
+ // so do the transpose here using f instead
+
+ // transpose and permute to Field with transform dim first
+ f[in_dom] = (*tempR)[tempR->getLayout().getDomain()];
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = &f; // Field* management aid
+ }
+
+
+
+ // Loop over all the Vnodes, working on the LField in each.
+ typename RealField_t::const_iterator_if l_i, l_end = tempR->end_if();
+ for (l_i = tempR->begin_if(); l_i != l_end; ++l_i) {
+
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
+
+ // Do 1D real-to-real FFT's on all the strips in the LField:
+ int nstrips = 1, length = ldf->size(0);
+ for (d=1; d<Dim; ++d) nstrips *= ldf->size(d);
+ for (int istrip=0; istrip<nstrips; ++istrip) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(idim, direction, localdataR);
+ // advance the data pointer
+ localdataR += length;
+ } // loop over 1D strips
+ } // loop over all the LFields
+
+ } // loop over all transformed dimensions
+
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &f) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ f[in_dom] = (*tempR)[tempR->getLayout().getDomain()];
+ if (this->compressTemps()) *tempR = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1) f = f * this->getNormFact();
+
+ return;
}
@@ -4033,30 +3293,30 @@ FFT<SineTransform,Dim,T>::transform(
template <class T>
FFT<SineTransform,1U,T>::FFT(
- const typename FFT<SineTransform,1U,T>::Domain_t& rdomain,
- const bool sineTransformDims[1U], const bool& compressTemps)
- : FFTBase<1U,T>(FFT<SineTransform,1U,T>::sineFFT, rdomain,
- sineTransformDims, compressTemps)
+ const typename FFT<SineTransform,1U,T>::Domain_t& rdomain,
+ const bool sineTransformDims[1U], const bool& compressTemps)
+: FFTBase<1U,T>(FFT<SineTransform,1U,T>::sineFFT, rdomain,
+ sineTransformDims, compressTemps)
{
- // Tau profiling
+ // Tau profiling
- // get axis length
- int length;
- length = rdomain[0].length();
+ // get axis length
+ int length;
+ length = rdomain[0].length();
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::sineFFT; // sine transform
- T& normFact = this->getNormFact();
- normFact = 1.0 / (2.0 * (length + 1));
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::sineFFT; // sine transform
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / (2.0 * (length + 1));
- // set up FFT Engine
- this->getEngine().setup(1U, &transformType, &length);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(1U, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -4066,29 +3326,29 @@ FFT<SineTransform,1U,T>::FFT(
template <class T>
FFT<SineTransform,1U,T>::FFT(
- const typename FFT<SineTransform,1U,T>::Domain_t& rdomain,
- const bool& compressTemps)
- : FFTBase<1U,T>(FFT<SineTransform,1U,T>::sineFFT, rdomain, compressTemps)
+ const typename FFT<SineTransform,1U,T>::Domain_t& rdomain,
+ const bool& compressTemps)
+: FFTBase<1U,T>(FFT<SineTransform,1U,T>::sineFFT, rdomain, compressTemps)
{
- // Tau profiling
+ // Tau profiling
- // get axis length
- int length;
- length = rdomain[0].length();
+ // get axis length
+ int length;
+ length = rdomain[0].length();
- // get transform type for FFT Engine, compute normalization
- int transformType;
- transformType = FFTBase<1U,T>::sineFFT; // sine transform
- T& normFact = this->getNormFact();
- normFact = 1.0 / (2.0 * (length + 1));
+ // get transform type for FFT Engine, compute normalization
+ int transformType;
+ transformType = FFTBase<1U,T>::sineFFT; // sine transform
+ T& normFact = this->getNormFact();
+ normFact = 1.0 / (2.0 * (length + 1));
- // set up FFT Engine
- this->getEngine().setup(1U, &transformType, &length);
- // set up the temporary fields
- setup();
+ // set up FFT Engine
+ this->getEngine().setup(1U, &transformType, &length);
+ // set up the temporary fields
+ setup();
}
//-----------------------------------------------------------------------------
@@ -4100,21 +3360,21 @@ template <class T>
void
FFT<SineTransform,1U,T>::setup(void) {
- // Tau profiling
+ // Tau profiling
- const Domain_t& domain = this->getDomain(); // get real Field domain
+ const Domain_t& domain = this->getDomain(); // get real Field domain
- // generate temporary field layout
- tempRLayouts_m = new Layout_t(domain[0], PARALLEL, 1);
- // create temporary real Field
- tempRFields_m = new RealField_t(*tempRLayouts_m);
- // If user requests no intermediate compression, uncompress right now:
- if (!this->compressTemps()) tempRFields_m->Uncompress();
+ // generate temporary field layout
+ tempRLayouts_m = new Layout_t(domain[0], PARALLEL, 1);
+ // create temporary real Field
+ tempRFields_m = new RealField_t(*tempRLayouts_m);
+ // If user requests no intermediate compression, uncompress right now:
+ if (!this->compressTemps()) tempRFields_m->Uncompress();
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -4124,14 +3384,14 @@ FFT<SineTransform,1U,T>::setup(void) {
template <class T>
FFT<SineTransform,1U,T>::~FFT(void) {
- // Tau profiling
+ // Tau profiling
- // delete temporary field and layout
- delete tempRFields_m;
- delete tempRLayouts_m;
+ // delete temporary field and layout
+ delete tempRFields_m;
+ delete tempRLayouts_m;
}
//-----------------------------------------------------------------------------
@@ -4141,103 +3401,103 @@ FFT<SineTransform,1U,T>::~FFT(void) {
template <class T>
void
FFT<SineTransform,1U,T>::transform(
- int direction,
- FFT<SineTransform,1U,T>::RealField_t& f,
- FFT<SineTransform,1U,T>::RealField_t& g,
- const bool& constInput)
+ int direction,
+ FFT<SineTransform,1U,T>::RealField_t& f,
+ FFT<SineTransform,1U,T>::RealField_t& g,
+ const bool& constInput)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Fields
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- const Layout_t& out_layout = g.getLayout();
- const Domain_t& out_dom = out_layout.getDomain();
- PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
- this->checkDomain(this->getDomain(),out_dom), true);
-
- // Field* management aid
- RealField_t* tempR = &f;
- // Local work array passed to FFT:
- T* localdataR;
-
-
- // Now do the serial transform along this dimension:
-
- // get the domain for comparison
- const Domain_t& temp_dom = tempRLayouts_m->getDomain();
- bool skipTranspose = false;
- // we might be able
- // to skip the first temporary and just use f
- if (!constInput) {
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
+
+ // Check domain of incoming Fields
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ const Layout_t& out_layout = g.getLayout();
+ const Domain_t& out_dom = out_layout.getDomain();
+ PAssert_EQ( this->checkDomain(this->getDomain(),in_dom) &&
+ this->checkDomain(this->getDomain(),out_dom), true);
+
+ // Field* management aid
+ RealField_t* tempR = &f;
+ // Local work array passed to FFT:
+ T* localdataR;
+
+
+ // Now do the serial transform along this dimension:
+
+ // get the domain for comparison
+ const Domain_t& temp_dom = tempRLayouts_m->getDomain();
+ bool skipTranspose = false;
+ // we might be able
+ // to skip the first temporary and just use f
+ if (!constInput) {
+ // check that zeroth axis is the same, has one vnode
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
+ (in_dom[0].length() == temp_dom[0].length()) &&
+ (in_layout.numVnodes() == 1) &&
+ (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
+ }
+
+ bool skipFinal = false;
+ // we might be able
+ // to skip the last temporary and transpose right into g
+
// check that zeroth axis is the same, has one vnode
// and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
- (in_dom[0].length() == temp_dom[0].length()) &&
- (in_layout.numVnodes() == 1) &&
- (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
- }
-
- bool skipFinal = false;
- // we might be able
- // to skip the last temporary and transpose right into g
-
- // check that zeroth axis is the same, has one vnode
- // and that there are no guard cells
- skipFinal = ( (out_dom[0].sameBase(temp_dom[0])) &&
- (out_dom[0].length() == temp_dom[0].length()) &&
- (out_layout.numVnodes() == 1) &&
- (g.getGC() == FFT<SineTransform,1U,T>::nullGC) );
-
- if (!skipTranspose) {
- // assign to Field with proper layout
- (*tempRFields_m) = (*tempR);
- tempR = tempRFields_m; // Field* management aid
- }
- if (skipFinal) {
- // we can skip the last temporary field
- // so do the transpose here using g instead
-
- // assign to Field with proper layout
- g = (*tempR);
+ skipFinal = ( (out_dom[0].sameBase(temp_dom[0])) &&
+ (out_dom[0].length() == temp_dom[0].length()) &&
+ (out_layout.numVnodes() == 1) &&
+ (g.getGC() == FFT<SineTransform,1U,T>::nullGC) );
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = &g; // Field* management aid
- }
+ if (!skipTranspose) {
+ // assign to Field with proper layout
+ (*tempRFields_m) = (*tempR);
+ tempR = tempRFields_m; // Field* management aid
+ }
+ if (skipFinal) {
+ // we can skip the last temporary field
+ // so do the transpose here using g instead
+ // assign to Field with proper layout
+ g = (*tempR);
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = &g; // Field* management aid
+ }
- // There should be just one LField.
- typename RealField_t::const_iterator_if l_i = tempR->begin_if();
- if (l_i != tempR->end_if()) {
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
- // Do the 1D FFT:
- this->getEngine().callFFT(0, direction, localdataR);
- }
+ // There should be just one LField.
+ typename RealField_t::const_iterator_if l_i = tempR->begin_if();
+ if (l_i != tempR->end_if()) {
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &g) {
+ // Do the 1D FFT:
+ this->getEngine().callFFT(0, direction, localdataR);
+ }
- // Now assign into output Field, and compress last temp's storage:
- g = (*tempR);
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- }
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &g) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ g = (*tempR);
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+
+ }
- // Normalize:
- if (direction == +1) g = g * this->getNormFact();
+ // Normalize:
+ if (direction == +1) g = g * this->getNormFact();
- return;
+ return;
}
//-----------------------------------------------------------------------------
@@ -4247,104 +3507,103 @@ FFT<SineTransform,1U,T>::transform(
template <class T>
void
FFT<SineTransform,1U,T>::transform(
- int direction,
- FFT<SineTransform,1U,T>::RealField_t& f)
+ int direction,
+ FFT<SineTransform,1U,T>::RealField_t& f)
{
- // indicate we're doing another FFT
- //INCIPPLSTAT(incFFTs);
-
- // Check domain of incoming Field
- const Layout_t& in_layout = f.getLayout();
- const Domain_t& in_dom = in_layout.getDomain();
- PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
-
- // Field* management aid
- RealField_t* tempR = &f;
- // Local work array passed to FFT:
- T* localdataR;
-
-
- // Now do the serial transform along this dimension:
-
- // get domain for comparison
- const Domain_t& temp_dom = tempRLayouts_m->getDomain();
- bool skipTranspose = false;
- // we might be able
- // to skip the transpose into the first temporary Field
-
- // check that zeroth axis is the same and is serial
- // and that there are no guard cells
- skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
- (in_dom[0].length() == temp_dom[0].length()) &&
- (in_layout.numVnodes() == 1) &&
- (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
-
- bool skipFinal = false;
- // we might be able
- // to skip the last temporary and transpose right into f
-
- // check that zeroth axis is the same, has one vnode
- // and that there are no guard cells
- skipFinal = ( (in_dom[0].sameBase(temp_dom[0])) &&
- (in_dom[0].length() == temp_dom[0].length()) &&
- (in_layout.numVnodes() == 1) &&
- (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
-
- if (!skipTranspose) {
- // assign to Field with proper layout
- (*tempRFields_m) = (*tempR);
-
- tempR = tempRFields_m; // Field* management aid
- }
- if (skipFinal) {
- // we can skip the last temporary field
- // so do the transpose here using f instead
-
- // assign to Field with proper layout
- f = (*tempR);
+ // indicate we're doing another FFT
+ //INCIPPLSTAT(incFFTs);
- // Compress out previous iterate's storage:
- if (this->compressTemps() && tempR != &f) *tempR = 0;
- tempR = &f; // Field* management aid
- }
+ // Check domain of incoming Field
+ const Layout_t& in_layout = f.getLayout();
+ const Domain_t& in_dom = in_layout.getDomain();
+ PAssert_EQ(this->checkDomain(this->getDomain(),in_dom), true);
+ // Field* management aid
+ RealField_t* tempR = &f;
+ // Local work array passed to FFT:
+ T* localdataR;
- // There should be just one LField.
- typename RealField_t::const_iterator_if l_i = tempR->begin_if();
- if (l_i != tempR->end_if()) {
+ // Now do the serial transform along this dimension:
- // Get the LField
- RealLField_t* ldf = (*l_i).second.get();
- // make sure we are uncompressed
- ldf->Uncompress();
- // get the raw data pointer
- localdataR = ldf->getP();
+ // get domain for comparison
+ const Domain_t& temp_dom = tempRLayouts_m->getDomain();
+ bool skipTranspose = false;
+ // we might be able
+ // to skip the transpose into the first temporary Field
- // Do the 1D FFT:
- this->getEngine().callFFT(0, direction, localdataR);
- }
+ // check that zeroth axis is the same and is serial
+ // and that there are no guard cells
+ skipTranspose = ( (in_dom[0].sameBase(temp_dom[0])) &&
+ (in_dom[0].length() == temp_dom[0].length()) &&
+ (in_layout.numVnodes() == 1) &&
+ (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
+ bool skipFinal = false;
+ // we might be able
+ // to skip the last temporary and transpose right into f
- // skip final assignment and compress if we used g as final temporary
- if (tempR != &f) {
+ // check that zeroth axis is the same, has one vnode
+ // and that there are no guard cells
+ skipFinal = ( (in_dom[0].sameBase(temp_dom[0])) &&
+ (in_dom[0].length() == temp_dom[0].length()) &&
+ (in_layout.numVnodes() == 1) &&
+ (f.getGC() == FFT<SineTransform,1U,T>::nullGC) );
- // Now assign into output Field, and compress last temp's storage:
- f = (*tempR);
- if (this->compressTemps()) *tempR = 0;
+ if (!skipTranspose) {
+ // assign to Field with proper layout
+ (*tempRFields_m) = (*tempR);
- }
+ tempR = tempRFields_m; // Field* management aid
+ }
+ if (skipFinal) {
+ // we can skip the last temporary field
+ // so do the transpose here using f instead
- // Normalize:
- if (direction == +1) f = f * this->getNormFact();
+ // assign to Field with proper layout
+ f = (*tempR);
+
+ // Compress out previous iterate's storage:
+ if (this->compressTemps() && tempR != &f) *tempR = 0;
+ tempR = &f; // Field* management aid
+ }
- return;
-}
+ // There should be just one LField.
+ typename RealField_t::const_iterator_if l_i = tempR->begin_if();
+ if (l_i != tempR->end_if()) {
+
+ // Get the LField
+ RealLField_t* ldf = (*l_i).second.get();
+ // make sure we are uncompressed
+ ldf->Uncompress();
+ // get the raw data pointer
+ localdataR = ldf->getP();
+
+ // Do the 1D FFT:
+ this->getEngine().callFFT(0, direction, localdataR);
+ }
+
+
+ // skip final assignment and compress if we used g as final temporary
+ if (tempR != &f) {
+
+ // Now assign into output Field, and compress last temp's storage:
+ f = (*tempR);
+ if (this->compressTemps()) *tempR = 0;
+
+ }
+
+ // Normalize:
+ if (direction == +1) f = f * this->getNormFact();
+
+ return;
+}
-/***************************************************************************
- * $RCSfile: FFT.cpp,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
- * IPPL_VERSION_ID: $Id: FFT.cpp,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
- ***************************************************************************/
\ No newline at end of file
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/FFTBase.h b/ippl/src/FFT/FFTBase.h
index 6d51ce8d4c0f698c7431ccc3752db005f1d1709b..eddc59ebb48a39bacdbb09584cd7f532ec6016af 100644
--- a/ippl/src/FFT/FFTBase.h
+++ b/ippl/src/FFT/FFTBase.h
@@ -1,12 +1,12 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
//--------------------------------------------------------------------------
// Class FFTBase
@@ -20,11 +20,7 @@
#include "Index/NDIndex.h"
#include "Field/GuardCellSizes.h"
-#if defined(IPPL_USE_SCSL_FFT)
-#include "FFT/SCSL_FFT.h"
-#else
#include "FFT/fftpack_FFT.h"
-#endif
#include <map>
#include <iostream>
@@ -67,11 +63,7 @@ public:
enum FFT_e { ccFFT, rcFFT, sineFFT, cosineFFT };
// Type used for performing 1D FFTs
-#if defined(IPPL_USE_SCSL_FFT)
- typedef SCSL<T> InternalFFT_t;
-#else
typedef FFTPACK<T> InternalFFT_t;
-#endif
FFTBase() {}
@@ -308,9 +300,9 @@ FFTBase<Dim,T>::checkDomain(const FFTBase<Dim,T>::Domain_t& dom1,
#endif // IPPL_FFT_FFTBASE_H
-/***************************************************************************
- * $RCSfile: FFTBase.h,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
- * IPPL_VERSION_ID: $Id: FFTBase.h,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
- ***************************************************************************/
-
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/FFTBase.hpp b/ippl/src/FFT/FFTBase.hpp
index b5e64a0fa86c7361f5e14ff5a19b949d3cb52ef8..120a8c3e048f5042175361cb33ca1b3ce6ff289c 100644
--- a/ippl/src/FFT/FFTBase.hpp
+++ b/ippl/src/FFT/FFTBase.hpp
@@ -1,36 +1,19 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- * This program was prepared by PSI.
- * All rights in the program are reserved by PSI.
- * Neither PSI nor the author(s)
- * makes any warranty, express or implied, or assumes any liability or
- * responsibility for the use of this software
- *
- * Visit www.amas.web.psi for more details
- *
- ***************************************************************************/
-
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
-
-// include files
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
#include "FFT/FFTBase.h"
#include "Utility/PAssert.h"
/***************************************************************************
FFTBase.cpp: constructors and write() member function for FFTBase class
- ***************************************************************************/
-
+***************************************************************************/
// instantiate static null GC object
template <unsigned Dim, class T>
@@ -44,31 +27,29 @@ template <unsigned Dim, class T>
FFTBase<Dim,T>::FFTBase(FFTBase<Dim,T>::FFT_e transform,
const FFTBase<Dim,T>::Domain_t& domain,
const bool transformTheseDims[Dim], bool compressTemps)
- : transformType_m(transform), // transform type
+: transformType_m(transform), // transform type
Domain_m(domain), // field domain
compressTempFields_m(compressTemps) // compress temp fields?
{
- // Tau profiling
+ // Tau profiling
-
-
- // Store which dims are transformed, and count up how many there are
- nTransformDims_m = 0;
- unsigned d;
- for (d=0; d<Dim; ++d) {
- transformDims_m[d] = transformTheseDims[d];
- if (transformTheseDims[d]) ++nTransformDims_m;
- }
-
- // check that at least one dimension is being transformed
- PInsist(nTransformDims_m>0,"Must transform at least one axis!!");
-
- // Store the "names" (0,1,2) of these dims in an array of computed size
- activeDims_m = new unsigned[nTransformDims_m];
- int icount = 0;
- for (d=0; d<Dim; ++d)
- if (transformDims_m[d]) activeDims_m[icount++] = d;
+ // Store which dims are transformed, and count up how many there are
+ nTransformDims_m = 0;
+ unsigned d;
+ for (d=0; d<Dim; ++d) {
+ transformDims_m[d] = transformTheseDims[d];
+ if (transformTheseDims[d]) ++nTransformDims_m;
+ }
+
+ // check that at least one dimension is being transformed
+ PInsist(nTransformDims_m>0,"Must transform at least one axis!!");
+
+ // Store the "names" (0,1,2) of these dims in an array of computed size
+ activeDims_m = new unsigned[nTransformDims_m];
+ int icount = 0;
+ for (d=0; d<Dim; ++d)
+ if (transformDims_m[d]) activeDims_m[icount++] = d;
}
@@ -76,22 +57,20 @@ template <unsigned Dim, class T>
FFTBase<Dim,T>::FFTBase(FFTBase<Dim,T>::FFT_e transform,
const FFTBase<Dim,T>::Domain_t& domain,
bool compressTemps)
- : transformType_m(transform), // transform type
+: transformType_m(transform), // transform type
Domain_m(domain), // field domain
compressTempFields_m(compressTemps) // compress temp fields?
{
- // Tau profiling
-
-
+ // Tau profiling
- // Default, transform all dims:
- nTransformDims_m = Dim;
- activeDims_m = new unsigned[Dim];
- for (unsigned d=0; d<Dim; d++) {
- transformDims_m[d] = true;
- activeDims_m[d] = d;
- }
+ // Default, transform all dims:
+ nTransformDims_m = Dim;
+ activeDims_m = new unsigned[Dim];
+ for (unsigned d=0; d<Dim; d++) {
+ transformDims_m[d] = true;
+ activeDims_m[d] = d;
+ }
}
@@ -102,38 +81,36 @@ FFTBase<Dim,T>::FFTBase(FFTBase<Dim,T>::FFT_e transform,
template <unsigned Dim, class T>
void FFTBase<Dim,T>::write(std::ostream& out) const {
- // Tau profiling
-
+ // Tau profiling
-
- // Dump contents of FFT object
- out << "---------------FFT Object Dump Begin-------------------" << std::endl;
- // Output the user-defined names for transform directions:
- out << "Map of transform direction names:" << std::endl;
- std::map<char*,unsigned>::const_iterator mi, m_end = directions_m.end();
- for (mi = directions_m.begin(); mi != m_end; ++mi)
- out << "[" << (*mi).first << "," << (*mi).second << "]" << std::endl;
- // Output type of transform
- out << "Transform type = " << getTransformType(transformType_m) << std::endl;
- // Output which dims are transformed:
- out << "Transform dimensions: ";
- for (unsigned d=0; d<Dim; ++d) {
- out << "dim " << d << " = ";
- if (transformDims_m[d])
- out << "true" << std::endl;
- else
- out << "false" << std::endl;
- }
- out << std::endl;
- out << "Input Field domain = " << Domain_m << std::endl; // Output the domain.
- out << "---------------FFT Object Dump End---------------------" << std::endl;
-
- return;
+ // Dump contents of FFT object
+ out << "---------------FFT Object Dump Begin-------------------" << std::endl;
+ // Output the user-defined names for transform directions:
+ out << "Map of transform direction names:" << std::endl;
+ std::map<char*,unsigned>::const_iterator mi, m_end = directions_m.end();
+ for (mi = directions_m.begin(); mi != m_end; ++mi)
+ out << "[" << (*mi).first << "," << (*mi).second << "]" << std::endl;
+ // Output type of transform
+ out << "Transform type = " << getTransformType(transformType_m) << std::endl;
+ // Output which dims are transformed:
+ out << "Transform dimensions: ";
+ for (unsigned d=0; d<Dim; ++d) {
+ out << "dim " << d << " = ";
+ if (transformDims_m[d])
+ out << "true" << std::endl;
+ else
+ out << "false" << std::endl;
+ }
+ out << std::endl;
+ out << "Input Field domain = " << Domain_m << std::endl; // Output the domain.
+ out << "---------------FFT Object Dump End---------------------" << std::endl;
+
+ return;
}
-
-/***************************************************************************
- * $RCSfile: FFTBase.cpp,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
- * IPPL_VERSION_ID: $Id: FFTBase.cpp,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
- ***************************************************************************/
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/SCSL_FFT.F b/ippl/src/FFT/SCSL_FFT.F
deleted file mode 100644
index 357436468e267c9f43c4e237e893051314a8d49b..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/SCSL_FFT.F
+++ /dev/null
@@ -1,93 +0,0 @@
-c========================================================================
-c
-c The IPPL Framework
-c
-c This program was prepared by the Regents of the University of
-c Visit http://people.web.psi.ch/adelmann/ for more details
-c
-c========================================================================
-
-c========================================================================
-c=========================SCSL_FFT.F=====================================
-c========================================================================
-c I'm adding some Fortran wrappers around SGI/Cray Scientific Library
-c routines, so we can call them from C++ code
-c========================================================================
-
-
-c The Cray T3E does not have double-precision versions of these
-c routines, since its routines are double-precision by default.
-
-#ifndef IPPL_T3E
-
- subroutine sgizzfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit double precision (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call zzfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
- subroutine sgidzfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit double precision (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call dzfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
- subroutine sgizdfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit double precision (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call zdfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
-#endif
-
-
-c Single-precision routines
-
- subroutine sgiccfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call ccfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
- subroutine sgiscfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call scfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
- subroutine sgicsfft(isign,n,scale,x,y,table,work,isys)
-c------------------------------------------------------------------------
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c------------------------------------------------------------------------
- dimension x(*), y(*), table(*), work(*)
- call csfft(isign,n,scale,x,y,table,work,isys)
- return
- end
-
-
-c========================================================================
-c $RCSfile: SCSL_FFT.F,v $ $Author: adelmann $
-c $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
-c IPPL_VERSION_ID: $Id: SCSL_FFT.F,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
-c========================================================================
diff --git a/ippl/src/FFT/SCSL_FFT.h b/ippl/src/FFT/SCSL_FFT.h
deleted file mode 100644
index 690447b39c36c1bc793b7193d688a26251a829ff..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/SCSL_FFT.h
+++ /dev/null
@@ -1,359 +0,0 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
-
-#ifndef IPPL_FFT_SCSL_FFT_H
-#define IPPL_FFT_SCSL_FFT_H
-
-// include files
-#include "Utility/PAssert.h"
-#include "Utility/IpplInfo.h"
-
-
-/**************************************************************************
- * SCSL_FFT.h: Prototypes for accessing Fortran 1D FFT routines from the
- * SGI/Cray Scientific Library, and definitions for templated class SCSL,
- * which acts as an FFT engine for the FFT class, providing storage for
- * trigonometric information and performing the 1D FFTs as needed.
- **************************************************************************/
-
-
-// For platforms that do Fortran symbols in all caps.
-#if defined(IPPL_T3E)
-
-#define sgiccfft_ SGICCFFT
-#define sgiscfft_ SGISCFFT
-#define sgicsfft_ SGICSFFT
-
-#endif
-
-// SCSL Fortran wrapper function prototypes
-// These functions are defined in file SCSL_FFT.F, and they just turn
-// around and call the native library routines.
-
-#if defined(IPPL_T3E)
-
-// Cray T3E has only "real" routines, which are automatically double precision
-extern "C" {
- // double-precision CC FFT
- void sgiccfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
- // double-precision RC FFT
- void sgiscfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
- // double-precision CR FFT
- void sgicsfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
-}
-
-#else
-
-// SGI offers separate single- and double-precision routines
-extern "C" {
- // double-precision CC FFT
- void sgizzfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
- // double-precision RC FFT
- void sgidzfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
- // double-precision CR FFT
- void sgizdfft_(int& isign, int& n, double& scale, double& x, double& y,
- double& table, double& work, int& isys);
- // single-precision CC FFT
- void sgiccfft_(int& isign, int& n, float& scale, float& x, float& y,
- float& table, float& work, int& isys);
- // single-precision RC FFT
- void sgiscfft_(int& isign, int& n, float& scale, float& x, float& y,
- float& table, float& work, int& isys);
- // single-precision CR FFT
- void sgicsfft_(int& isign, int& n, float& scale, float& x, float& y,
- float& table, float& work, int& isys);
-}
-
-#endif // defined(IPPL_T3E)
-
-
-// SCSL_wrap provides static functions that wrap the Fortran functions
-// in a common interface. We specialize this class on precision type.
-template <class T>
-class SCSL_wrap {};
-
-// Specialization for float
-template <>
-class SCSL_wrap<float> {
-
-public:
- // interface functions used by class SCSL
-
- // complex-to-complex FFT
- static void ccfft(int isign, int n, float scale, float* x, float* y,
- float* table, float* work, int isys) {
- sgiccfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // real-to-complex FFT
- static void rcfft(int isign, int n, float scale, float* x, float* y,
- float* table, float* work, int isys) {
- sgiscfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // complex-to-real FFT
- static void crfft(int isign, int n, float scale, float* x, float* y,
- float* table, float* work, int isys) {
- sgicsfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
-
-};
-
-
-// Specialization for double
-
-#if defined(IPPL_T3E)
-
-// Cray T3E uses single-precision function names for double-precision FFTs
-template <>
-class SCSL_wrap<double> {
-
-public:
- // interface functions used by class SCSL
-
- // complex-to-complex FFT
- static void ccfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgiccfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // real-to-complex FFT
- static void rcfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgiscfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // complex-to-real FFT
- static void crfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgicsfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
-
-};
-
-#else
-
-// SGI has special FFT routines for double-precision
-template <>
-class SCSL_wrap<double> {
-
-public:
- // interface functions used by class SCSL
-
- // complex-to-complex FFT
- static void ccfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgizzfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // real-to-complex FFT
- static void rcfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgidzfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
- // complex-to-real FFT
- static void crfft(int isign, int n, double scale, double* x, double* y,
- double* table, double* work, int isys) {
- sgizdfft_(isign, n, scale, *x, *y, *table, *work, isys);
- }
-
-};
-
-#endif // defined(IPPL_T3E)
-
-
-// Definition of FFT engine class SCSL
-template <class T>
-class SCSL {
-
-public:
-
- // definition of complex type
- typedef complex<T> Complex_t;
-
- // Trivial constructor. Do the real work in setup function.
- SCSL(void) {}
-
- // destructor
- ~SCSL(void);
-
- // setup internal storage and prepare to do FFTs
- void setup(unsigned numTransformDims, const int* transformTypes,
- const int* axisLengths);
-
- // invoke FFT on complex data for given dimension and direction
- void callFFT(unsigned transformDim, int direction, Complex_t* data);
-
- // invoke FFT on real data for given dimension and direction
- void callFFT(unsigned transformDim, int direction, T* data);
-
-private:
-
- unsigned numTransformDims_m; // number of dimensions to transform
- int* transformType_m; // type of transform for each dimension
- int* axisLength_m; // length along each dimension
- T** table_m; // tables of trigonometric factors
- T** work_m; // workspace arrays
-
-};
-
-
-// Inline member function definitions
-
-// destructor
-template <class T>
-inline
-SCSL<T>::~SCSL(void) {
- // delete storage
- for (unsigned d=0; d<numTransformDims_m; ++d) {
- delete [] table_m[d];
- delete [] work_m[d];
- }
- delete [] table_m;
- delete [] work_m;
- delete [] transformType_m;
- delete [] axisLength_m;
-}
-
-// setup internal storage and prepare for doing FFTs
-template <class T>
-inline void
-SCSL<T>::setup(unsigned numTransformDims, const int* transformTypes,
- const int* axisLengths) {
-
- // store transform types and lengths for each transform dim
- numTransformDims_m = numTransformDims;
- transformType_m = new int[numTransformDims_m];
- axisLength_m = new int[numTransformDims_m];
- unsigned d;
- for (d=0; d<numTransformDims_m; ++d) {
- transformType_m[d] = transformTypes[d];
- axisLength_m[d] = axisLengths[d];
- }
-
- // allocate trig table and workspace arrays and initialize table
- // table and work array sizes vary for the Cray T3E and the SGI Origin 2000
- table_m = new T*[numTransformDims_m];
- work_m = new T*[numTransformDims_m];
- T dummy = T();
- for (d=0; d<numTransformDims_m; ++d) {
- switch (transformType_m[d]) {
- case 0: // CC FFT
-#if defined(IPPL_T3E)
- table_m[d] = new T[2 * axisLength_m[d]];
- work_m[d] = new T[4 * axisLength_m[d]];
-#else
- table_m[d] = new T[2 * axisLength_m[d] + 30];
- work_m[d] = new T[2 * axisLength_m[d]];
-#endif
- // this call initializes table
- SCSL_wrap<T>::ccfft(0, axisLength_m[d], dummy, &dummy, &dummy,
- table_m[d], &dummy, 0);
- break;
- case 1: // RC FFT
-#if defined(IPPL_T3E)
- table_m[d] = new T[2 * axisLength_m[d]];
- work_m[d] = new T[2 * (axisLength_m[d] + 2)]; // extend length by two
-#else
- table_m[d] = new T[axisLength_m[d] + 15];
- work_m[d] = new T[axisLength_m[d] + 2]; // extend length by two reals
-#endif
- // this call initializes table
- SCSL_wrap<T>::rcfft(0, axisLength_m[d], dummy, &dummy, &dummy,
- table_m[d], &dummy, 0);
- break;
- case 2: // Sine transform
- ERRORMSG("No sine transforms available in SCSL!!" << endl);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
- }
- }
-
- return;
-}
-
-// invoke FFT on complex data for given dimension and direction
-template <class T>
-inline void
-SCSL<T>::callFFT(unsigned transformDim, int direction,
- SCSL<T>::Complex_t* data) {
-
- // check transform dimension and direction arguments
- PAssert_LT(transformDim, numTransformDims_m);
- PAssert_EQ(std::abs(direction), 1);
-
- // cast complex number pointer to T* for calling Fortran routines
- T* rdata = reinterpret_cast<T*>(data);
-
- // branch on type of transform for this dimension
- switch (transformType_m[transformDim]) {
- case 0: // CC FFT
- SCSL_wrap<T>::ccfft(direction, axisLength_m[transformDim], 1.0, rdata,
- rdata, table_m[transformDim], work_m[transformDim], 0);
- break;
- case 1: // RC FFT
- if (direction == +1) { // real-to-complex
- SCSL_wrap<T>::rcfft(+1, axisLength_m[transformDim], 1.0, rdata, rdata,
- table_m[transformDim], work_m[transformDim], 0);
- }
- else { // complex-to-real
- SCSL_wrap<T>::crfft(-1, axisLength_m[transformDim], 1.0, rdata, rdata,
- table_m[transformDim], work_m[transformDim], 0);
- }
- break;
- case 2: // Sine transform
- ERRORMSG("No sine transforms available in SCSL!!" << endl);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
- }
-
- return;
-}
-
-
-// invoke FFT on real data for given dimension and direction
-template <class T>
-inline void
-SCSL<T>::callFFT(unsigned transformDim, int direction, T* data) {
-
- // check transform dimension argument
- PAssert_LT(transformDim, numTransformDims_m);
- // branch on transform type for this dimension
- switch (transformType_m[transformDim]) {
- case 0: // CC FFT
- ERRORMSG("complex-to-complex FFT uses complex data!!" << endl);
- break;
- case 1: // RC FFT
- ERRORMSG("real-to-complex FFT uses complex data!!" << endl);
- break;
- case 2: // Sine transform
- ERRORMSG("No sine transforms available in SCSL!!" << endl);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
- }
-
- return;
-}
-
-
-#endif // IPPL_FFT_SCSL_FFT_H
-
-/***************************************************************************
- * $RCSfile: SCSL_FFT.h,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
- * IPPL_VERSION_ID: $Id: SCSL_FFT.h,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
- ***************************************************************************/
-
diff --git a/ippl/src/FFT/c_utils.c b/ippl/src/FFT/c_utils.c
deleted file mode 100644
index 30255557d4e433d7bb77e7317e1d4138882cd0aa..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/c_utils.c
+++ /dev/null
@@ -1,65 +0,0 @@
-/*
-Copyright (c) 2008-2011, Max-Planck-Society
-All rights reserved.
-
-Redistribution and use in source and binary forms, with or without
-modification, are permitted provided that the following conditions are met:
- * Redistributions of source code must retain the above copyright
- notice, this list of conditions and the following disclaimer.
- * Redistributions in binary form must reproduce the above copyright
- notice, this list of conditions and the following disclaimer in the
- documentation and/or other materials provided with the distribution.
- * Neither the name of the Max-Planck-Society nor the
- names of its contributors may be used to endorse or promote products
- derived from this software without specific prior written permission.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
-ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
-WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
-DISCLAIMED. IN NO EVENT SHALL THE MAX-PLANCK-SOCIETY BE LIABLE FOR ANY
-DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
-(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
-LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
-ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
-SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-*/
-
-/*
- * libc_utils is being developed at the Max-Planck-Institut fuer Astrophysik
- * and financially supported by the Deutsches Zentrum fuer Luft- und Raumfahrt
- * (DLR).
- */
-
-/*
- * Convenience functions
- *
- * Author: Martin Reinecke
- */
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include "c_utils.h"
-
-void util_fail_ (const char *file, int line, const char *func, const char *msg)
- {
- fprintf(stderr,"%s, %i (%s):\n%s\n",file,line,func,msg);
- exit(1);
- }
-void util_warn_ (const char *file, int line, const char *func, const char *msg)
- {
- fprintf(stderr,"%s, %i (%s):\n%s\n",file,line,func,msg);
- exit(1);
- }
-
-void *util_malloc_ (size_t sz)
- {
- void *res;
- if (sz==0) return NULL;
- res = malloc(sz);
- UTIL_ASSERT(res,"malloc() failed");
- return res;
- }
-void util_free_ (void *ptr)
- { if ((ptr)!=NULL) free(ptr); }
diff --git a/ippl/src/FFT/c_utils.h b/ippl/src/FFT/c_utils.h
deleted file mode 100644
index a42915e4754022efda87a9716d1fe81f2f6b2374..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/c_utils.h
+++ /dev/null
@@ -1,137 +0,0 @@
-/*
-Copyright (c) 2008-2011, Max-Planck-Society
-All rights reserved.
-
-Redistribution and use in source and binary forms, with or without
-modification, are permitted provided that the following conditions are met:
- * Redistributions of source code must retain the above copyright
- notice, this list of conditions and the following disclaimer.
- * Redistributions in binary form must reproduce the above copyright
- notice, this list of conditions and the following disclaimer in the
- documentation and/or other materials provided with the distribution.
- * Neither the name of the Max-Planck-Society nor the
- names of its contributors may be used to endorse or promote products
- derived from this software without specific prior written permission.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
-ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
-WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
-DISCLAIMED. IN NO EVENT SHALL THE MAX-PLANCK-SOCIETY BE LIABLE FOR ANY
-DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
-(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
-LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
-ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
-SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-*/
-
-/*
- * libc_utils is being developed at the Max-Planck-Institut fuer Astrophysik
- * and financially supported by the Deutsches Zentrum fuer Luft- und Raumfahrt
- * (DLR).
- */
-
-/*! \file c_utils.h
- * Convenience functions
- *
- * \author Martin Reinecke
- * \note This file should only be included from .c files, NOT from .h files.
- */
-
-#ifndef PLANCK_C_UTILS_H
-#define PLANCK_C_UTILS_H
-
-#include <math.h>
-#include <stdlib.h>
-#include <stddef.h>
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-void util_fail_ (const char *file, int line, const char *func, const char *msg);
-void util_warn_ (const char *file, int line, const char *func, const char *msg);
-void *util_malloc_ (size_t sz);
-void util_free_ (void *ptr);
-
-void announce_c (const char *name);
-void module_startup_c (const char *name, int argc, int argc_expected,
- const char *argv_expected, int verbose);
-
-#if defined (__GNUC__)
-#define UTIL_FUNC_NAME__ __func__
-#else
-#define UTIL_FUNC_NAME__ "unknown"
-#endif
-
-#define UTIL_ASSERT(cond,msg) \
- if(!(cond)) util_fail_(__FILE__,__LINE__,UTIL_FUNC_NAME__,msg)
-#define UTIL_WARN(cond,msg) \
- if(!(cond)) util_warn_(__FILE__,__LINE__,UTIL_FUNC_NAME__,msg)
-#define UTIL_FAIL(msg) \
- util_fail_(__FILE__,__LINE__,UTIL_FUNC_NAME__,msg)
-
-#define ALLOC(ptr,type,num) \
- do { (ptr)=(type *)util_malloc_((num)*sizeof(type)); } while (0)
-#define RALLOC(type,num) \
- ((type *)util_malloc_((num)*sizeof(type)))
-#define DEALLOC(ptr) \
- do { util_free_(ptr); (ptr)=NULL; } while(0)
-#define RESIZE(ptr,type,num) \
- do { util_free_(ptr); ALLOC(ptr,type,num); } while(0)
-#define REALLOC(ptr,type,num) \
- do { \
- ptr = (type *)realloc(ptr,(num)*sizeof(type)); \
- UTIL_ASSERT(ptr,"realloc() failed"); \
- } while(0)
-#define GROW(ptr,type,sz_old,sz_new) \
- do { \
- if ((sz_new)>(sz_old)) \
- { RESIZE(ptr,type,2*(sz_new));sz_old=2*(sz_new); } \
- } while(0)
-#define SET_ARRAY(ptr,i1,i2,val) \
- do { \
- ptrdiff_t cnt_; \
- for (cnt_=(i1);cnt_<(i2);++cnt_) (ptr)[cnt_]=(val); \
- } while(0)
-#define COPY_ARRAY(src,dest,i1,i2) \
- do { \
- ptrdiff_t cnt_; \
- for (cnt_=(i1);cnt_<(i2);++cnt_) (dest)[cnt_]=(src)[cnt_]; \
- } while(0)
-
-#define ALLOC2D(ptr,type,num1,num2) \
- do { \
- size_t cnt_, num1_=(num1), num2_=(num2); \
- ALLOC(ptr,type *,num1_); \
- ALLOC(ptr[0],type,num1_*num2_); \
- for (cnt_=1; cnt_<num1_; ++cnt_) \
- ptr[cnt_]=ptr[cnt_-1]+num2_; \
- } while(0)
-#define DEALLOC2D(ptr) \
- do { if(ptr) DEALLOC((ptr)[0]); DEALLOC(ptr); } while(0)
-
-#define FAPPROX(a,b,eps) \
- (fabs((a)-(b))<((eps)*fabs(b)))
-#define ABSAPPROX(a,b,eps) \
- (fabs((a)-(b))<(eps))
-#define IMAX(a,b) \
- (((a)>(b)) ? (a) : (b))
-#define IMIN(a,b) \
- (((a)<(b)) ? (a) : (b))
-
-#define SWAP(a,b,type) \
- do { type tmp_=(a); (a)=(b); (b)=tmp_; } while(0)
-
-#define CHECK_STACK_ALIGN(align) \
- do { \
- double foo; \
- UTIL_WARN((((size_t)(&foo))&(align-1))==0, \
- "WARNING: stack not sufficiently aligned!"); \
- } while(0)
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif
diff --git a/ippl/src/FFT/f2c.h b/ippl/src/FFT/f2c.h
deleted file mode 100644
index 0eb44f12da07b73ad8041327c1529e6bddab7b6c..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/f2c.h
+++ /dev/null
@@ -1,245 +0,0 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- * This program was prepared by PSI.
- * All rights in the program are reserved by PSI.
- * Neither PSI nor the author(s)
- * makes any warranty, express or implied, or assumes any liability or
- * responsibility for the use of this software
- *
- * Visit www.amas.web.psi for more details
- *
- ***************************************************************************/
-
-/* f2c.h -- Standard Fortran to C header file */
-
-/** barf [ba:rf] 2. "He suggested using FORTRAN, and everybody barfed."
-
- - From The Shogakukan DICTIONARY OF NEW ENGLISH (Second edition) */
-
-#ifndef F2C_INCLUDE
-#define F2C_INCLUDE
-
-typedef long int integer;
-typedef unsigned long uinteger;
-typedef char *address;
-typedef short int shortint;
-typedef float real;
-typedef double doublereal;
-typedef struct { real r, i; } complex;
-typedef struct { doublereal r, i; } doublecomplex;
-typedef long int logical;
-typedef short int shortlogical;
-typedef char logical1;
-typedef char integer1;
-#if 0 /* Adjust for integer*8. */
-typedef long long longint; /* system-dependent */
-typedef unsigned long long ulongint; /* system-dependent */
-#define qbit_clear(a,b) ((a) & ~((ulongint)1 << (b)))
-#define qbit_set(a,b) ((a) | ((ulongint)1 << (b)))
-#endif
-
-#define TRUE_ (1)
-#define FALSE_ (0)
-
-/* Extern is for use with -E */
-#ifndef Extern
-#define Extern extern
-#endif
-
-/* I/O stuff */
-
-#ifdef f2c_i2
-/* for -i2 */
-typedef short flag;
-typedef short ftnlen;
-typedef short ftnint;
-#else
-typedef long int flag;
-typedef long int ftnlen;
-typedef long int ftnint;
-#endif
-
-/*external read, write*/
-typedef struct
-{ flag cierr;
- ftnint ciunit;
- flag ciend;
- char *cifmt;
- ftnint cirec;
-} cilist;
-
-/*internal read, write*/
-typedef struct
-{ flag icierr;
- char *iciunit;
- flag iciend;
- char *icifmt;
- ftnint icirlen;
- ftnint icirnum;
-} icilist;
-
-/*open*/
-typedef struct
-{ flag oerr;
- ftnint ounit;
- char *ofnm;
- ftnlen ofnmlen;
- char *osta;
- char *oacc;
- char *ofm;
- ftnint orl;
- char *oblnk;
-} olist;
-
-/*close*/
-typedef struct
-{ flag cerr;
- ftnint cunit;
- char *csta;
-} cllist;
-
-/*rewind, backspace, endfile*/
-typedef struct
-{ flag aerr;
- ftnint aunit;
-} alist;
-
-/* inquire */
-typedef struct
-{ flag inerr;
- ftnint inunit;
- char *infile;
- ftnlen infilen;
- ftnint *inex; /*parameters in standard's order*/
- ftnint *inopen;
- ftnint *innum;
- ftnint *innamed;
- char *inname;
- ftnlen innamlen;
- char *inacc;
- ftnlen inacclen;
- char *inseq;
- ftnlen inseqlen;
- char *indir;
- ftnlen indirlen;
- char *infmt;
- ftnlen infmtlen;
- char *inform;
- ftnint informlen;
- char *inunf;
- ftnlen inunflen;
- ftnint *inrecl;
- ftnint *innrec;
- char *inblank;
- ftnlen inblanklen;
-} inlist;
-
-#define VOID void
-
-union Multitype { /* for multipple entry points */
- integer1 g;
- shortint h;
- integer i;
- /* longint j; */
- real r;
- doublereal d;
- complex c;
- doublecomplex z;
- };
-
-typedef union Multitype Multitype;
-
-/*typedef long int Long;*/ /* No longer used; formerly in Namelist */
-
-struct Vardesc { /* for Namelist */
- char *name;
- char *addr;
- ftnlen *dims;
- int type;
- };
-typedef struct Vardesc Vardesc;
-
-struct Namelist {
- char *name;
- Vardesc **vars;
- int nvars;
- };
-typedef struct Namelist Namelist;
-
-#define abs(x) ((x) >= 0 ? (x) : -(x))
-#define dabs(x) (doublereal)abs(x)
-#define min(a,b) ((a) <= (b) ? (a) : (b))
-#define max(a,b) ((a) >= (b) ? (a) : (b))
-#define dmin(a,b) (doublereal)min(a,b)
-#define dmax(a,b) (doublereal)max(a,b)
-#define bit_test(a,b) ((a) >> (b) & 1)
-#define bit_clear(a,b) ((a) & ~((uinteger)1 << (b)))
-#define bit_set(a,b) ((a) | ((uinteger)1 << (b)))
-
-/* procedure parameter types for -A and -C++ */
-
-#define F2C_proc_par_types 1
-#ifdef __cplusplus
-typedef int /* Unknown procedure type */ (*U_fp)(...);
-typedef shortint (*J_fp)(...);
-typedef integer (*I_fp)(...);
-typedef real (*R_fp)(...);
-typedef doublereal (*D_fp)(...), (*E_fp)(...);
-typedef /* Complex */ VOID (*C_fp)(...);
-typedef /* Double Complex */ VOID (*Z_fp)(...);
-typedef logical (*L_fp)(...);
-typedef shortlogical (*K_fp)(...);
-typedef /* Character */ VOID (*H_fp)(...);
-typedef /* Subroutine */ int (*S_fp)(...);
-#else
-typedef int /* Unknown procedure type */ (*U_fp)();
-typedef shortint (*J_fp)();
-typedef integer (*I_fp)();
-typedef real (*R_fp)();
-typedef doublereal (*D_fp)(), (*E_fp)();
-typedef /* Complex */ VOID (*C_fp)();
-typedef /* Double Complex */ VOID (*Z_fp)();
-typedef logical (*L_fp)();
-typedef shortlogical (*K_fp)();
-typedef /* Character */ VOID (*H_fp)();
-typedef /* Subroutine */ int (*S_fp)();
-#endif
-/* E_fp is for real functions when -R is not specified */
-typedef VOID C_f; /* complex function */
-typedef VOID H_f; /* character function */
-typedef VOID Z_f; /* double complex function */
-typedef doublereal E_f; /* real function with -R not specified */
-
-/* undef any lower-case symbols that your C compiler predefines, e.g.: */
-
-#ifndef Skip_f2c_Undefs
-#undef cray
-#undef gcos
-#undef mc68010
-#undef mc68020
-#undef mips
-#undef pdp11
-#undef sgi
-#undef sparc
-#undef sun
-#undef sun2
-#undef sun3
-#undef sun4
-#undef u370
-#undef u3b
-#undef u3b2
-#undef u3b5
-#undef unix
-#undef vax
-#endif
-#endif
-
-/***************************************************************************
- * $RCSfile: addheaderfooter,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:17 $
- * IPPL_VERSION_ID: $Id: addheaderfooter,v 1.1.1.1 2003/01/23 07:40:17 adelmann Exp $
- ***************************************************************************/
-
diff --git a/ippl/src/FFT/fftpack.F b/ippl/src/FFT/fftpack.F
deleted file mode 100644
index c47c95de3ca2760231ef16fa9a808344df097898..0000000000000000000000000000000000000000
--- a/ippl/src/FFT/fftpack.F
+++ /dev/null
@@ -1,3957 +0,0 @@
-c========================================================================
-c
-c The IPPL Framework
-c
-c This program was prepared by the Regents of the University of
-c Visit http://people.web.psi.ch/adelmann/ for more details
-c
-c========================================================================
-
-c=======================================================================
-c============================fftpack.F==================================
-c=============contains routines rffti, rfftf, rfftf,====================
-c==========cffti, cfftf, cfftb, sinti, and sint (from Netlib)===========
-c=======================================================================
-
-c Cray machines do not support double precision type. Use real.
-
-#if defined(IPPL_T3E)
-#define FLOAT REAL
-#else
-#define FLOAT DOUBLE PRECISION
-#endif
-
-c---------------------------DOUBLE PRECISION----------------------------
- subroutine rffti (n,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- if (n .eq. 1) return
- call rffti1 (n,wsave(n+1),wsave(2*n+1))
- return
- end
-
-
- subroutine rffti1 (n,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wa(*) ,xxifac(*) ,ntryh(4)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- data ntryh(1),ntryh(2),ntryh(3),ntryh(4)/4,2,3,5/
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nl = n
- nf = 0
- j = 0
- 101 j = j+1
- if (j-4) 102,102,103
- 102 ntry = ntryh(j)
- go to 104
- 103 ntry = ntry+2
- 104 nq = nl/ntry
- nr = nl-ntry*nq
- if (nr) 101,105,101
- 105 nf = nf+1
- jfac(nf+2) = ntry
- nl = nq
- if (ntry .ne. 2) go to 107
- if (nf .eq. 1) go to 107
- do 106 i=2,nf
- ib = nf-i+2
- jfac(ib+2) = jfac(ib+1)
- 106 continue
- jfac(3) = 2
- 107 if (nl .ne. 1) go to 104
- jfac(1) = n
- jfac(2) = nf
- tpi = 6.28318530717959
- argh = tpi/float(n)
- is = 0
- nfm1 = nf-1
- l1 = 1
- if (nfm1 .eq. 0) return
- do 110 k1=1,nfm1
- ip = jfac(k1+2)
- ld = 0
- l2 = l1*ip
- ido = n/l2
- ipm = ip-1
- do 109 j=1,ipm
- ld = ld+l1
- i = is
- argld = float(ld)*argh
- fi = 0.
- do 108 ii=3,ido,2
- i = i+2
- fi = fi+1.
- arg = fi*argld
- wa(i-1) = cos(arg)
- wa(i) = sin(arg)
- 108 continue
- is = is+ido
- 109 continue
- l1 = l2
- 110 continue
- do 111 i = 1, 15
- xxifac(i) = rfac(i)
- 111 continue
- return
- end
-
-
-
- subroutine rfftf (n,r,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension r(*) ,wsave(*)
- if (n .eq. 1) return
- call rfftf1 (n,r,wsave,wsave(n+1),wsave(2*n+1))
- return
- end
- subroutine rfftf1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 1
- l2 = n
- iw = n
- do 111 k1=1,nf
- kh = nf-k1
- ip = jfac(kh+3)
- l1 = l2/ip
- ido = n/l2
- idl1 = ido*l1
- iw = iw-(ip-1)*ido
- na = 1-na
- if (ip .ne. 4) go to 102
- ix2 = iw+ido
- ix3 = ix2+ido
- if (na .ne. 0) go to 101
- call radf4 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 110
- 101 call radf4 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- go to 110
- 102 if (ip .ne. 2) go to 104
- if (na .ne. 0) go to 103
- call radf2 (ido,l1,c,ch,wa(iw))
- go to 110
- 103 call radf2 (ido,l1,ch,c,wa(iw))
- go to 110
- 104 if (ip .ne. 3) go to 106
- ix2 = iw+ido
- if (na .ne. 0) go to 105
- call radf3 (ido,l1,c,ch,wa(iw),wa(ix2))
- go to 110
- 105 call radf3 (ido,l1,ch,c,wa(iw),wa(ix2))
- go to 110
- 106 if (ip .ne. 5) go to 108
- ix2 = iw+ido
- ix3 = ix2+ido
- ix4 = ix3+ido
- if (na .ne. 0) go to 107
- call radf5 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 110
- 107 call radf5 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 110
- 108 if (ido .eq. 1) na = 1-na
- if (na .ne. 0) go to 109
- call radfg (ido,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- na = 1
- go to 110
- 109 call radfg (ido,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- na = 0
- 110 l2 = l1
- 111 continue
- if (na .eq. 1) return
- do 112 i=1,n
- c(i) = ch(i)
- 112 continue
- return
- end
- subroutine radfg (ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,c2(idl1,ip),
- 2 ch2(idl1,ip) ,wa(*)
- data tpi/6.28318530717959/
- arg = tpi/float(ip)
- dcp = cos(arg)
- dsp = sin(arg)
- ipph = (ip+1)/2
- ipp2 = ip+2
- idp2 = ido+2
- nbd = (ido-1)/2
- if (ido .eq. 1) go to 119
- do 101 ik=1,idl1
- ch2(ik,1) = c2(ik,1)
- 101 continue
- do 103 j=2,ip
- do 102 k=1,l1
- ch(1,k,j) = c1(1,k,j)
- 102 continue
- 103 continue
- if (nbd .gt. l1) go to 107
- is = -ido
- do 106 j=2,ip
- is = is+ido
- idij = is
- do 105 i=3,ido,2
- idij = idij+2
- do 104 k=1,l1
- ch(i-1,k,j) = wa(idij-1)*c1(i-1,k,j)+wa(idij)*c1(i,k,j)
- ch(i,k,j) = wa(idij-1)*c1(i,k,j)-wa(idij)*c1(i-1,k,j)
- 104 continue
- 105 continue
- 106 continue
- go to 111
- 107 is = -ido
- do 110 j=2,ip
- is = is+ido
- do 109 k=1,l1
- idij = is
- do 108 i=3,ido,2
- idij = idij+2
- ch(i-1,k,j) = wa(idij-1)*c1(i-1,k,j)+wa(idij)*c1(i,k,j)
- ch(i,k,j) = wa(idij-1)*c1(i,k,j)-wa(idij)*c1(i-1,k,j)
- 108 continue
- 109 continue
- 110 continue
- 111 if (nbd .lt. l1) go to 115
- do 114 j=2,ipph
- jc = ipp2-j
- do 113 k=1,l1
- do 112 i=3,ido,2
- c1(i-1,k,j) = ch(i-1,k,j)+ch(i-1,k,jc)
- c1(i-1,k,jc) = ch(i,k,j)-ch(i,k,jc)
- c1(i,k,j) = ch(i,k,j)+ch(i,k,jc)
- c1(i,k,jc) = ch(i-1,k,jc)-ch(i-1,k,j)
- 112 continue
- 113 continue
- 114 continue
- go to 121
- 115 do 118 j=2,ipph
- jc = ipp2-j
- do 117 i=3,ido,2
- do 116 k=1,l1
- c1(i-1,k,j) = ch(i-1,k,j)+ch(i-1,k,jc)
- c1(i-1,k,jc) = ch(i,k,j)-ch(i,k,jc)
- c1(i,k,j) = ch(i,k,j)+ch(i,k,jc)
- c1(i,k,jc) = ch(i-1,k,jc)-ch(i-1,k,j)
- 116 continue
- 117 continue
- 118 continue
- go to 121
- 119 do 120 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 120 continue
- 121 do 123 j=2,ipph
- jc = ipp2-j
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)+ch(1,k,jc)
- c1(1,k,jc) = ch(1,k,jc)-ch(1,k,j)
- 122 continue
- 123 continue
-c
- ar1 = 1.
- ai1 = 0.
- do 127 l=2,ipph
- lc = ipp2-l
- ar1h = dcp*ar1-dsp*ai1
- ai1 = dcp*ai1+dsp*ar1
- ar1 = ar1h
- do 124 ik=1,idl1
- ch2(ik,l) = c2(ik,1)+ar1*c2(ik,2)
- ch2(ik,lc) = ai1*c2(ik,ip)
- 124 continue
- dc2 = ar1
- ds2 = ai1
- ar2 = ar1
- ai2 = ai1
- do 126 j=3,ipph
- jc = ipp2-j
- ar2h = dc2*ar2-ds2*ai2
- ai2 = dc2*ai2+ds2*ar2
- ar2 = ar2h
- do 125 ik=1,idl1
- ch2(ik,l) = ch2(ik,l)+ar2*c2(ik,j)
- ch2(ik,lc) = ch2(ik,lc)+ai2*c2(ik,jc)
- 125 continue
- 126 continue
- 127 continue
- do 129 j=2,ipph
- do 128 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+c2(ik,j)
- 128 continue
- 129 continue
-c
- if (ido .lt. l1) go to 132
- do 131 k=1,l1
- do 130 i=1,ido
- cc(i,1,k) = ch(i,k,1)
- 130 continue
- 131 continue
- go to 135
- 132 do 134 i=1,ido
- do 133 k=1,l1
- cc(i,1,k) = ch(i,k,1)
- 133 continue
- 134 continue
- 135 do 137 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 136 k=1,l1
- cc(ido,j2-2,k) = ch(1,k,j)
- cc(1,j2-1,k) = ch(1,k,jc)
- 136 continue
- 137 continue
- if (ido .eq. 1) return
- if (nbd .lt. l1) go to 141
- do 140 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 139 k=1,l1
- do 138 i=3,ido,2
- ic = idp2-i
- cc(i-1,j2-1,k) = ch(i-1,k,j)+ch(i-1,k,jc)
- cc(ic-1,j2-2,k) = ch(i-1,k,j)-ch(i-1,k,jc)
- cc(i,j2-1,k) = ch(i,k,j)+ch(i,k,jc)
- cc(ic,j2-2,k) = ch(i,k,jc)-ch(i,k,j)
- 138 continue
- 139 continue
- 140 continue
- return
- 141 do 144 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 143 i=3,ido,2
- ic = idp2-i
- do 142 k=1,l1
- cc(i-1,j2-1,k) = ch(i-1,k,j)+ch(i-1,k,jc)
- cc(ic-1,j2-2,k) = ch(i-1,k,j)-ch(i-1,k,jc)
- cc(i,j2-1,k) = ch(i,k,j)+ch(i,k,jc)
- cc(ic,j2-2,k) = ch(i,k,jc)-ch(i,k,j)
- 142 continue
- 143 continue
- 144 continue
- return
- end
- subroutine radf5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,l1,5) ,ch(ido,5,l1) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- do 101 k=1,l1
- cr2 = cc(1,k,5)+cc(1,k,2)
- ci5 = cc(1,k,5)-cc(1,k,2)
- cr3 = cc(1,k,4)+cc(1,k,3)
- ci4 = cc(1,k,4)-cc(1,k,3)
- ch(1,1,k) = cc(1,k,1)+cr2+cr3
- ch(ido,2,k) = cc(1,k,1)+tr11*cr2+tr12*cr3
- ch(1,3,k) = ti11*ci5+ti12*ci4
- ch(ido,4,k) = cc(1,k,1)+tr12*cr2+tr11*cr3
- ch(1,5,k) = ti12*ci5-ti11*ci4
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- dr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- di2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- dr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- di3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- dr4 = wa3(i-2)*cc(i-1,k,4)+wa3(i-1)*cc(i,k,4)
- di4 = wa3(i-2)*cc(i,k,4)-wa3(i-1)*cc(i-1,k,4)
- dr5 = wa4(i-2)*cc(i-1,k,5)+wa4(i-1)*cc(i,k,5)
- di5 = wa4(i-2)*cc(i,k,5)-wa4(i-1)*cc(i-1,k,5)
- cr2 = dr2+dr5
- ci5 = dr5-dr2
- cr5 = di2-di5
- ci2 = di2+di5
- cr3 = dr3+dr4
- ci4 = dr4-dr3
- cr4 = di3-di4
- ci3 = di3+di4
- ch(i-1,1,k) = cc(i-1,k,1)+cr2+cr3
- ch(i,1,k) = cc(i,k,1)+ci2+ci3
- tr2 = cc(i-1,k,1)+tr11*cr2+tr12*cr3
- ti2 = cc(i,k,1)+tr11*ci2+tr12*ci3
- tr3 = cc(i-1,k,1)+tr12*cr2+tr11*cr3
- ti3 = cc(i,k,1)+tr12*ci2+tr11*ci3
- tr5 = ti11*cr5+ti12*cr4
- ti5 = ti11*ci5+ti12*ci4
- tr4 = ti12*cr5-ti11*cr4
- ti4 = ti12*ci5-ti11*ci4
- ch(i-1,3,k) = tr2+tr5
- ch(ic-1,2,k) = tr2-tr5
- ch(i,3,k) = ti2+ti5
- ch(ic,2,k) = ti5-ti2
- ch(i-1,5,k) = tr3+tr4
- ch(ic-1,4,k) = tr3-tr4
- ch(i,5,k) = ti3+ti4
- ch(ic,4,k) = ti4-ti3
- 102 continue
- 103 continue
- return
- end
- subroutine radf3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,3,l1) ,cc(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- do 101 k=1,l1
- cr2 = cc(1,k,2)+cc(1,k,3)
- ch(1,1,k) = cc(1,k,1)+cr2
- ch(1,3,k) = taui*(cc(1,k,3)-cc(1,k,2))
- ch(ido,2,k) = cc(1,k,1)+taur*cr2
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- dr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- di2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- dr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- di3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- cr2 = dr2+dr3
- ci2 = di2+di3
- ch(i-1,1,k) = cc(i-1,k,1)+cr2
- ch(i,1,k) = cc(i,k,1)+ci2
- tr2 = cc(i-1,k,1)+taur*cr2
- ti2 = cc(i,k,1)+taur*ci2
- tr3 = taui*(di2-di3)
- ti3 = taui*(dr3-dr2)
- ch(i-1,3,k) = tr2+tr3
- ch(ic-1,2,k) = tr2-tr3
- ch(i,3,k) = ti2+ti3
- ch(ic,2,k) = ti3-ti2
- 102 continue
- 103 continue
- return
- end
- subroutine radf2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,2,l1) ,cc(ido,l1,2) ,
- 1 wa1(*)
- do 101 k=1,l1
- ch(1,1,k) = cc(1,k,1)+cc(1,k,2)
- ch(ido,2,k) = cc(1,k,1)-cc(1,k,2)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- tr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- ti2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- ch(i,1,k) = cc(i,k,1)+ti2
- ch(ic,2,k) = ti2-cc(i,k,1)
- ch(i-1,1,k) = cc(i-1,k,1)+tr2
- ch(ic-1,2,k) = cc(i-1,k,1)-tr2
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 do 106 k=1,l1
- ch(1,2,k) = -cc(ido,k,2)
- ch(ido,1,k) = cc(ido,k,1)
- 106 continue
- 107 return
- end
- subroutine radf4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,l1,4) ,ch(ido,4,l1) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- data hsqt2 /.7071067811865475/
- do 101 k=1,l1
- tr1 = cc(1,k,2)+cc(1,k,4)
- tr2 = cc(1,k,1)+cc(1,k,3)
- ch(1,1,k) = tr1+tr2
- ch(ido,4,k) = tr2-tr1
- ch(ido,2,k) = cc(1,k,1)-cc(1,k,3)
- ch(1,3,k) = cc(1,k,4)-cc(1,k,2)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- cr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- ci2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- cr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- ci3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- cr4 = wa3(i-2)*cc(i-1,k,4)+wa3(i-1)*cc(i,k,4)
- ci4 = wa3(i-2)*cc(i,k,4)-wa3(i-1)*cc(i-1,k,4)
- tr1 = cr2+cr4
- tr4 = cr4-cr2
- ti1 = ci2+ci4
- ti4 = ci2-ci4
- ti2 = cc(i,k,1)+ci3
- ti3 = cc(i,k,1)-ci3
- tr2 = cc(i-1,k,1)+cr3
- tr3 = cc(i-1,k,1)-cr3
- ch(i-1,1,k) = tr1+tr2
- ch(ic-1,4,k) = tr2-tr1
- ch(i,1,k) = ti1+ti2
- ch(ic,4,k) = ti1-ti2
- ch(i-1,3,k) = ti4+tr3
- ch(ic-1,2,k) = tr3-ti4
- ch(i,3,k) = tr4+ti3
- ch(ic,2,k) = tr4-ti3
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 continue
- do 106 k=1,l1
- ti1 = -hsqt2*(cc(ido,k,2)+cc(ido,k,4))
- tr1 = hsqt2*(cc(ido,k,2)-cc(ido,k,4))
- ch(ido,1,k) = tr1+cc(ido,k,1)
- ch(ido,3,k) = cc(ido,k,1)-tr1
- ch(1,2,k) = ti1-cc(ido,k,3)
- ch(1,4,k) = ti1+cc(ido,k,3)
- 106 continue
- 107 return
- end
-
-
-
- subroutine rfftb (n,r,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension r(*) ,wsave(*)
- if (n .eq. 1) return
- call rfftb1 (n,r,wsave,wsave(n+1),wsave(2*n+1))
- return
- end
- subroutine rfftb1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idl1 = ido*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+ido
- ix3 = ix2+ido
- if (na .ne. 0) go to 101
- call radb4 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call radb4 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call radb2 (ido,l1,c,ch,wa(iw))
- go to 105
- 104 call radb2 (ido,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+ido
- if (na .ne. 0) go to 107
- call radb3 (ido,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call radb3 (ido,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+ido
- ix3 = ix2+ido
- ix4 = ix3+ido
- if (na .ne. 0) go to 110
- call radb5 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call radb5 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call radbg (ido,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call radbg (ido,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (ido .eq. 1) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*ido
- 116 continue
- if (na .eq. 0) return
- do 117 i=1,n
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine radbg (ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,c2(idl1,ip),
- 2 ch2(idl1,ip) ,wa(*)
- data tpi/6.28318530717959/
- arg = tpi/float(ip)
- dcp = cos(arg)
- dsp = sin(arg)
- idp2 = ido+2
- nbd = (ido-1)/2
- ipp2 = ip+2
- ipph = (ip+1)/2
- if (ido .lt. l1) go to 103
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 101 continue
- 102 continue
- go to 106
- 103 do 105 i=1,ido
- do 104 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- 106 do 108 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 107 k=1,l1
- ch(1,k,j) = cc(ido,j2-2,k)+cc(ido,j2-2,k)
- ch(1,k,jc) = cc(1,j2-1,k)+cc(1,j2-1,k)
- 107 continue
- 108 continue
- if (ido .eq. 1) go to 116
- if (nbd .lt. l1) go to 112
- do 111 j=2,ipph
- jc = ipp2-j
- do 110 k=1,l1
- do 109 i=3,ido,2
- ic = idp2-i
- ch(i-1,k,j) = cc(i-1,2*j-1,k)+cc(ic-1,2*j-2,k)
- ch(i-1,k,jc) = cc(i-1,2*j-1,k)-cc(ic-1,2*j-2,k)
- ch(i,k,j) = cc(i,2*j-1,k)-cc(ic,2*j-2,k)
- ch(i,k,jc) = cc(i,2*j-1,k)+cc(ic,2*j-2,k)
- 109 continue
- 110 continue
- 111 continue
- go to 116
- 112 do 115 j=2,ipph
- jc = ipp2-j
- do 114 i=3,ido,2
- ic = idp2-i
- do 113 k=1,l1
- ch(i-1,k,j) = cc(i-1,2*j-1,k)+cc(ic-1,2*j-2,k)
- ch(i-1,k,jc) = cc(i-1,2*j-1,k)-cc(ic-1,2*j-2,k)
- ch(i,k,j) = cc(i,2*j-1,k)-cc(ic,2*j-2,k)
- ch(i,k,jc) = cc(i,2*j-1,k)+cc(ic,2*j-2,k)
- 113 continue
- 114 continue
- 115 continue
- 116 ar1 = 1.
- ai1 = 0.
- do 120 l=2,ipph
- lc = ipp2-l
- ar1h = dcp*ar1-dsp*ai1
- ai1 = dcp*ai1+dsp*ar1
- ar1 = ar1h
- do 117 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+ar1*ch2(ik,2)
- c2(ik,lc) = ai1*ch2(ik,ip)
- 117 continue
- dc2 = ar1
- ds2 = ai1
- ar2 = ar1
- ai2 = ai1
- do 119 j=3,ipph
- jc = ipp2-j
- ar2h = dc2*ar2-ds2*ai2
- ai2 = dc2*ai2+ds2*ar2
- ar2 = ar2h
- do 118 ik=1,idl1
- c2(ik,l) = c2(ik,l)+ar2*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)+ai2*ch2(ik,jc)
- 118 continue
- 119 continue
- 120 continue
- do 122 j=2,ipph
- do 121 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 121 continue
- 122 continue
- do 124 j=2,ipph
- jc = ipp2-j
- do 123 k=1,l1
- ch(1,k,j) = c1(1,k,j)-c1(1,k,jc)
- ch(1,k,jc) = c1(1,k,j)+c1(1,k,jc)
- 123 continue
- 124 continue
- if (ido .eq. 1) go to 132
- if (nbd .lt. l1) go to 128
- do 127 j=2,ipph
- jc = ipp2-j
- do 126 k=1,l1
- do 125 i=3,ido,2
- ch(i-1,k,j) = c1(i-1,k,j)-c1(i,k,jc)
- ch(i-1,k,jc) = c1(i-1,k,j)+c1(i,k,jc)
- ch(i,k,j) = c1(i,k,j)+c1(i-1,k,jc)
- ch(i,k,jc) = c1(i,k,j)-c1(i-1,k,jc)
- 125 continue
- 126 continue
- 127 continue
- go to 132
- 128 do 131 j=2,ipph
- jc = ipp2-j
- do 130 i=3,ido,2
- do 129 k=1,l1
- ch(i-1,k,j) = c1(i-1,k,j)-c1(i,k,jc)
- ch(i-1,k,jc) = c1(i-1,k,j)+c1(i,k,jc)
- ch(i,k,j) = c1(i,k,j)+c1(i-1,k,jc)
- ch(i,k,jc) = c1(i,k,j)-c1(i-1,k,jc)
- 129 continue
- 130 continue
- 131 continue
- 132 continue
- if (ido .eq. 1) return
- do 133 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 133 continue
- do 135 j=2,ip
- do 134 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- 134 continue
- 135 continue
- if (nbd .gt. l1) go to 139
- is = -ido
- do 138 j=2,ip
- is = is+ido
- idij = is
- do 137 i=3,ido,2
- idij = idij+2
- do 136 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 136 continue
- 137 continue
- 138 continue
- go to 143
- 139 is = -ido
- do 142 j=2,ip
- is = is+ido
- do 141 k=1,l1
- idij = is
- do 140 i=3,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 140 continue
- 141 continue
- 142 continue
- 143 return
- end
- subroutine radb5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- do 101 k=1,l1
- ti5 = cc(1,3,k)+cc(1,3,k)
- ti4 = cc(1,5,k)+cc(1,5,k)
- tr2 = cc(ido,2,k)+cc(ido,2,k)
- tr3 = cc(ido,4,k)+cc(ido,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci5 = ti11*ti5+ti12*ti4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(1,k,5) = cr2+ci5
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- ti5 = cc(i,3,k)+cc(ic,2,k)
- ti2 = cc(i,3,k)-cc(ic,2,k)
- ti4 = cc(i,5,k)+cc(ic,4,k)
- ti3 = cc(i,5,k)-cc(ic,4,k)
- tr5 = cc(i-1,3,k)-cc(ic-1,2,k)
- tr2 = cc(i-1,3,k)+cc(ic-1,2,k)
- tr4 = cc(i-1,5,k)-cc(ic-1,4,k)
- tr3 = cc(i-1,5,k)+cc(ic-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-2)*dr2-wa1(i-1)*di2
- ch(i,k,2) = wa1(i-2)*di2+wa1(i-1)*dr2
- ch(i-1,k,3) = wa2(i-2)*dr3-wa2(i-1)*di3
- ch(i,k,3) = wa2(i-2)*di3+wa2(i-1)*dr3
- ch(i-1,k,4) = wa3(i-2)*dr4-wa3(i-1)*di4
- ch(i,k,4) = wa3(i-2)*di4+wa3(i-1)*dr4
- ch(i-1,k,5) = wa4(i-2)*dr5-wa4(i-1)*di5
- ch(i,k,5) = wa4(i-2)*di5+wa4(i-1)*dr5
- 102 continue
- 103 continue
- return
- end
- subroutine radb3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- do 101 k=1,l1
- tr2 = cc(ido,2,k)+cc(ido,2,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ci3 = taui*(cc(1,3,k)+cc(1,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- tr2 = cc(i-1,3,k)+cc(ic-1,2,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,3,k)-cc(ic,2,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,3,k)-cc(ic-1,2,k))
- ci3 = taui*(cc(i,3,k)+cc(ic,2,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i-1,k,2) = wa1(i-2)*dr2-wa1(i-1)*di2
- ch(i,k,2) = wa1(i-2)*di2+wa1(i-1)*dr2
- ch(i-1,k,3) = wa2(i-2)*dr3-wa2(i-1)*di3
- ch(i,k,3) = wa2(i-2)*di3+wa2(i-1)*dr3
- 102 continue
- 103 continue
- return
- end
- subroutine radb2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(ido,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(ido,2,k)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- ch(i-1,k,1) = cc(i-1,1,k)+cc(ic-1,2,k)
- tr2 = cc(i-1,1,k)-cc(ic-1,2,k)
- ch(i,k,1) = cc(i,1,k)-cc(ic,2,k)
- ti2 = cc(i,1,k)+cc(ic,2,k)
- ch(i-1,k,2) = wa1(i-2)*tr2-wa1(i-1)*ti2
- ch(i,k,2) = wa1(i-2)*ti2+wa1(i-1)*tr2
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 do 106 k=1,l1
- ch(ido,k,1) = cc(ido,1,k)+cc(ido,1,k)
- ch(ido,k,2) = -(cc(1,2,k)+cc(1,2,k))
- 106 continue
- 107 return
- end
- subroutine radb4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- data sqrt2 /1.414213562373095/
- do 101 k=1,l1
- tr1 = cc(1,1,k)-cc(ido,4,k)
- tr2 = cc(1,1,k)+cc(ido,4,k)
- tr3 = cc(ido,2,k)+cc(ido,2,k)
- tr4 = cc(1,3,k)+cc(1,3,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,2) = tr1-tr4
- ch(1,k,3) = tr2-tr3
- ch(1,k,4) = tr1+tr4
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- ti1 = cc(i,1,k)+cc(ic,4,k)
- ti2 = cc(i,1,k)-cc(ic,4,k)
- ti3 = cc(i,3,k)-cc(ic,2,k)
- tr4 = cc(i,3,k)+cc(ic,2,k)
- tr1 = cc(i-1,1,k)-cc(ic-1,4,k)
- tr2 = cc(i-1,1,k)+cc(ic-1,4,k)
- ti4 = cc(i-1,3,k)-cc(ic-1,2,k)
- tr3 = cc(i-1,3,k)+cc(ic-1,2,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1-tr4
- cr4 = tr1+tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-2)*cr2-wa1(i-1)*ci2
- ch(i,k,2) = wa1(i-2)*ci2+wa1(i-1)*cr2
- ch(i-1,k,3) = wa2(i-2)*cr3-wa2(i-1)*ci3
- ch(i,k,3) = wa2(i-2)*ci3+wa2(i-1)*cr3
- ch(i-1,k,4) = wa3(i-2)*cr4-wa3(i-1)*ci4
- ch(i,k,4) = wa3(i-2)*ci4+wa3(i-1)*cr4
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 continue
- do 106 k=1,l1
- ti1 = cc(1,2,k)+cc(1,4,k)
- ti2 = cc(1,4,k)-cc(1,2,k)
- tr1 = cc(ido,1,k)-cc(ido,3,k)
- tr2 = cc(ido,1,k)+cc(ido,3,k)
- ch(ido,k,1) = tr2+tr2
- ch(ido,k,2) = sqrt2*(tr1-ti1)
- ch(ido,k,3) = ti2+ti2
- ch(ido,k,4) = -sqrt2*(tr1+ti1)
- 106 continue
- 107 return
- end
-
-
-
- subroutine cffti (n,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call cffti1 (n,wsave(iw1),wsave(iw2))
- return
- end
- subroutine cffti1 (n,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wa(*) ,xxifac(*) ,ntryh(4)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- data ntryh(1),ntryh(2),ntryh(3),ntryh(4)/3,4,2,5/
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nl = n
- nf = 0
- j = 0
- 101 j = j+1
- if (j-4) 102,102,103
- 102 ntry = ntryh(j)
- go to 104
- 103 ntry = ntry+2
- 104 nq = nl/ntry
- nr = nl-ntry*nq
- if (nr) 101,105,101
- 105 nf = nf+1
- jfac(nf+2) = ntry
- nl = nq
- if (ntry .ne. 2) go to 107
- if (nf .eq. 1) go to 107
- do 106 i=2,nf
- ib = nf-i+2
- jfac(ib+2) = jfac(ib+1)
- 106 continue
- jfac(3) = 2
- 107 if (nl .ne. 1) go to 104
- jfac(1) = n
- jfac(2) = nf
- tpi = 6.28318530717959
- argh = tpi/float(n)
- i = 2
- l1 = 1
- do 110 k1=1,nf
- ip = jfac(k1+2)
- ld = 0
- l2 = l1*ip
- ido = n/l2
- idot = ido+ido+2
- ipm = ip-1
- do 109 j=1,ipm
- i1 = i
- wa(i-1) = 1.
- wa(i) = 0.
- ld = ld+l1
- fi = 0.
- argld = float(ld)*argh
- do 108 ii=4,idot,2
- i = i+2
- fi = fi+1.
- arg = fi*argld
- wa(i-1) = cos(arg)
- wa(i) = sin(arg)
- 108 continue
- if (ip .le. 5) go to 109
- wa(i1-1) = wa(i-1)
- wa(i1) = wa(i)
- 109 continue
- l1 = l2
- 110 continue
- do 111 i = 1, 15
- xxifac(i) = rfac(i)
- 111 continue
- return
- end
-
-
-
- subroutine cfftf (n,c,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension c(*) ,wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call cfftf1 (n,c,wsave,wsave(iw1),wsave(iw2))
- return
- end
- subroutine cfftf1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idot = ido+ido
- idl1 = idot*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+idot
- ix3 = ix2+idot
- if (na .ne. 0) go to 101
- call passf4 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call passf4 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call passf2 (idot,l1,c,ch,wa(iw))
- go to 105
- 104 call passf2 (idot,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+idot
- if (na .ne. 0) go to 107
- call passf3 (idot,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call passf3 (idot,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+idot
- ix3 = ix2+idot
- ix4 = ix3+idot
- if (na .ne. 0) go to 110
- call passf5 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call passf5 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call passf (nac,idot,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call passf (nac,idot,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (nac .ne. 0) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*idot
- 116 continue
- if (na .eq. 0) return
- n2 = n+n
- do 117 i=1,n2
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine passf (nac,ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,wa(*) ,c2(idl1,ip),
- 2 ch2(idl1,ip)
- idot = ido/2
- nt = ip*idl1
- ipp2 = ip+2
- ipph = (ip+1)/2
- idp = ip*ido
-c
- if (ido .lt. l1) go to 106
- do 103 j=2,ipph
- jc = ipp2-j
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 101 continue
- 102 continue
- 103 continue
- do 105 k=1,l1
- do 104 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- go to 112
- 106 do 109 j=2,ipph
- jc = ipp2-j
- do 108 i=1,ido
- do 107 k=1,l1
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 107 continue
- 108 continue
- 109 continue
- do 111 i=1,ido
- do 110 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 110 continue
- 111 continue
- 112 idl = 2-ido
- inc = 0
- do 116 l=2,ipph
- lc = ipp2-l
- idl = idl+ido
- do 113 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+wa(idl-1)*ch2(ik,2)
- c2(ik,lc) = -wa(idl)*ch2(ik,ip)
- 113 continue
- idlj = idl
- inc = inc+ido
- do 115 j=3,ipph
- jc = ipp2-j
- idlj = idlj+inc
- if (idlj .gt. idp) idlj = idlj-idp
- war = wa(idlj-1)
- wai = wa(idlj)
- do 114 ik=1,idl1
- c2(ik,l) = c2(ik,l)+war*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)-wai*ch2(ik,jc)
- 114 continue
- 115 continue
- 116 continue
- do 118 j=2,ipph
- do 117 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 117 continue
- 118 continue
- do 120 j=2,ipph
- jc = ipp2-j
- do 119 ik=2,idl1,2
- ch2(ik-1,j) = c2(ik-1,j)-c2(ik,jc)
- ch2(ik-1,jc) = c2(ik-1,j)+c2(ik,jc)
- ch2(ik,j) = c2(ik,j)+c2(ik-1,jc)
- ch2(ik,jc) = c2(ik,j)-c2(ik-1,jc)
- 119 continue
- 120 continue
- nac = 1
- if (ido .eq. 2) return
- nac = 0
- do 121 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 121 continue
- do 123 j=2,ip
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- c1(2,k,j) = ch(2,k,j)
- 122 continue
- 123 continue
- if (idot .gt. l1) go to 127
- idij = 0
- do 126 j=2,ip
- idij = idij+2
- do 125 i=4,ido,2
- idij = idij+2
- do 124 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)+wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)-wa(idij)*ch(i-1,k,j)
- 124 continue
- 125 continue
- 126 continue
- return
- 127 idj = 2-ido
- do 130 j=2,ip
- idj = idj+ido
- do 129 k=1,l1
- idij = idj
- do 128 i=4,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)+wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)-wa(idij)*ch(i-1,k,j)
- 128 continue
- 129 continue
- 130 continue
- return
- end
- subroutine passf5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,-.951056516295154,
- 1-.809016994374947,-.587785252292473/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti5 = cc(2,2,k)-cc(2,5,k)
- ti2 = cc(2,2,k)+cc(2,5,k)
- ti4 = cc(2,3,k)-cc(2,4,k)
- ti3 = cc(2,3,k)+cc(2,4,k)
- tr5 = cc(1,2,k)-cc(1,5,k)
- tr2 = cc(1,2,k)+cc(1,5,k)
- tr4 = cc(1,3,k)-cc(1,4,k)
- tr3 = cc(1,3,k)+cc(1,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- ch(2,k,1) = cc(2,1,k)+ti2+ti3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(2,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(2,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,5) = cr2+ci5
- ch(2,k,2) = ci2+cr5
- ch(2,k,3) = ci3+cr4
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(2,k,4) = ci3-cr4
- ch(2,k,5) = ci2-cr5
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti5 = cc(i,2,k)-cc(i,5,k)
- ti2 = cc(i,2,k)+cc(i,5,k)
- ti4 = cc(i,3,k)-cc(i,4,k)
- ti3 = cc(i,3,k)+cc(i,4,k)
- tr5 = cc(i-1,2,k)-cc(i-1,5,k)
- tr2 = cc(i-1,2,k)+cc(i-1,5,k)
- tr4 = cc(i-1,3,k)-cc(i-1,4,k)
- tr3 = cc(i-1,3,k)+cc(i-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-1)*dr2+wa1(i)*di2
- ch(i,k,2) = wa1(i-1)*di2-wa1(i)*dr2
- ch(i-1,k,3) = wa2(i-1)*dr3+wa2(i)*di3
- ch(i,k,3) = wa2(i-1)*di3-wa2(i)*dr3
- ch(i-1,k,4) = wa3(i-1)*dr4+wa3(i)*di4
- ch(i,k,4) = wa3(i-1)*di4-wa3(i)*dr4
- ch(i-1,k,5) = wa4(i-1)*dr5+wa4(i)*di5
- ch(i,k,5) = wa4(i-1)*di5-wa4(i)*dr5
- 103 continue
- 104 continue
- return
- end
- subroutine passf3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,-.866025403784439/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- tr2 = cc(1,2,k)+cc(1,3,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ti2 = cc(2,2,k)+cc(2,3,k)
- ci2 = cc(2,1,k)+taur*ti2
- ch(2,k,1) = cc(2,1,k)+ti2
- cr3 = taui*(cc(1,2,k)-cc(1,3,k))
- ci3 = taui*(cc(2,2,k)-cc(2,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- ch(2,k,2) = ci2+cr3
- ch(2,k,3) = ci2-cr3
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- tr2 = cc(i-1,2,k)+cc(i-1,3,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,2,k)+cc(i,3,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,2,k)-cc(i-1,3,k))
- ci3 = taui*(cc(i,2,k)-cc(i,3,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i,k,2) = wa1(i-1)*di2-wa1(i)*dr2
- ch(i-1,k,2) = wa1(i-1)*dr2+wa1(i)*di2
- ch(i,k,3) = wa2(i-1)*di3-wa2(i)*dr3
- ch(i-1,k,3) = wa2(i-1)*dr3+wa2(i)*di3
- 103 continue
- 104 continue
- return
- end
- subroutine passf2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- if (ido .gt. 2) go to 102
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(1,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(1,2,k)
- ch(2,k,1) = cc(2,1,k)+cc(2,2,k)
- ch(2,k,2) = cc(2,1,k)-cc(2,2,k)
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ch(i-1,k,1) = cc(i-1,1,k)+cc(i-1,2,k)
- tr2 = cc(i-1,1,k)-cc(i-1,2,k)
- ch(i,k,1) = cc(i,1,k)+cc(i,2,k)
- ti2 = cc(i,1,k)-cc(i,2,k)
- ch(i,k,2) = wa1(i-1)*ti2-wa1(i)*tr2
- ch(i-1,k,2) = wa1(i-1)*tr2+wa1(i)*ti2
- 103 continue
- 104 continue
- return
- end
- subroutine passf4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti1 = cc(2,1,k)-cc(2,3,k)
- ti2 = cc(2,1,k)+cc(2,3,k)
- tr4 = cc(2,2,k)-cc(2,4,k)
- ti3 = cc(2,2,k)+cc(2,4,k)
- tr1 = cc(1,1,k)-cc(1,3,k)
- tr2 = cc(1,1,k)+cc(1,3,k)
- ti4 = cc(1,4,k)-cc(1,2,k)
- tr3 = cc(1,2,k)+cc(1,4,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,3) = tr2-tr3
- ch(2,k,1) = ti2+ti3
- ch(2,k,3) = ti2-ti3
- ch(1,k,2) = tr1+tr4
- ch(1,k,4) = tr1-tr4
- ch(2,k,2) = ti1+ti4
- ch(2,k,4) = ti1-ti4
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti1 = cc(i,1,k)-cc(i,3,k)
- ti2 = cc(i,1,k)+cc(i,3,k)
- ti3 = cc(i,2,k)+cc(i,4,k)
- tr4 = cc(i,2,k)-cc(i,4,k)
- tr1 = cc(i-1,1,k)-cc(i-1,3,k)
- tr2 = cc(i-1,1,k)+cc(i-1,3,k)
- ti4 = cc(i-1,4,k)-cc(i-1,2,k)
- tr3 = cc(i-1,2,k)+cc(i-1,4,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1+tr4
- cr4 = tr1-tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-1)*cr2+wa1(i)*ci2
- ch(i,k,2) = wa1(i-1)*ci2-wa1(i)*cr2
- ch(i-1,k,3) = wa2(i-1)*cr3+wa2(i)*ci3
- ch(i,k,3) = wa2(i-1)*ci3-wa2(i)*cr3
- ch(i-1,k,4) = wa3(i-1)*cr4+wa3(i)*ci4
- ch(i,k,4) = wa3(i-1)*ci4-wa3(i)*cr4
- 103 continue
- 104 continue
- return
- end
-
-
-
- subroutine cfftb (n,c,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension c(*) ,wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call cfftb1 (n,c,wsave,wsave(iw1),wsave(iw2))
- return
- end
- subroutine cfftb1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idot = ido+ido
- idl1 = idot*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+idot
- ix3 = ix2+idot
- if (na .ne. 0) go to 101
- call passb4 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call passb4 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call passb2 (idot,l1,c,ch,wa(iw))
- go to 105
- 104 call passb2 (idot,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+idot
- if (na .ne. 0) go to 107
- call passb3 (idot,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call passb3 (idot,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+idot
- ix3 = ix2+idot
- ix4 = ix3+idot
- if (na .ne. 0) go to 110
- call passb5 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call passb5 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call passb (nac,idot,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call passb (nac,idot,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (nac .ne. 0) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*idot
- 116 continue
- if (na .eq. 0) return
- n2 = n+n
- do 117 i=1,n2
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine passb (nac,ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,wa(*) ,c2(idl1,ip),
- 2 ch2(idl1,ip)
- idot = ido/2
- nt = ip*idl1
- ipp2 = ip+2
- ipph = (ip+1)/2
- idp = ip*ido
-c
- if (ido .lt. l1) go to 106
- do 103 j=2,ipph
- jc = ipp2-j
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 101 continue
- 102 continue
- 103 continue
- do 105 k=1,l1
- do 104 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- go to 112
- 106 do 109 j=2,ipph
- jc = ipp2-j
- do 108 i=1,ido
- do 107 k=1,l1
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 107 continue
- 108 continue
- 109 continue
- do 111 i=1,ido
- do 110 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 110 continue
- 111 continue
- 112 idl = 2-ido
- inc = 0
- do 116 l=2,ipph
- lc = ipp2-l
- idl = idl+ido
- do 113 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+wa(idl-1)*ch2(ik,2)
- c2(ik,lc) = wa(idl)*ch2(ik,ip)
- 113 continue
- idlj = idl
- inc = inc+ido
- do 115 j=3,ipph
- jc = ipp2-j
- idlj = idlj+inc
- if (idlj .gt. idp) idlj = idlj-idp
- war = wa(idlj-1)
- wai = wa(idlj)
- do 114 ik=1,idl1
- c2(ik,l) = c2(ik,l)+war*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)+wai*ch2(ik,jc)
- 114 continue
- 115 continue
- 116 continue
- do 118 j=2,ipph
- do 117 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 117 continue
- 118 continue
- do 120 j=2,ipph
- jc = ipp2-j
- do 119 ik=2,idl1,2
- ch2(ik-1,j) = c2(ik-1,j)-c2(ik,jc)
- ch2(ik-1,jc) = c2(ik-1,j)+c2(ik,jc)
- ch2(ik,j) = c2(ik,j)+c2(ik-1,jc)
- ch2(ik,jc) = c2(ik,j)-c2(ik-1,jc)
- 119 continue
- 120 continue
- nac = 1
- if (ido .eq. 2) return
- nac = 0
- do 121 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 121 continue
- do 123 j=2,ip
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- c1(2,k,j) = ch(2,k,j)
- 122 continue
- 123 continue
- if (idot .gt. l1) go to 127
- idij = 0
- do 126 j=2,ip
- idij = idij+2
- do 125 i=4,ido,2
- idij = idij+2
- do 124 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 124 continue
- 125 continue
- 126 continue
- return
- 127 idj = 2-ido
- do 130 j=2,ip
- idj = idj+ido
- do 129 k=1,l1
- idij = idj
- do 128 i=4,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 128 continue
- 129 continue
- 130 continue
- return
- end
- subroutine passb5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti5 = cc(2,2,k)-cc(2,5,k)
- ti2 = cc(2,2,k)+cc(2,5,k)
- ti4 = cc(2,3,k)-cc(2,4,k)
- ti3 = cc(2,3,k)+cc(2,4,k)
- tr5 = cc(1,2,k)-cc(1,5,k)
- tr2 = cc(1,2,k)+cc(1,5,k)
- tr4 = cc(1,3,k)-cc(1,4,k)
- tr3 = cc(1,3,k)+cc(1,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- ch(2,k,1) = cc(2,1,k)+ti2+ti3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(2,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(2,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,5) = cr2+ci5
- ch(2,k,2) = ci2+cr5
- ch(2,k,3) = ci3+cr4
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(2,k,4) = ci3-cr4
- ch(2,k,5) = ci2-cr5
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti5 = cc(i,2,k)-cc(i,5,k)
- ti2 = cc(i,2,k)+cc(i,5,k)
- ti4 = cc(i,3,k)-cc(i,4,k)
- ti3 = cc(i,3,k)+cc(i,4,k)
- tr5 = cc(i-1,2,k)-cc(i-1,5,k)
- tr2 = cc(i-1,2,k)+cc(i-1,5,k)
- tr4 = cc(i-1,3,k)-cc(i-1,4,k)
- tr3 = cc(i-1,3,k)+cc(i-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-1)*dr2-wa1(i)*di2
- ch(i,k,2) = wa1(i-1)*di2+wa1(i)*dr2
- ch(i-1,k,3) = wa2(i-1)*dr3-wa2(i)*di3
- ch(i,k,3) = wa2(i-1)*di3+wa2(i)*dr3
- ch(i-1,k,4) = wa3(i-1)*dr4-wa3(i)*di4
- ch(i,k,4) = wa3(i-1)*di4+wa3(i)*dr4
- ch(i-1,k,5) = wa4(i-1)*dr5-wa4(i)*di5
- ch(i,k,5) = wa4(i-1)*di5+wa4(i)*dr5
- 103 continue
- 104 continue
- return
- end
- subroutine passb3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- tr2 = cc(1,2,k)+cc(1,3,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ti2 = cc(2,2,k)+cc(2,3,k)
- ci2 = cc(2,1,k)+taur*ti2
- ch(2,k,1) = cc(2,1,k)+ti2
- cr3 = taui*(cc(1,2,k)-cc(1,3,k))
- ci3 = taui*(cc(2,2,k)-cc(2,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- ch(2,k,2) = ci2+cr3
- ch(2,k,3) = ci2-cr3
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- tr2 = cc(i-1,2,k)+cc(i-1,3,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,2,k)+cc(i,3,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,2,k)-cc(i-1,3,k))
- ci3 = taui*(cc(i,2,k)-cc(i,3,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i,k,2) = wa1(i-1)*di2+wa1(i)*dr2
- ch(i-1,k,2) = wa1(i-1)*dr2-wa1(i)*di2
- ch(i,k,3) = wa2(i-1)*di3+wa2(i)*dr3
- ch(i-1,k,3) = wa2(i-1)*dr3-wa2(i)*di3
- 103 continue
- 104 continue
- return
- end
- subroutine passb2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- if (ido .gt. 2) go to 102
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(1,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(1,2,k)
- ch(2,k,1) = cc(2,1,k)+cc(2,2,k)
- ch(2,k,2) = cc(2,1,k)-cc(2,2,k)
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ch(i-1,k,1) = cc(i-1,1,k)+cc(i-1,2,k)
- tr2 = cc(i-1,1,k)-cc(i-1,2,k)
- ch(i,k,1) = cc(i,1,k)+cc(i,2,k)
- ti2 = cc(i,1,k)-cc(i,2,k)
- ch(i,k,2) = wa1(i-1)*ti2+wa1(i)*tr2
- ch(i-1,k,2) = wa1(i-1)*tr2-wa1(i)*ti2
- 103 continue
- 104 continue
- return
- end
- subroutine passb4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti1 = cc(2,1,k)-cc(2,3,k)
- ti2 = cc(2,1,k)+cc(2,3,k)
- tr4 = cc(2,4,k)-cc(2,2,k)
- ti3 = cc(2,2,k)+cc(2,4,k)
- tr1 = cc(1,1,k)-cc(1,3,k)
- tr2 = cc(1,1,k)+cc(1,3,k)
- ti4 = cc(1,2,k)-cc(1,4,k)
- tr3 = cc(1,2,k)+cc(1,4,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,3) = tr2-tr3
- ch(2,k,1) = ti2+ti3
- ch(2,k,3) = ti2-ti3
- ch(1,k,2) = tr1+tr4
- ch(1,k,4) = tr1-tr4
- ch(2,k,2) = ti1+ti4
- ch(2,k,4) = ti1-ti4
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti1 = cc(i,1,k)-cc(i,3,k)
- ti2 = cc(i,1,k)+cc(i,3,k)
- ti3 = cc(i,2,k)+cc(i,4,k)
- tr4 = cc(i,4,k)-cc(i,2,k)
- tr1 = cc(i-1,1,k)-cc(i-1,3,k)
- tr2 = cc(i-1,1,k)+cc(i-1,3,k)
- ti4 = cc(i-1,2,k)-cc(i-1,4,k)
- tr3 = cc(i-1,2,k)+cc(i-1,4,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1+tr4
- cr4 = tr1-tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-1)*cr2-wa1(i)*ci2
- ch(i,k,2) = wa1(i-1)*ci2+wa1(i)*cr2
- ch(i-1,k,3) = wa2(i-1)*cr3-wa2(i)*ci3
- ch(i,k,3) = wa2(i-1)*ci3+wa2(i)*cr3
- ch(i-1,k,4) = wa3(i-1)*cr4-wa3(i)*ci4
- ch(i,k,4) = wa3(i-1)*ci4+wa3(i)*cr4
- 103 continue
- 104 continue
- return
- end
-
-
-
- subroutine sinti (n,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- data pi /3.14159265358979/
- if (n .le. 1) return
- ns2 = n/2
- np1 = n+1
- dt = pi/float(np1)
- do 101 k=1,ns2
- wsave(k) = 2.*sin(k*dt)
- 101 continue
- call rffti (np1,wsave(ns2+1))
- return
- end
-
-
-
- subroutine sint (n,x,wsave)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension x(*) ,wsave(*)
-
-c with n = 20
-c iw1 = 11
-c iw2 = 11+20+1=32
-c iw3 = 32+21=53
- np1 = n+1
- iw1 = n/2+1
- iw2 = iw1+np1
- iw3 = iw2+np1
- call sint1(n,x,wsave,wsave(iw1),wsave(iw2),wsave(iw3))
- return
- end
-
-
- subroutine sint1(n,war,was,xh,x,xxifac)
-c----------------------------TJW
- implicit FLOAT (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension war(*),was(*),x(*),xh(*),xxifac(*)
- data sqrt3 /1.73205080756888/
- do 100 i=1,n
- xh(i) = war(i)
- war(i) = x(i)
- 100 continue
- if (n-2) 101,102,103
- 101 xh(1) = xh(1)+xh(1)
- go to 106
- 102 xhold = sqrt3*(xh(1)+xh(2))
- xh(2) = sqrt3*(xh(1)-xh(2))
- xh(1) = xhold
- go to 106
- 103 np1 = n+1
- ns2 = n/2
- x(1) = 0.
- do 104 k=1,ns2
- kc = np1-k
- t1 = xh(k)-xh(kc)
- t2 = was(k)*(xh(k)+xh(kc))
- x(k+1) = t1+t2
- x(kc+1) = t2-t1
- 104 continue
- modn = mod(n,2)
- if (modn .ne. 0) x(ns2+2) = 4.*xh(ns2+1)
- call rfftf1 (np1,x,xh,war,xxifac)
- xh(1) = .5*x(1)
- do 105 i=3,n,2
- xh(i-1) = -x(i)
- xh(i) = xh(i-2)+x(i-1)
- 105 continue
- if (modn .ne. 0) go to 106
- xh(n) = -x(n+1)
- 106 do 107 i=1,n
- x(i) = war(i)
- war(i) = xh(i)
- 107 continue
- return
- end
-
-c---------------------------SINGLE PRECISION----------------------------
- subroutine frffti (n,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- if (n .eq. 1) return
- call frffti1 (n,wsave(n+1),wsave(2*n+1))
- return
- end
-
- subroutine frffti1 (n,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wa(*) ,xxifac(*) ,ntryh(4)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- data ntryh(1),ntryh(2),ntryh(3),ntryh(4)/4,2,3,5/
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nl = n
- nf = 0
- j = 0
- 101 j = j+1
- if (j-4) 102,102,103
- 102 ntry = ntryh(j)
- go to 104
- 103 ntry = ntry+2
- 104 nq = nl/ntry
- nr = nl-ntry*nq
- if (nr) 101,105,101
- 105 nf = nf+1
- jfac(nf+2) = ntry
- nl = nq
- if (ntry .ne. 2) go to 107
- if (nf .eq. 1) go to 107
- do 106 i=2,nf
- ib = nf-i+2
- jfac(ib+2) = jfac(ib+1)
- 106 continue
- jfac(3) = 2
- 107 if (nl .ne. 1) go to 104
- jfac(1) = n
- jfac(2) = nf
- tpi = 6.28318530717959
- argh = tpi/float(n)
- is = 0
- nfm1 = nf-1
- l1 = 1
- if (nfm1 .eq. 0) return
- do 110 k1=1,nfm1
- ip = jfac(k1+2)
- ld = 0
- l2 = l1*ip
- ido = n/l2
- ipm = ip-1
- do 109 j=1,ipm
- ld = ld+l1
- i = is
- argld = float(ld)*argh
- fi = 0.
- do 108 ii=3,ido,2
- i = i+2
- fi = fi+1.
- arg = fi*argld
- wa(i-1) = cos(arg)
- wa(i) = sin(arg)
- 108 continue
- is = is+ido
- 109 continue
- l1 = l2
- 110 continue
- do 111 i = 1, 15
- xxifac(i) = rfac(i)
- 111 continue
- return
- end
-
-
-
- subroutine frfftf (n,r,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension r(*) ,wsave(*)
- if (n .eq. 1) return
- call frfftf1 (n,r,wsave,wsave(n+1),wsave(2*n+1))
- return
- end
- subroutine frfftf1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 1
- l2 = n
- iw = n
- do 111 k1=1,nf
- kh = nf-k1
- ip = jfac(kh+3)
- l1 = l2/ip
- ido = n/l2
- idl1 = ido*l1
- iw = iw-(ip-1)*ido
- na = 1-na
- if (ip .ne. 4) go to 102
- ix2 = iw+ido
- ix3 = ix2+ido
- if (na .ne. 0) go to 101
- call fradf4 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 110
- 101 call fradf4 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- go to 110
- 102 if (ip .ne. 2) go to 104
- if (na .ne. 0) go to 103
- call fradf2 (ido,l1,c,ch,wa(iw))
- go to 110
- 103 call fradf2 (ido,l1,ch,c,wa(iw))
- go to 110
- 104 if (ip .ne. 3) go to 106
- ix2 = iw+ido
- if (na .ne. 0) go to 105
- call fradf3 (ido,l1,c,ch,wa(iw),wa(ix2))
- go to 110
- 105 call fradf3 (ido,l1,ch,c,wa(iw),wa(ix2))
- go to 110
- 106 if (ip .ne. 5) go to 108
- ix2 = iw+ido
- ix3 = ix2+ido
- ix4 = ix3+ido
- if (na .ne. 0) go to 107
- call fradf5 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 110
- 107 call fradf5 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 110
- 108 if (ido .eq. 1) na = 1-na
- if (na .ne. 0) go to 109
- call fradfg (ido,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- na = 1
- go to 110
- 109 call fradfg (ido,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- na = 0
- 110 l2 = l1
- 111 continue
- if (na .eq. 1) return
- do 112 i=1,n
- c(i) = ch(i)
- 112 continue
- return
- end
- subroutine fradfg (ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,c2(idl1,ip),
- 2 ch2(idl1,ip) ,wa(*)
- data tpi/6.28318530717959/
- arg = tpi/float(ip)
- dcp = cos(arg)
- dsp = sin(arg)
- ipph = (ip+1)/2
- ipp2 = ip+2
- idp2 = ido+2
- nbd = (ido-1)/2
- if (ido .eq. 1) go to 119
- do 101 ik=1,idl1
- ch2(ik,1) = c2(ik,1)
- 101 continue
- do 103 j=2,ip
- do 102 k=1,l1
- ch(1,k,j) = c1(1,k,j)
- 102 continue
- 103 continue
- if (nbd .gt. l1) go to 107
- is = -ido
- do 106 j=2,ip
- is = is+ido
- idij = is
- do 105 i=3,ido,2
- idij = idij+2
- do 104 k=1,l1
- ch(i-1,k,j) = wa(idij-1)*c1(i-1,k,j)+wa(idij)*c1(i,k,j)
- ch(i,k,j) = wa(idij-1)*c1(i,k,j)-wa(idij)*c1(i-1,k,j)
- 104 continue
- 105 continue
- 106 continue
- go to 111
- 107 is = -ido
- do 110 j=2,ip
- is = is+ido
- do 109 k=1,l1
- idij = is
- do 108 i=3,ido,2
- idij = idij+2
- ch(i-1,k,j) = wa(idij-1)*c1(i-1,k,j)+wa(idij)*c1(i,k,j)
- ch(i,k,j) = wa(idij-1)*c1(i,k,j)-wa(idij)*c1(i-1,k,j)
- 108 continue
- 109 continue
- 110 continue
- 111 if (nbd .lt. l1) go to 115
- do 114 j=2,ipph
- jc = ipp2-j
- do 113 k=1,l1
- do 112 i=3,ido,2
- c1(i-1,k,j) = ch(i-1,k,j)+ch(i-1,k,jc)
- c1(i-1,k,jc) = ch(i,k,j)-ch(i,k,jc)
- c1(i,k,j) = ch(i,k,j)+ch(i,k,jc)
- c1(i,k,jc) = ch(i-1,k,jc)-ch(i-1,k,j)
- 112 continue
- 113 continue
- 114 continue
- go to 121
- 115 do 118 j=2,ipph
- jc = ipp2-j
- do 117 i=3,ido,2
- do 116 k=1,l1
- c1(i-1,k,j) = ch(i-1,k,j)+ch(i-1,k,jc)
- c1(i-1,k,jc) = ch(i,k,j)-ch(i,k,jc)
- c1(i,k,j) = ch(i,k,j)+ch(i,k,jc)
- c1(i,k,jc) = ch(i-1,k,jc)-ch(i-1,k,j)
- 116 continue
- 117 continue
- 118 continue
- go to 121
- 119 do 120 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 120 continue
- 121 do 123 j=2,ipph
- jc = ipp2-j
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)+ch(1,k,jc)
- c1(1,k,jc) = ch(1,k,jc)-ch(1,k,j)
- 122 continue
- 123 continue
-c
- ar1 = 1.
- ai1 = 0.
- do 127 l=2,ipph
- lc = ipp2-l
- ar1h = dcp*ar1-dsp*ai1
- ai1 = dcp*ai1+dsp*ar1
- ar1 = ar1h
- do 124 ik=1,idl1
- ch2(ik,l) = c2(ik,1)+ar1*c2(ik,2)
- ch2(ik,lc) = ai1*c2(ik,ip)
- 124 continue
- dc2 = ar1
- ds2 = ai1
- ar2 = ar1
- ai2 = ai1
- do 126 j=3,ipph
- jc = ipp2-j
- ar2h = dc2*ar2-ds2*ai2
- ai2 = dc2*ai2+ds2*ar2
- ar2 = ar2h
- do 125 ik=1,idl1
- ch2(ik,l) = ch2(ik,l)+ar2*c2(ik,j)
- ch2(ik,lc) = ch2(ik,lc)+ai2*c2(ik,jc)
- 125 continue
- 126 continue
- 127 continue
- do 129 j=2,ipph
- do 128 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+c2(ik,j)
- 128 continue
- 129 continue
-c
- if (ido .lt. l1) go to 132
- do 131 k=1,l1
- do 130 i=1,ido
- cc(i,1,k) = ch(i,k,1)
- 130 continue
- 131 continue
- go to 135
- 132 do 134 i=1,ido
- do 133 k=1,l1
- cc(i,1,k) = ch(i,k,1)
- 133 continue
- 134 continue
- 135 do 137 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 136 k=1,l1
- cc(ido,j2-2,k) = ch(1,k,j)
- cc(1,j2-1,k) = ch(1,k,jc)
- 136 continue
- 137 continue
- if (ido .eq. 1) return
- if (nbd .lt. l1) go to 141
- do 140 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 139 k=1,l1
- do 138 i=3,ido,2
- ic = idp2-i
- cc(i-1,j2-1,k) = ch(i-1,k,j)+ch(i-1,k,jc)
- cc(ic-1,j2-2,k) = ch(i-1,k,j)-ch(i-1,k,jc)
- cc(i,j2-1,k) = ch(i,k,j)+ch(i,k,jc)
- cc(ic,j2-2,k) = ch(i,k,jc)-ch(i,k,j)
- 138 continue
- 139 continue
- 140 continue
- return
- 141 do 144 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 143 i=3,ido,2
- ic = idp2-i
- do 142 k=1,l1
- cc(i-1,j2-1,k) = ch(i-1,k,j)+ch(i-1,k,jc)
- cc(ic-1,j2-2,k) = ch(i-1,k,j)-ch(i-1,k,jc)
- cc(i,j2-1,k) = ch(i,k,j)+ch(i,k,jc)
- cc(ic,j2-2,k) = ch(i,k,jc)-ch(i,k,j)
- 142 continue
- 143 continue
- 144 continue
- return
- end
- subroutine fradf5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,l1,5) ,ch(ido,5,l1) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- do 101 k=1,l1
- cr2 = cc(1,k,5)+cc(1,k,2)
- ci5 = cc(1,k,5)-cc(1,k,2)
- cr3 = cc(1,k,4)+cc(1,k,3)
- ci4 = cc(1,k,4)-cc(1,k,3)
- ch(1,1,k) = cc(1,k,1)+cr2+cr3
- ch(ido,2,k) = cc(1,k,1)+tr11*cr2+tr12*cr3
- ch(1,3,k) = ti11*ci5+ti12*ci4
- ch(ido,4,k) = cc(1,k,1)+tr12*cr2+tr11*cr3
- ch(1,5,k) = ti12*ci5-ti11*ci4
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- dr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- di2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- dr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- di3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- dr4 = wa3(i-2)*cc(i-1,k,4)+wa3(i-1)*cc(i,k,4)
- di4 = wa3(i-2)*cc(i,k,4)-wa3(i-1)*cc(i-1,k,4)
- dr5 = wa4(i-2)*cc(i-1,k,5)+wa4(i-1)*cc(i,k,5)
- di5 = wa4(i-2)*cc(i,k,5)-wa4(i-1)*cc(i-1,k,5)
- cr2 = dr2+dr5
- ci5 = dr5-dr2
- cr5 = di2-di5
- ci2 = di2+di5
- cr3 = dr3+dr4
- ci4 = dr4-dr3
- cr4 = di3-di4
- ci3 = di3+di4
- ch(i-1,1,k) = cc(i-1,k,1)+cr2+cr3
- ch(i,1,k) = cc(i,k,1)+ci2+ci3
- tr2 = cc(i-1,k,1)+tr11*cr2+tr12*cr3
- ti2 = cc(i,k,1)+tr11*ci2+tr12*ci3
- tr3 = cc(i-1,k,1)+tr12*cr2+tr11*cr3
- ti3 = cc(i,k,1)+tr12*ci2+tr11*ci3
- tr5 = ti11*cr5+ti12*cr4
- ti5 = ti11*ci5+ti12*ci4
- tr4 = ti12*cr5-ti11*cr4
- ti4 = ti12*ci5-ti11*ci4
- ch(i-1,3,k) = tr2+tr5
- ch(ic-1,2,k) = tr2-tr5
- ch(i,3,k) = ti2+ti5
- ch(ic,2,k) = ti5-ti2
- ch(i-1,5,k) = tr3+tr4
- ch(ic-1,4,k) = tr3-tr4
- ch(i,5,k) = ti3+ti4
- ch(ic,4,k) = ti4-ti3
- 102 continue
- 103 continue
- return
- end
- subroutine fradf3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,3,l1) ,cc(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- do 101 k=1,l1
- cr2 = cc(1,k,2)+cc(1,k,3)
- ch(1,1,k) = cc(1,k,1)+cr2
- ch(1,3,k) = taui*(cc(1,k,3)-cc(1,k,2))
- ch(ido,2,k) = cc(1,k,1)+taur*cr2
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- dr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- di2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- dr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- di3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- cr2 = dr2+dr3
- ci2 = di2+di3
- ch(i-1,1,k) = cc(i-1,k,1)+cr2
- ch(i,1,k) = cc(i,k,1)+ci2
- tr2 = cc(i-1,k,1)+taur*cr2
- ti2 = cc(i,k,1)+taur*ci2
- tr3 = taui*(di2-di3)
- ti3 = taui*(dr3-dr2)
- ch(i-1,3,k) = tr2+tr3
- ch(ic-1,2,k) = tr2-tr3
- ch(i,3,k) = ti2+ti3
- ch(ic,2,k) = ti3-ti2
- 102 continue
- 103 continue
- return
- end
- subroutine fradf2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,2,l1) ,cc(ido,l1,2) ,
- 1 wa1(*)
- do 101 k=1,l1
- ch(1,1,k) = cc(1,k,1)+cc(1,k,2)
- ch(ido,2,k) = cc(1,k,1)-cc(1,k,2)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- tr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- ti2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- ch(i,1,k) = cc(i,k,1)+ti2
- ch(ic,2,k) = ti2-cc(i,k,1)
- ch(i-1,1,k) = cc(i-1,k,1)+tr2
- ch(ic-1,2,k) = cc(i-1,k,1)-tr2
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 do 106 k=1,l1
- ch(1,2,k) = -cc(ido,k,2)
- ch(ido,1,k) = cc(ido,k,1)
- 106 continue
- 107 return
- end
- subroutine fradf4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,l1,4) ,ch(ido,4,l1) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- data hsqt2 /.7071067811865475/
- do 101 k=1,l1
- tr1 = cc(1,k,2)+cc(1,k,4)
- tr2 = cc(1,k,1)+cc(1,k,3)
- ch(1,1,k) = tr1+tr2
- ch(ido,4,k) = tr2-tr1
- ch(ido,2,k) = cc(1,k,1)-cc(1,k,3)
- ch(1,3,k) = cc(1,k,4)-cc(1,k,2)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- cr2 = wa1(i-2)*cc(i-1,k,2)+wa1(i-1)*cc(i,k,2)
- ci2 = wa1(i-2)*cc(i,k,2)-wa1(i-1)*cc(i-1,k,2)
- cr3 = wa2(i-2)*cc(i-1,k,3)+wa2(i-1)*cc(i,k,3)
- ci3 = wa2(i-2)*cc(i,k,3)-wa2(i-1)*cc(i-1,k,3)
- cr4 = wa3(i-2)*cc(i-1,k,4)+wa3(i-1)*cc(i,k,4)
- ci4 = wa3(i-2)*cc(i,k,4)-wa3(i-1)*cc(i-1,k,4)
- tr1 = cr2+cr4
- tr4 = cr4-cr2
- ti1 = ci2+ci4
- ti4 = ci2-ci4
- ti2 = cc(i,k,1)+ci3
- ti3 = cc(i,k,1)-ci3
- tr2 = cc(i-1,k,1)+cr3
- tr3 = cc(i-1,k,1)-cr3
- ch(i-1,1,k) = tr1+tr2
- ch(ic-1,4,k) = tr2-tr1
- ch(i,1,k) = ti1+ti2
- ch(ic,4,k) = ti1-ti2
- ch(i-1,3,k) = ti4+tr3
- ch(ic-1,2,k) = tr3-ti4
- ch(i,3,k) = tr4+ti3
- ch(ic,2,k) = tr4-ti3
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 continue
- do 106 k=1,l1
- ti1 = -hsqt2*(cc(ido,k,2)+cc(ido,k,4))
- tr1 = hsqt2*(cc(ido,k,2)-cc(ido,k,4))
- ch(ido,1,k) = tr1+cc(ido,k,1)
- ch(ido,3,k) = cc(ido,k,1)-tr1
- ch(1,2,k) = ti1-cc(ido,k,3)
- ch(1,4,k) = ti1+cc(ido,k,3)
- 106 continue
- 107 return
- end
-
-
-
- subroutine frfftb (n,r,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension r(*) ,wsave(*)
- if (n .eq. 1) return
- call frfftb1 (n,r,wsave,wsave(n+1),wsave(2*n+1))
- return
- end
- subroutine frfftb1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idl1 = ido*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+ido
- ix3 = ix2+ido
- if (na .ne. 0) go to 101
- call fradb4 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call fradb4 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call fradb2 (ido,l1,c,ch,wa(iw))
- go to 105
- 104 call fradb2 (ido,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+ido
- if (na .ne. 0) go to 107
- call fradb3 (ido,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call fradb3 (ido,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+ido
- ix3 = ix2+ido
- ix4 = ix3+ido
- if (na .ne. 0) go to 110
- call fradb5 (ido,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call fradb5 (ido,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call fradbg (ido,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call fradbg (ido,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (ido .eq. 1) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*ido
- 116 continue
- if (na .eq. 0) return
- do 117 i=1,n
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine fradbg (ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,c2(idl1,ip),
- 2 ch2(idl1,ip) ,wa(*)
- data tpi/6.28318530717959/
- arg = tpi/float(ip)
- dcp = cos(arg)
- dsp = sin(arg)
- idp2 = ido+2
- nbd = (ido-1)/2
- ipp2 = ip+2
- ipph = (ip+1)/2
- if (ido .lt. l1) go to 103
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 101 continue
- 102 continue
- go to 106
- 103 do 105 i=1,ido
- do 104 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- 106 do 108 j=2,ipph
- jc = ipp2-j
- j2 = j+j
- do 107 k=1,l1
- ch(1,k,j) = cc(ido,j2-2,k)+cc(ido,j2-2,k)
- ch(1,k,jc) = cc(1,j2-1,k)+cc(1,j2-1,k)
- 107 continue
- 108 continue
- if (ido .eq. 1) go to 116
- if (nbd .lt. l1) go to 112
- do 111 j=2,ipph
- jc = ipp2-j
- do 110 k=1,l1
- do 109 i=3,ido,2
- ic = idp2-i
- ch(i-1,k,j) = cc(i-1,2*j-1,k)+cc(ic-1,2*j-2,k)
- ch(i-1,k,jc) = cc(i-1,2*j-1,k)-cc(ic-1,2*j-2,k)
- ch(i,k,j) = cc(i,2*j-1,k)-cc(ic,2*j-2,k)
- ch(i,k,jc) = cc(i,2*j-1,k)+cc(ic,2*j-2,k)
- 109 continue
- 110 continue
- 111 continue
- go to 116
- 112 do 115 j=2,ipph
- jc = ipp2-j
- do 114 i=3,ido,2
- ic = idp2-i
- do 113 k=1,l1
- ch(i-1,k,j) = cc(i-1,2*j-1,k)+cc(ic-1,2*j-2,k)
- ch(i-1,k,jc) = cc(i-1,2*j-1,k)-cc(ic-1,2*j-2,k)
- ch(i,k,j) = cc(i,2*j-1,k)-cc(ic,2*j-2,k)
- ch(i,k,jc) = cc(i,2*j-1,k)+cc(ic,2*j-2,k)
- 113 continue
- 114 continue
- 115 continue
- 116 ar1 = 1.
- ai1 = 0.
- do 120 l=2,ipph
- lc = ipp2-l
- ar1h = dcp*ar1-dsp*ai1
- ai1 = dcp*ai1+dsp*ar1
- ar1 = ar1h
- do 117 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+ar1*ch2(ik,2)
- c2(ik,lc) = ai1*ch2(ik,ip)
- 117 continue
- dc2 = ar1
- ds2 = ai1
- ar2 = ar1
- ai2 = ai1
- do 119 j=3,ipph
- jc = ipp2-j
- ar2h = dc2*ar2-ds2*ai2
- ai2 = dc2*ai2+ds2*ar2
- ar2 = ar2h
- do 118 ik=1,idl1
- c2(ik,l) = c2(ik,l)+ar2*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)+ai2*ch2(ik,jc)
- 118 continue
- 119 continue
- 120 continue
- do 122 j=2,ipph
- do 121 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 121 continue
- 122 continue
- do 124 j=2,ipph
- jc = ipp2-j
- do 123 k=1,l1
- ch(1,k,j) = c1(1,k,j)-c1(1,k,jc)
- ch(1,k,jc) = c1(1,k,j)+c1(1,k,jc)
- 123 continue
- 124 continue
- if (ido .eq. 1) go to 132
- if (nbd .lt. l1) go to 128
- do 127 j=2,ipph
- jc = ipp2-j
- do 126 k=1,l1
- do 125 i=3,ido,2
- ch(i-1,k,j) = c1(i-1,k,j)-c1(i,k,jc)
- ch(i-1,k,jc) = c1(i-1,k,j)+c1(i,k,jc)
- ch(i,k,j) = c1(i,k,j)+c1(i-1,k,jc)
- ch(i,k,jc) = c1(i,k,j)-c1(i-1,k,jc)
- 125 continue
- 126 continue
- 127 continue
- go to 132
- 128 do 131 j=2,ipph
- jc = ipp2-j
- do 130 i=3,ido,2
- do 129 k=1,l1
- ch(i-1,k,j) = c1(i-1,k,j)-c1(i,k,jc)
- ch(i-1,k,jc) = c1(i-1,k,j)+c1(i,k,jc)
- ch(i,k,j) = c1(i,k,j)+c1(i-1,k,jc)
- ch(i,k,jc) = c1(i,k,j)-c1(i-1,k,jc)
- 129 continue
- 130 continue
- 131 continue
- 132 continue
- if (ido .eq. 1) return
- do 133 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 133 continue
- do 135 j=2,ip
- do 134 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- 134 continue
- 135 continue
- if (nbd .gt. l1) go to 139
- is = -ido
- do 138 j=2,ip
- is = is+ido
- idij = is
- do 137 i=3,ido,2
- idij = idij+2
- do 136 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 136 continue
- 137 continue
- 138 continue
- go to 143
- 139 is = -ido
- do 142 j=2,ip
- is = is+ido
- do 141 k=1,l1
- idij = is
- do 140 i=3,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 140 continue
- 141 continue
- 142 continue
- 143 return
- end
- subroutine fradb5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- do 101 k=1,l1
- ti5 = cc(1,3,k)+cc(1,3,k)
- ti4 = cc(1,5,k)+cc(1,5,k)
- tr2 = cc(ido,2,k)+cc(ido,2,k)
- tr3 = cc(ido,4,k)+cc(ido,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci5 = ti11*ti5+ti12*ti4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(1,k,5) = cr2+ci5
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- ti5 = cc(i,3,k)+cc(ic,2,k)
- ti2 = cc(i,3,k)-cc(ic,2,k)
- ti4 = cc(i,5,k)+cc(ic,4,k)
- ti3 = cc(i,5,k)-cc(ic,4,k)
- tr5 = cc(i-1,3,k)-cc(ic-1,2,k)
- tr2 = cc(i-1,3,k)+cc(ic-1,2,k)
- tr4 = cc(i-1,5,k)-cc(ic-1,4,k)
- tr3 = cc(i-1,5,k)+cc(ic-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-2)*dr2-wa1(i-1)*di2
- ch(i,k,2) = wa1(i-2)*di2+wa1(i-1)*dr2
- ch(i-1,k,3) = wa2(i-2)*dr3-wa2(i-1)*di3
- ch(i,k,3) = wa2(i-2)*di3+wa2(i-1)*dr3
- ch(i-1,k,4) = wa3(i-2)*dr4-wa3(i-1)*di4
- ch(i,k,4) = wa3(i-2)*di4+wa3(i-1)*dr4
- ch(i-1,k,5) = wa4(i-2)*dr5-wa4(i-1)*di5
- ch(i,k,5) = wa4(i-2)*di5+wa4(i-1)*dr5
- 102 continue
- 103 continue
- return
- end
- subroutine fradb3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- do 101 k=1,l1
- tr2 = cc(ido,2,k)+cc(ido,2,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ci3 = taui*(cc(1,3,k)+cc(1,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- 101 continue
- if (ido .eq. 1) return
- idp2 = ido+2
- do 103 k=1,l1
- do 102 i=3,ido,2
- ic = idp2-i
- tr2 = cc(i-1,3,k)+cc(ic-1,2,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,3,k)-cc(ic,2,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,3,k)-cc(ic-1,2,k))
- ci3 = taui*(cc(i,3,k)+cc(ic,2,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i-1,k,2) = wa1(i-2)*dr2-wa1(i-1)*di2
- ch(i,k,2) = wa1(i-2)*di2+wa1(i-1)*dr2
- ch(i-1,k,3) = wa2(i-2)*dr3-wa2(i-1)*di3
- ch(i,k,3) = wa2(i-2)*di3+wa2(i-1)*dr3
- 102 continue
- 103 continue
- return
- end
- subroutine fradb2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(ido,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(ido,2,k)
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- ch(i-1,k,1) = cc(i-1,1,k)+cc(ic-1,2,k)
- tr2 = cc(i-1,1,k)-cc(ic-1,2,k)
- ch(i,k,1) = cc(i,1,k)-cc(ic,2,k)
- ti2 = cc(i,1,k)+cc(ic,2,k)
- ch(i-1,k,2) = wa1(i-2)*tr2-wa1(i-1)*ti2
- ch(i,k,2) = wa1(i-2)*ti2+wa1(i-1)*tr2
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 do 106 k=1,l1
- ch(ido,k,1) = cc(ido,1,k)+cc(ido,1,k)
- ch(ido,k,2) = -(cc(1,2,k)+cc(1,2,k))
- 106 continue
- 107 return
- end
- subroutine fradb4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- data sqrt2 /1.414213562373095/
- do 101 k=1,l1
- tr1 = cc(1,1,k)-cc(ido,4,k)
- tr2 = cc(1,1,k)+cc(ido,4,k)
- tr3 = cc(ido,2,k)+cc(ido,2,k)
- tr4 = cc(1,3,k)+cc(1,3,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,2) = tr1-tr4
- ch(1,k,3) = tr2-tr3
- ch(1,k,4) = tr1+tr4
- 101 continue
- if (ido-2) 107,105,102
- 102 idp2 = ido+2
- do 104 k=1,l1
- do 103 i=3,ido,2
- ic = idp2-i
- ti1 = cc(i,1,k)+cc(ic,4,k)
- ti2 = cc(i,1,k)-cc(ic,4,k)
- ti3 = cc(i,3,k)-cc(ic,2,k)
- tr4 = cc(i,3,k)+cc(ic,2,k)
- tr1 = cc(i-1,1,k)-cc(ic-1,4,k)
- tr2 = cc(i-1,1,k)+cc(ic-1,4,k)
- ti4 = cc(i-1,3,k)-cc(ic-1,2,k)
- tr3 = cc(i-1,3,k)+cc(ic-1,2,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1-tr4
- cr4 = tr1+tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-2)*cr2-wa1(i-1)*ci2
- ch(i,k,2) = wa1(i-2)*ci2+wa1(i-1)*cr2
- ch(i-1,k,3) = wa2(i-2)*cr3-wa2(i-1)*ci3
- ch(i,k,3) = wa2(i-2)*ci3+wa2(i-1)*cr3
- ch(i-1,k,4) = wa3(i-2)*cr4-wa3(i-1)*ci4
- ch(i,k,4) = wa3(i-2)*ci4+wa3(i-1)*cr4
- 103 continue
- 104 continue
- if (mod(ido,2) .eq. 1) return
- 105 continue
- do 106 k=1,l1
- ti1 = cc(1,2,k)+cc(1,4,k)
- ti2 = cc(1,4,k)-cc(1,2,k)
- tr1 = cc(ido,1,k)-cc(ido,3,k)
- tr2 = cc(ido,1,k)+cc(ido,3,k)
- ch(ido,k,1) = tr2+tr2
- ch(ido,k,2) = sqrt2*(tr1-ti1)
- ch(ido,k,3) = ti2+ti2
- ch(ido,k,4) = -sqrt2*(tr1+ti1)
- 106 continue
- 107 return
- end
-
-
-
- subroutine fcffti (n,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call fcffti1 (n,wsave(iw1),wsave(iw2))
- return
- end
- subroutine fcffti1 (n,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wa(*) ,xxifac(*) ,ntryh(4)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- data ntryh(1),ntryh(2),ntryh(3),ntryh(4)/3,4,2,5/
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nl = n
- nf = 0
- j = 0
- 101 j = j+1
- if (j-4) 102,102,103
- 102 ntry = ntryh(j)
- go to 104
- 103 ntry = ntry+2
- 104 nq = nl/ntry
- nr = nl-ntry*nq
- if (nr) 101,105,101
- 105 nf = nf+1
- jfac(nf+2) = ntry
- nl = nq
- if (ntry .ne. 2) go to 107
- if (nf .eq. 1) go to 107
- do 106 i=2,nf
- ib = nf-i+2
- jfac(ib+2) = jfac(ib+1)
- 106 continue
- jfac(3) = 2
- 107 if (nl .ne. 1) go to 104
- jfac(1) = n
- jfac(2) = nf
- tpi = 6.28318530717959
- argh = tpi/float(n)
- i = 2
- l1 = 1
- do 110 k1=1,nf
- ip = jfac(k1+2)
- ld = 0
- l2 = l1*ip
- ido = n/l2
- idot = ido+ido+2
- ipm = ip-1
- do 109 j=1,ipm
- i1 = i
- wa(i-1) = 1.
- wa(i) = 0.
- ld = ld+l1
- fi = 0.
- argld = float(ld)*argh
- do 108 ii=4,idot,2
- i = i+2
- fi = fi+1.
- arg = fi*argld
- wa(i-1) = cos(arg)
- wa(i) = sin(arg)
- 108 continue
- if (ip .le. 5) go to 109
- wa(i1-1) = wa(i-1)
- wa(i1) = wa(i)
- 109 continue
- l1 = l2
- 110 continue
- do 111 i = 1, 15
- xxifac(i) = rfac(i)
- 111 continue
- return
- end
-
-
-
- subroutine fcfftf (n,c,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension c(*) ,wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call fcfftf1 (n,c,wsave,wsave(iw1),wsave(iw2))
- return
- end
- subroutine fcfftf1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idot = ido+ido
- idl1 = idot*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+idot
- ix3 = ix2+idot
- if (na .ne. 0) go to 101
- call fpassf4 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call fpassf4 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call fpassf2 (idot,l1,c,ch,wa(iw))
- go to 105
- 104 call fpassf2 (idot,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+idot
- if (na .ne. 0) go to 107
- call fpassf3 (idot,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call fpassf3 (idot,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+idot
- ix3 = ix2+idot
- ix4 = ix3+idot
- if (na .ne. 0) go to 110
- call fpassf5 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call fpassf5 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call fpassf (nac,idot,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call fpassf (nac,idot,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (nac .ne. 0) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*idot
- 116 continue
- if (na .eq. 0) return
- n2 = n+n
- do 117 i=1,n2
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine fpassf (nac,ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,wa(*) ,c2(idl1,ip),
- 2 ch2(idl1,ip)
- idot = ido/2
- nt = ip*idl1
- ipp2 = ip+2
- ipph = (ip+1)/2
- idp = ip*ido
-c
- if (ido .lt. l1) go to 106
- do 103 j=2,ipph
- jc = ipp2-j
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 101 continue
- 102 continue
- 103 continue
- do 105 k=1,l1
- do 104 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- go to 112
- 106 do 109 j=2,ipph
- jc = ipp2-j
- do 108 i=1,ido
- do 107 k=1,l1
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 107 continue
- 108 continue
- 109 continue
- do 111 i=1,ido
- do 110 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 110 continue
- 111 continue
- 112 idl = 2-ido
- inc = 0
- do 116 l=2,ipph
- lc = ipp2-l
- idl = idl+ido
- do 113 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+wa(idl-1)*ch2(ik,2)
- c2(ik,lc) = -wa(idl)*ch2(ik,ip)
- 113 continue
- idlj = idl
- inc = inc+ido
- do 115 j=3,ipph
- jc = ipp2-j
- idlj = idlj+inc
- if (idlj .gt. idp) idlj = idlj-idp
- war = wa(idlj-1)
- wai = wa(idlj)
- do 114 ik=1,idl1
- c2(ik,l) = c2(ik,l)+war*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)-wai*ch2(ik,jc)
- 114 continue
- 115 continue
- 116 continue
- do 118 j=2,ipph
- do 117 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 117 continue
- 118 continue
- do 120 j=2,ipph
- jc = ipp2-j
- do 119 ik=2,idl1,2
- ch2(ik-1,j) = c2(ik-1,j)-c2(ik,jc)
- ch2(ik-1,jc) = c2(ik-1,j)+c2(ik,jc)
- ch2(ik,j) = c2(ik,j)+c2(ik-1,jc)
- ch2(ik,jc) = c2(ik,j)-c2(ik-1,jc)
- 119 continue
- 120 continue
- nac = 1
- if (ido .eq. 2) return
- nac = 0
- do 121 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 121 continue
- do 123 j=2,ip
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- c1(2,k,j) = ch(2,k,j)
- 122 continue
- 123 continue
- if (idot .gt. l1) go to 127
- idij = 0
- do 126 j=2,ip
- idij = idij+2
- do 125 i=4,ido,2
- idij = idij+2
- do 124 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)+wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)-wa(idij)*ch(i-1,k,j)
- 124 continue
- 125 continue
- 126 continue
- return
- 127 idj = 2-ido
- do 130 j=2,ip
- idj = idj+ido
- do 129 k=1,l1
- idij = idj
- do 128 i=4,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)+wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)-wa(idij)*ch(i-1,k,j)
- 128 continue
- 129 continue
- 130 continue
- return
- end
- subroutine fpassf5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,-.951056516295154,
- 1-.809016994374947,-.587785252292473/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti5 = cc(2,2,k)-cc(2,5,k)
- ti2 = cc(2,2,k)+cc(2,5,k)
- ti4 = cc(2,3,k)-cc(2,4,k)
- ti3 = cc(2,3,k)+cc(2,4,k)
- tr5 = cc(1,2,k)-cc(1,5,k)
- tr2 = cc(1,2,k)+cc(1,5,k)
- tr4 = cc(1,3,k)-cc(1,4,k)
- tr3 = cc(1,3,k)+cc(1,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- ch(2,k,1) = cc(2,1,k)+ti2+ti3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(2,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(2,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,5) = cr2+ci5
- ch(2,k,2) = ci2+cr5
- ch(2,k,3) = ci3+cr4
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(2,k,4) = ci3-cr4
- ch(2,k,5) = ci2-cr5
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti5 = cc(i,2,k)-cc(i,5,k)
- ti2 = cc(i,2,k)+cc(i,5,k)
- ti4 = cc(i,3,k)-cc(i,4,k)
- ti3 = cc(i,3,k)+cc(i,4,k)
- tr5 = cc(i-1,2,k)-cc(i-1,5,k)
- tr2 = cc(i-1,2,k)+cc(i-1,5,k)
- tr4 = cc(i-1,3,k)-cc(i-1,4,k)
- tr3 = cc(i-1,3,k)+cc(i-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-1)*dr2+wa1(i)*di2
- ch(i,k,2) = wa1(i-1)*di2-wa1(i)*dr2
- ch(i-1,k,3) = wa2(i-1)*dr3+wa2(i)*di3
- ch(i,k,3) = wa2(i-1)*di3-wa2(i)*dr3
- ch(i-1,k,4) = wa3(i-1)*dr4+wa3(i)*di4
- ch(i,k,4) = wa3(i-1)*di4-wa3(i)*dr4
- ch(i-1,k,5) = wa4(i-1)*dr5+wa4(i)*di5
- ch(i,k,5) = wa4(i-1)*di5-wa4(i)*dr5
- 103 continue
- 104 continue
- return
- end
- subroutine fpassf3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,-.866025403784439/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- tr2 = cc(1,2,k)+cc(1,3,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ti2 = cc(2,2,k)+cc(2,3,k)
- ci2 = cc(2,1,k)+taur*ti2
- ch(2,k,1) = cc(2,1,k)+ti2
- cr3 = taui*(cc(1,2,k)-cc(1,3,k))
- ci3 = taui*(cc(2,2,k)-cc(2,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- ch(2,k,2) = ci2+cr3
- ch(2,k,3) = ci2-cr3
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- tr2 = cc(i-1,2,k)+cc(i-1,3,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,2,k)+cc(i,3,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,2,k)-cc(i-1,3,k))
- ci3 = taui*(cc(i,2,k)-cc(i,3,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i,k,2) = wa1(i-1)*di2-wa1(i)*dr2
- ch(i-1,k,2) = wa1(i-1)*dr2+wa1(i)*di2
- ch(i,k,3) = wa2(i-1)*di3-wa2(i)*dr3
- ch(i-1,k,3) = wa2(i-1)*dr3+wa2(i)*di3
- 103 continue
- 104 continue
- return
- end
- subroutine fpassf2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- if (ido .gt. 2) go to 102
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(1,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(1,2,k)
- ch(2,k,1) = cc(2,1,k)+cc(2,2,k)
- ch(2,k,2) = cc(2,1,k)-cc(2,2,k)
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ch(i-1,k,1) = cc(i-1,1,k)+cc(i-1,2,k)
- tr2 = cc(i-1,1,k)-cc(i-1,2,k)
- ch(i,k,1) = cc(i,1,k)+cc(i,2,k)
- ti2 = cc(i,1,k)-cc(i,2,k)
- ch(i,k,2) = wa1(i-1)*ti2-wa1(i)*tr2
- ch(i-1,k,2) = wa1(i-1)*tr2+wa1(i)*ti2
- 103 continue
- 104 continue
- return
- end
- subroutine fpassf4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti1 = cc(2,1,k)-cc(2,3,k)
- ti2 = cc(2,1,k)+cc(2,3,k)
- tr4 = cc(2,2,k)-cc(2,4,k)
- ti3 = cc(2,2,k)+cc(2,4,k)
- tr1 = cc(1,1,k)-cc(1,3,k)
- tr2 = cc(1,1,k)+cc(1,3,k)
- ti4 = cc(1,4,k)-cc(1,2,k)
- tr3 = cc(1,2,k)+cc(1,4,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,3) = tr2-tr3
- ch(2,k,1) = ti2+ti3
- ch(2,k,3) = ti2-ti3
- ch(1,k,2) = tr1+tr4
- ch(1,k,4) = tr1-tr4
- ch(2,k,2) = ti1+ti4
- ch(2,k,4) = ti1-ti4
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti1 = cc(i,1,k)-cc(i,3,k)
- ti2 = cc(i,1,k)+cc(i,3,k)
- ti3 = cc(i,2,k)+cc(i,4,k)
- tr4 = cc(i,2,k)-cc(i,4,k)
- tr1 = cc(i-1,1,k)-cc(i-1,3,k)
- tr2 = cc(i-1,1,k)+cc(i-1,3,k)
- ti4 = cc(i-1,4,k)-cc(i-1,2,k)
- tr3 = cc(i-1,2,k)+cc(i-1,4,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1+tr4
- cr4 = tr1-tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-1)*cr2+wa1(i)*ci2
- ch(i,k,2) = wa1(i-1)*ci2-wa1(i)*cr2
- ch(i-1,k,3) = wa2(i-1)*cr3+wa2(i)*ci3
- ch(i,k,3) = wa2(i-1)*ci3-wa2(i)*cr3
- ch(i-1,k,4) = wa3(i-1)*cr4+wa3(i)*ci4
- ch(i,k,4) = wa3(i-1)*ci4-wa3(i)*cr4
- 103 continue
- 104 continue
- return
- end
-
-
-
- subroutine fcfftb (n,c,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension c(*) ,wsave(*)
- if (n .eq. 1) return
- iw1 = n+n+1
- iw2 = iw1+n+n
- call fcfftb1 (n,c,wsave,wsave(iw1),wsave(iw2))
- return
- end
- subroutine fcfftb1 (n,c,ch,wa,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(*) ,c(*) ,wa(*) ,xxifac(*)
- dimension rfac(15), jfac(15)
- equivalence (rfac(1), jfac(1))
- do 100 i = 1, 15
- rfac(i) = xxifac(i)
- 100 continue
- nf = jfac(2)
- na = 0
- l1 = 1
- iw = 1
- do 116 k1=1,nf
- ip = jfac(k1+2)
- l2 = ip*l1
- ido = n/l2
- idot = ido+ido
- idl1 = idot*l1
- if (ip .ne. 4) go to 103
- ix2 = iw+idot
- ix3 = ix2+idot
- if (na .ne. 0) go to 101
- call fpassb4 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3))
- go to 102
- 101 call fpassb4 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3))
- 102 na = 1-na
- go to 115
- 103 if (ip .ne. 2) go to 106
- if (na .ne. 0) go to 104
- call fpassb2 (idot,l1,c,ch,wa(iw))
- go to 105
- 104 call fpassb2 (idot,l1,ch,c,wa(iw))
- 105 na = 1-na
- go to 115
- 106 if (ip .ne. 3) go to 109
- ix2 = iw+idot
- if (na .ne. 0) go to 107
- call fpassb3 (idot,l1,c,ch,wa(iw),wa(ix2))
- go to 108
- 107 call fpassb3 (idot,l1,ch,c,wa(iw),wa(ix2))
- 108 na = 1-na
- go to 115
- 109 if (ip .ne. 5) go to 112
- ix2 = iw+idot
- ix3 = ix2+idot
- ix4 = ix3+idot
- if (na .ne. 0) go to 110
- call fpassb5 (idot,l1,c,ch,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- go to 111
- 110 call fpassb5 (idot,l1,ch,c,wa(iw),wa(ix2),wa(ix3),wa(ix4))
- 111 na = 1-na
- go to 115
- 112 if (na .ne. 0) go to 113
- call fpassb (nac,idot,ip,l1,idl1,c,c,c,ch,ch,wa(iw))
- go to 114
- 113 call fpassb (nac,idot,ip,l1,idl1,ch,ch,ch,c,c,wa(iw))
- 114 if (nac .ne. 0) na = 1-na
- 115 l1 = l2
- iw = iw+(ip-1)*idot
- 116 continue
- if (na .eq. 0) return
- n2 = n+n
- do 117 i=1,n2
- c(i) = ch(i)
- 117 continue
- return
- end
- subroutine fpassb (nac,ido,ip,l1,idl1,cc,c1,c2,ch,ch2,wa)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension ch(ido,l1,ip) ,cc(ido,ip,l1) ,
- 1 c1(ido,l1,ip) ,wa(*) ,c2(idl1,ip),
- 2 ch2(idl1,ip)
- idot = ido/2
- nt = ip*idl1
- ipp2 = ip+2
- ipph = (ip+1)/2
- idp = ip*ido
-c
- if (ido .lt. l1) go to 106
- do 103 j=2,ipph
- jc = ipp2-j
- do 102 k=1,l1
- do 101 i=1,ido
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 101 continue
- 102 continue
- 103 continue
- do 105 k=1,l1
- do 104 i=1,ido
- ch(i,k,1) = cc(i,1,k)
- 104 continue
- 105 continue
- go to 112
- 106 do 109 j=2,ipph
- jc = ipp2-j
- do 108 i=1,ido
- do 107 k=1,l1
- ch(i,k,j) = cc(i,j,k)+cc(i,jc,k)
- ch(i,k,jc) = cc(i,j,k)-cc(i,jc,k)
- 107 continue
- 108 continue
- 109 continue
- do 111 i=1,ido
- do 110 k=1,l1
- ch(i,k,1) = cc(i,1,k)
- 110 continue
- 111 continue
- 112 idl = 2-ido
- inc = 0
- do 116 l=2,ipph
- lc = ipp2-l
- idl = idl+ido
- do 113 ik=1,idl1
- c2(ik,l) = ch2(ik,1)+wa(idl-1)*ch2(ik,2)
- c2(ik,lc) = wa(idl)*ch2(ik,ip)
- 113 continue
- idlj = idl
- inc = inc+ido
- do 115 j=3,ipph
- jc = ipp2-j
- idlj = idlj+inc
- if (idlj .gt. idp) idlj = idlj-idp
- war = wa(idlj-1)
- wai = wa(idlj)
- do 114 ik=1,idl1
- c2(ik,l) = c2(ik,l)+war*ch2(ik,j)
- c2(ik,lc) = c2(ik,lc)+wai*ch2(ik,jc)
- 114 continue
- 115 continue
- 116 continue
- do 118 j=2,ipph
- do 117 ik=1,idl1
- ch2(ik,1) = ch2(ik,1)+ch2(ik,j)
- 117 continue
- 118 continue
- do 120 j=2,ipph
- jc = ipp2-j
- do 119 ik=2,idl1,2
- ch2(ik-1,j) = c2(ik-1,j)-c2(ik,jc)
- ch2(ik-1,jc) = c2(ik-1,j)+c2(ik,jc)
- ch2(ik,j) = c2(ik,j)+c2(ik-1,jc)
- ch2(ik,jc) = c2(ik,j)-c2(ik-1,jc)
- 119 continue
- 120 continue
- nac = 1
- if (ido .eq. 2) return
- nac = 0
- do 121 ik=1,idl1
- c2(ik,1) = ch2(ik,1)
- 121 continue
- do 123 j=2,ip
- do 122 k=1,l1
- c1(1,k,j) = ch(1,k,j)
- c1(2,k,j) = ch(2,k,j)
- 122 continue
- 123 continue
- if (idot .gt. l1) go to 127
- idij = 0
- do 126 j=2,ip
- idij = idij+2
- do 125 i=4,ido,2
- idij = idij+2
- do 124 k=1,l1
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 124 continue
- 125 continue
- 126 continue
- return
- 127 idj = 2-ido
- do 130 j=2,ip
- idj = idj+ido
- do 129 k=1,l1
- idij = idj
- do 128 i=4,ido,2
- idij = idij+2
- c1(i-1,k,j) = wa(idij-1)*ch(i-1,k,j)-wa(idij)*ch(i,k,j)
- c1(i,k,j) = wa(idij-1)*ch(i,k,j)+wa(idij)*ch(i-1,k,j)
- 128 continue
- 129 continue
- 130 continue
- return
- end
- subroutine fpassb5 (ido,l1,cc,ch,wa1,wa2,wa3,wa4)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,5,l1) ,ch(ido,l1,5) ,
- 1 wa1(*) ,wa2(*) ,wa3(*) ,wa4(*)
- data tr11,ti11,tr12,ti12 /.309016994374947,.951056516295154,
- 1-.809016994374947,.587785252292473/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti5 = cc(2,2,k)-cc(2,5,k)
- ti2 = cc(2,2,k)+cc(2,5,k)
- ti4 = cc(2,3,k)-cc(2,4,k)
- ti3 = cc(2,3,k)+cc(2,4,k)
- tr5 = cc(1,2,k)-cc(1,5,k)
- tr2 = cc(1,2,k)+cc(1,5,k)
- tr4 = cc(1,3,k)-cc(1,4,k)
- tr3 = cc(1,3,k)+cc(1,4,k)
- ch(1,k,1) = cc(1,1,k)+tr2+tr3
- ch(2,k,1) = cc(2,1,k)+ti2+ti3
- cr2 = cc(1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(2,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(2,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- ch(1,k,2) = cr2-ci5
- ch(1,k,5) = cr2+ci5
- ch(2,k,2) = ci2+cr5
- ch(2,k,3) = ci3+cr4
- ch(1,k,3) = cr3-ci4
- ch(1,k,4) = cr3+ci4
- ch(2,k,4) = ci3-cr4
- ch(2,k,5) = ci2-cr5
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti5 = cc(i,2,k)-cc(i,5,k)
- ti2 = cc(i,2,k)+cc(i,5,k)
- ti4 = cc(i,3,k)-cc(i,4,k)
- ti3 = cc(i,3,k)+cc(i,4,k)
- tr5 = cc(i-1,2,k)-cc(i-1,5,k)
- tr2 = cc(i-1,2,k)+cc(i-1,5,k)
- tr4 = cc(i-1,3,k)-cc(i-1,4,k)
- tr3 = cc(i-1,3,k)+cc(i-1,4,k)
- ch(i-1,k,1) = cc(i-1,1,k)+tr2+tr3
- ch(i,k,1) = cc(i,1,k)+ti2+ti3
- cr2 = cc(i-1,1,k)+tr11*tr2+tr12*tr3
- ci2 = cc(i,1,k)+tr11*ti2+tr12*ti3
- cr3 = cc(i-1,1,k)+tr12*tr2+tr11*tr3
- ci3 = cc(i,1,k)+tr12*ti2+tr11*ti3
- cr5 = ti11*tr5+ti12*tr4
- ci5 = ti11*ti5+ti12*ti4
- cr4 = ti12*tr5-ti11*tr4
- ci4 = ti12*ti5-ti11*ti4
- dr3 = cr3-ci4
- dr4 = cr3+ci4
- di3 = ci3+cr4
- di4 = ci3-cr4
- dr5 = cr2+ci5
- dr2 = cr2-ci5
- di5 = ci2-cr5
- di2 = ci2+cr5
- ch(i-1,k,2) = wa1(i-1)*dr2-wa1(i)*di2
- ch(i,k,2) = wa1(i-1)*di2+wa1(i)*dr2
- ch(i-1,k,3) = wa2(i-1)*dr3-wa2(i)*di3
- ch(i,k,3) = wa2(i-1)*di3+wa2(i)*dr3
- ch(i-1,k,4) = wa3(i-1)*dr4-wa3(i)*di4
- ch(i,k,4) = wa3(i-1)*di4+wa3(i)*dr4
- ch(i-1,k,5) = wa4(i-1)*dr5-wa4(i)*di5
- ch(i,k,5) = wa4(i-1)*di5+wa4(i)*dr5
- 103 continue
- 104 continue
- return
- end
- subroutine fpassb3 (ido,l1,cc,ch,wa1,wa2)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,3,l1) ,ch(ido,l1,3) ,
- 1 wa1(*) ,wa2(*)
- data taur,taui /-.5,.866025403784439/
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- tr2 = cc(1,2,k)+cc(1,3,k)
- cr2 = cc(1,1,k)+taur*tr2
- ch(1,k,1) = cc(1,1,k)+tr2
- ti2 = cc(2,2,k)+cc(2,3,k)
- ci2 = cc(2,1,k)+taur*ti2
- ch(2,k,1) = cc(2,1,k)+ti2
- cr3 = taui*(cc(1,2,k)-cc(1,3,k))
- ci3 = taui*(cc(2,2,k)-cc(2,3,k))
- ch(1,k,2) = cr2-ci3
- ch(1,k,3) = cr2+ci3
- ch(2,k,2) = ci2+cr3
- ch(2,k,3) = ci2-cr3
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- tr2 = cc(i-1,2,k)+cc(i-1,3,k)
- cr2 = cc(i-1,1,k)+taur*tr2
- ch(i-1,k,1) = cc(i-1,1,k)+tr2
- ti2 = cc(i,2,k)+cc(i,3,k)
- ci2 = cc(i,1,k)+taur*ti2
- ch(i,k,1) = cc(i,1,k)+ti2
- cr3 = taui*(cc(i-1,2,k)-cc(i-1,3,k))
- ci3 = taui*(cc(i,2,k)-cc(i,3,k))
- dr2 = cr2-ci3
- dr3 = cr2+ci3
- di2 = ci2+cr3
- di3 = ci2-cr3
- ch(i,k,2) = wa1(i-1)*di2+wa1(i)*dr2
- ch(i-1,k,2) = wa1(i-1)*dr2-wa1(i)*di2
- ch(i,k,3) = wa2(i-1)*di3+wa2(i)*dr3
- ch(i-1,k,3) = wa2(i-1)*dr3-wa2(i)*di3
- 103 continue
- 104 continue
- return
- end
- subroutine fpassb2 (ido,l1,cc,ch,wa1)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,2,l1) ,ch(ido,l1,2) ,
- 1 wa1(*)
- if (ido .gt. 2) go to 102
- do 101 k=1,l1
- ch(1,k,1) = cc(1,1,k)+cc(1,2,k)
- ch(1,k,2) = cc(1,1,k)-cc(1,2,k)
- ch(2,k,1) = cc(2,1,k)+cc(2,2,k)
- ch(2,k,2) = cc(2,1,k)-cc(2,2,k)
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ch(i-1,k,1) = cc(i-1,1,k)+cc(i-1,2,k)
- tr2 = cc(i-1,1,k)-cc(i-1,2,k)
- ch(i,k,1) = cc(i,1,k)+cc(i,2,k)
- ti2 = cc(i,1,k)-cc(i,2,k)
- ch(i,k,2) = wa1(i-1)*ti2+wa1(i)*tr2
- ch(i-1,k,2) = wa1(i-1)*tr2-wa1(i)*ti2
- 103 continue
- 104 continue
- return
- end
- subroutine fpassb4 (ido,l1,cc,ch,wa1,wa2,wa3)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension cc(ido,4,l1) ,ch(ido,l1,4) ,
- 1 wa1(*) ,wa2(*) ,wa3(*)
- if (ido .ne. 2) go to 102
- do 101 k=1,l1
- ti1 = cc(2,1,k)-cc(2,3,k)
- ti2 = cc(2,1,k)+cc(2,3,k)
- tr4 = cc(2,4,k)-cc(2,2,k)
- ti3 = cc(2,2,k)+cc(2,4,k)
- tr1 = cc(1,1,k)-cc(1,3,k)
- tr2 = cc(1,1,k)+cc(1,3,k)
- ti4 = cc(1,2,k)-cc(1,4,k)
- tr3 = cc(1,2,k)+cc(1,4,k)
- ch(1,k,1) = tr2+tr3
- ch(1,k,3) = tr2-tr3
- ch(2,k,1) = ti2+ti3
- ch(2,k,3) = ti2-ti3
- ch(1,k,2) = tr1+tr4
- ch(1,k,4) = tr1-tr4
- ch(2,k,2) = ti1+ti4
- ch(2,k,4) = ti1-ti4
- 101 continue
- return
- 102 do 104 k=1,l1
- do 103 i=2,ido,2
- ti1 = cc(i,1,k)-cc(i,3,k)
- ti2 = cc(i,1,k)+cc(i,3,k)
- ti3 = cc(i,2,k)+cc(i,4,k)
- tr4 = cc(i,4,k)-cc(i,2,k)
- tr1 = cc(i-1,1,k)-cc(i-1,3,k)
- tr2 = cc(i-1,1,k)+cc(i-1,3,k)
- ti4 = cc(i-1,2,k)-cc(i-1,4,k)
- tr3 = cc(i-1,2,k)+cc(i-1,4,k)
- ch(i-1,k,1) = tr2+tr3
- cr3 = tr2-tr3
- ch(i,k,1) = ti2+ti3
- ci3 = ti2-ti3
- cr2 = tr1+tr4
- cr4 = tr1-tr4
- ci2 = ti1+ti4
- ci4 = ti1-ti4
- ch(i-1,k,2) = wa1(i-1)*cr2-wa1(i)*ci2
- ch(i,k,2) = wa1(i-1)*ci2+wa1(i)*cr2
- ch(i-1,k,3) = wa2(i-1)*cr3-wa2(i)*ci3
- ch(i,k,3) = wa2(i-1)*ci3+wa2(i)*cr3
- ch(i-1,k,4) = wa3(i-1)*cr4-wa3(i)*ci4
- ch(i,k,4) = wa3(i-1)*ci4+wa3(i)*cr4
- 103 continue
- 104 continue
- return
- end
-
-
-
- subroutine fsinti (n,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension wsave(*)
- data pi /3.14159265358979/
- if (n .le. 1) return
- ns2 = n/2
- np1 = n+1
- dt = pi/float(np1)
- do 101 k=1,ns2
- wsave(k) = 2.*sin(k*dt)
- 101 continue
- call frffti (np1,wsave(ns2+1))
- return
- end
-
-
-
- subroutine fsint (n,x,wsave)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension x(*) ,wsave(*)
- np1 = n+1
- iw1 = n/2+1
- iw2 = iw1+np1
- iw3 = iw2+np1
- call fsint1(n,x,wsave,wsave(iw1),wsave(iw2),wsave(iw3))
- return
- end
- subroutine fsint1(n,war,was,xh,x,xxifac)
-c----------------------------TJW
- implicit real (a-h,o-z)
- implicit integer (i-n)
-c----------------------------TJW
- dimension war(*),was(*),x(*),xh(*),xxifac(*)
- data sqrt3 /1.73205080756888/
- do 100 i=1,n
- xh(i) = war(i)
- war(i) = x(i)
- 100 continue
- if (n-2) 101,102,103
- 101 xh(1) = xh(1)+xh(1)
- go to 106
- 102 xhold = sqrt3*(xh(1)+xh(2))
- xh(2) = sqrt3*(xh(1)-xh(2))
- xh(1) = xhold
- go to 106
- 103 np1 = n+1
- ns2 = n/2
- x(1) = 0.
- do 104 k=1,ns2
- kc = np1-k
- t1 = xh(k)-xh(kc)
- t2 = was(k)*(xh(k)+xh(kc))
- x(k+1) = t1+t2
- x(kc+1) = t2-t1
- 104 continue
- modn = mod(n,2)
- if (modn .ne. 0) x(ns2+2) = 4.*xh(ns2+1)
- call frfftf1 (np1,x,xh,war,xxifac)
- xh(1) = .5*x(1)
- do 105 i=3,n,2
- xh(i-1) = -x(i)
- xh(i) = xh(i-2)+x(i-1)
- 105 continue
- if (modn .ne. 0) go to 106
- xh(n) = -x(n+1)
- 106 do 107 i=1,n
- x(i) = war(i)
- war(i) = xh(i)
- 107 continue
- return
- end
-
-
-
-c========================================================================
-c $RCSfile: fftpack.F,v $ $Author: adelmann $
-c $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:25 $
-c IPPL_VERSION_ID: $Id: fftpack.F,v 1.1.1.1 2003/01/23 07:40:25 adelmann Exp $
-c========================================================================
-
diff --git a/ippl/src/FFT/fftpack.cpp b/ippl/src/FFT/fftpack.cpp
index fdf9706c0eaaed42279bce3ad1d1a120c30eb6e5..e8a01e72790b8ae38d081796f91ee4411d7f054d 100644
--- a/ippl/src/FFT/fftpack.cpp
+++ b/ippl/src/FFT/fftpack.cpp
@@ -1,3 +1,13 @@
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
/*
* This file is part of libfftpack.
*
@@ -30,6 +40,7 @@
C port by Martin Reinecke (2010)
*/
+#include <stdio.h>
#include <math.h>
#include <stdlib.h>
#include <string.h>
@@ -38,10 +49,46 @@
extern "C" {
#endif
+/******************************************************************************
+ some helper functions and macros formerly in c_utils.c and c_utils.h
+*/
+ void *util_malloc_ (size_t sz) {
+ void *res;
+ if (sz==0) return NULL;
+ res = malloc(sz);
+ if (!(res)) {
+ fprintf (stderr, "%s, %i (%s):\n%s\n",
+ __FILE__, __LINE__, __func__,
+ "malloc() failed");
+ exit (1);
+ }
+ return res;
+ }
-#include "fftpack.h"
+#define ALLOC(ptr,type,num) \
+ do { (ptr)=(type *)util_malloc_((num)*sizeof(type)); } while (0)
-#include "c_utils.c"
+#define RALLOC(type,num) \
+ ((type *)util_malloc_((num)*sizeof(type)))
+
+#define DEALLOC(ptr) \
+ { if ((ptr) != 0) free (ptr); (ptr) = NULL; }
+
+ void util_free_ (void *ptr) {
+ if ((ptr) != NULL) {
+ free(ptr);
+ }
+ }
+
+#define SWAP(a,b,type) \
+ do { type tmp_=(a); (a)=(b); (b)=tmp_; } while(0)
+/******************************************************************************/
+
+ typedef struct {
+ double r,i;
+ } cmplx;
+
+#include "fftpack.h"
#define WA(x,i) wa[(i)+(x)*ido]
#define CH(a,b,c) ch[(a)+ido*((b)+l1*(c))]
@@ -54,10 +101,6 @@ extern "C" {
/* (a+ib) = conj(c+id) * (e+if) */
#define MULPM(a,b,c,d,e,f) { a=c*e+d*f; b=c*f-d*e; }
- typedef struct {
- double r,i;
- } cmplx;
-
#define CONCAT(a,b) a ## b
#define X(arg) CONCAT(passb,arg)
@@ -75,155 +118,155 @@ extern "C" {
#define CC(a,b,c) cc[(a)+ido*((b)+l1*(c))]
#define CH(a,b,c) ch[(a)+ido*((b)+cdim*(c))]
- static void radf2 (size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=2;
- size_t i, k, ic;
- double ti2, tr2;
-
- for (k=0; k<l1; k++)
- PM (CH(0,0,k),CH(ido-1,1,k),CC(0,k,0),CC(0,k,1))
- if ((ido&1)==0)
- for (k=0; k<l1; k++)
- {
- CH( 0,1,k) = -CC(ido-1,k,1);
- CH(ido-1,0,k) = CC(ido-1,k,0);
- }
- if (ido<=2) return;
- for (k=0; k<l1; k++)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- MULPM (tr2,ti2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
- PM (CH(i-1,0,k),CH(ic-1,1,k),CC(i-1,k,0),tr2)
- PM (CH(i ,0,k),CH(ic ,1,k),ti2,CC(i ,k,0))
- }
- }
-
- static void radf3(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=3;
- static const double taur=-0.5, taui=0.86602540378443864676;
- size_t i, k, ic;
- double ci2, di2, di3, cr2, dr2, dr3, ti2, ti3, tr2, tr3;
-
- for (k=0; k<l1; k++)
- {
- cr2=CC(0,k,1)+CC(0,k,2);
- CH(0,0,k) = CC(0,k,0)+cr2;
- CH(0,2,k) = taui*(CC(0,k,2)-CC(0,k,1));
- CH(ido-1,1,k) = CC(0,k,0)+taur*cr2;
- }
- if (ido==1) return;
- for (k=0; k<l1; k++)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- MULPM (dr2,di2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
- MULPM (dr3,di3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
- cr2=dr2+dr3;
- ci2=di2+di3;
- CH(i-1,0,k) = CC(i-1,k,0)+cr2;
- CH(i ,0,k) = CC(i ,k,0)+ci2;
- tr2 = CC(i-1,k,0)+taur*cr2;
- ti2 = CC(i ,k,0)+taur*ci2;
- tr3 = taui*(di2-di3);
- ti3 = taui*(dr3-dr2);
- PM(CH(i-1,2,k),CH(ic-1,1,k),tr2,tr3)
- PM(CH(i ,2,k),CH(ic ,1,k),ti3,ti2)
- }
- }
-
- static void radf4(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=4;
- static const double hsqt2=0.70710678118654752440;
- size_t i, k, ic;
- double ci2, ci3, ci4, cr2, cr3, cr4, ti1, ti2, ti3, ti4, tr1, tr2, tr3, tr4;
-
- for (k=0; k<l1; k++)
- {
- PM (tr1,CH(0,2,k),CC(0,k,3),CC(0,k,1))
- PM (tr2,CH(ido-1,1,k),CC(0,k,0),CC(0,k,2))
- PM (CH(0,0,k),CH(ido-1,3,k),tr2,tr1)
- }
- if ((ido&1)==0)
- for (k=0; k<l1; k++)
- {
- ti1=-hsqt2*(CC(ido-1,k,1)+CC(ido-1,k,3));
- tr1= hsqt2*(CC(ido-1,k,1)-CC(ido-1,k,3));
- PM (CH(ido-1,0,k),CH(ido-1,2,k),CC(ido-1,k,0),tr1)
- PM (CH( 0,3,k),CH( 0,1,k),ti1,CC(ido-1,k,2))
- }
- if (ido<=2) return;
- for (k=0; k<l1; k++)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- MULPM(cr2,ci2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
- MULPM(cr3,ci3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
- MULPM(cr4,ci4,WA(2,i-2),WA(2,i-1),CC(i-1,k,3),CC(i,k,3))
- PM(tr1,tr4,cr4,cr2)
- PM(ti1,ti4,ci2,ci4)
- PM(tr2,tr3,CC(i-1,k,0),cr3)
- PM(ti2,ti3,CC(i ,k,0),ci3)
- PM(CH(i-1,0,k),CH(ic-1,3,k),tr2,tr1)
- PM(CH(i ,0,k),CH(ic ,3,k),ti1,ti2)
- PM(CH(i-1,2,k),CH(ic-1,1,k),tr3,ti4)
- PM(CH(i ,2,k),CH(ic ,1,k),tr4,ti3)
- }
- }
-
- static void radf5(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=5;
- static const double tr11= 0.3090169943749474241, ti11=0.95105651629515357212,
- tr12=-0.8090169943749474241, ti12=0.58778525229247312917;
- size_t i, k, ic;
- double ci2, di2, ci4, ci5, di3, di4, di5, ci3, cr2, cr3, dr2, dr3,
- dr4, dr5, cr5, cr4, ti2, ti3, ti5, ti4, tr2, tr3, tr4, tr5;
-
- for (k=0; k<l1; k++)
- {
- PM (cr2,ci5,CC(0,k,4),CC(0,k,1))
- PM (cr3,ci4,CC(0,k,3),CC(0,k,2))
- CH(0,0,k)=CC(0,k,0)+cr2+cr3;
- CH(ido-1,1,k)=CC(0,k,0)+tr11*cr2+tr12*cr3;
- CH(0,2,k)=ti11*ci5+ti12*ci4;
- CH(ido-1,3,k)=CC(0,k,0)+tr12*cr2+tr11*cr3;
- CH(0,4,k)=ti12*ci5-ti11*ci4;
- }
- if (ido==1) return;
- for (k=0; k<l1;++k)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- MULPM (dr2,di2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
- MULPM (dr3,di3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
- MULPM (dr4,di4,WA(2,i-2),WA(2,i-1),CC(i-1,k,3),CC(i,k,3))
- MULPM (dr5,di5,WA(3,i-2),WA(3,i-1),CC(i-1,k,4),CC(i,k,4))
- PM(cr2,ci5,dr5,dr2)
- PM(ci2,cr5,di2,di5)
- PM(cr3,ci4,dr4,dr3)
- PM(ci3,cr4,di3,di4)
- CH(i-1,0,k)=CC(i-1,k,0)+cr2+cr3;
- CH(i ,0,k)=CC(i ,k,0)+ci2+ci3;
- tr2=CC(i-1,k,0)+tr11*cr2+tr12*cr3;
- ti2=CC(i ,k,0)+tr11*ci2+tr12*ci3;
- tr3=CC(i-1,k,0)+tr12*cr2+tr11*cr3;
- ti3=CC(i ,k,0)+tr12*ci2+tr11*ci3;
- MULPM(tr5,tr4,cr5,cr4,ti11,ti12)
- MULPM(ti5,ti4,ci5,ci4,ti11,ti12)
- PM(CH(i-1,2,k),CH(ic-1,1,k),tr2,tr5)
- PM(CH(i ,2,k),CH(ic ,1,k),ti5,ti2)
- PM(CH(i-1,4,k),CH(ic-1,3,k),tr3,tr4)
- PM(CH(i ,4,k),CH(ic ,3,k),ti4,ti3)
- }
- }
+ static void radf2 (size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=2;
+ size_t i, k, ic;
+ double ti2, tr2;
+
+ for (k=0; k<l1; k++)
+ PM (CH(0,0,k),CH(ido-1,1,k),CC(0,k,0),CC(0,k,1))
+ if ((ido&1)==0)
+ for (k=0; k<l1; k++)
+ {
+ CH( 0,1,k) = -CC(ido-1,k,1);
+ CH(ido-1,0,k) = CC(ido-1,k,0);
+ }
+ if (ido<=2) return;
+ for (k=0; k<l1; k++)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ MULPM (tr2,ti2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
+ PM (CH(i-1,0,k),CH(ic-1,1,k),CC(i-1,k,0),tr2)
+ PM (CH(i ,0,k),CH(ic ,1,k),ti2,CC(i ,k,0))
+ }
+ }
+
+ static void radf3(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=3;
+ static const double taur=-0.5, taui=0.86602540378443864676;
+ size_t i, k, ic;
+ double ci2, di2, di3, cr2, dr2, dr3, ti2, ti3, tr2, tr3;
+
+ for (k=0; k<l1; k++)
+ {
+ cr2=CC(0,k,1)+CC(0,k,2);
+ CH(0,0,k) = CC(0,k,0)+cr2;
+ CH(0,2,k) = taui*(CC(0,k,2)-CC(0,k,1));
+ CH(ido-1,1,k) = CC(0,k,0)+taur*cr2;
+ }
+ if (ido==1) return;
+ for (k=0; k<l1; k++)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ MULPM (dr2,di2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
+ MULPM (dr3,di3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
+ cr2=dr2+dr3;
+ ci2=di2+di3;
+ CH(i-1,0,k) = CC(i-1,k,0)+cr2;
+ CH(i ,0,k) = CC(i ,k,0)+ci2;
+ tr2 = CC(i-1,k,0)+taur*cr2;
+ ti2 = CC(i ,k,0)+taur*ci2;
+ tr3 = taui*(di2-di3);
+ ti3 = taui*(dr3-dr2);
+ PM(CH(i-1,2,k),CH(ic-1,1,k),tr2,tr3)
+ PM(CH(i ,2,k),CH(ic ,1,k),ti3,ti2)
+ }
+ }
+
+ static void radf4(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=4;
+ static const double hsqt2=0.70710678118654752440;
+ size_t i, k, ic;
+ double ci2, ci3, ci4, cr2, cr3, cr4, ti1, ti2, ti3, ti4, tr1, tr2, tr3, tr4;
+
+ for (k=0; k<l1; k++)
+ {
+ PM (tr1,CH(0,2,k),CC(0,k,3),CC(0,k,1))
+ PM (tr2,CH(ido-1,1,k),CC(0,k,0),CC(0,k,2))
+ PM (CH(0,0,k),CH(ido-1,3,k),tr2,tr1)
+ }
+ if ((ido&1)==0)
+ for (k=0; k<l1; k++)
+ {
+ ti1=-hsqt2*(CC(ido-1,k,1)+CC(ido-1,k,3));
+ tr1= hsqt2*(CC(ido-1,k,1)-CC(ido-1,k,3));
+ PM (CH(ido-1,0,k),CH(ido-1,2,k),CC(ido-1,k,0),tr1)
+ PM (CH( 0,3,k),CH( 0,1,k),ti1,CC(ido-1,k,2))
+ }
+ if (ido<=2) return;
+ for (k=0; k<l1; k++)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ MULPM(cr2,ci2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
+ MULPM(cr3,ci3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
+ MULPM(cr4,ci4,WA(2,i-2),WA(2,i-1),CC(i-1,k,3),CC(i,k,3))
+ PM(tr1,tr4,cr4,cr2)
+ PM(ti1,ti4,ci2,ci4)
+ PM(tr2,tr3,CC(i-1,k,0),cr3)
+ PM(ti2,ti3,CC(i ,k,0),ci3)
+ PM(CH(i-1,0,k),CH(ic-1,3,k),tr2,tr1)
+ PM(CH(i ,0,k),CH(ic ,3,k),ti1,ti2)
+ PM(CH(i-1,2,k),CH(ic-1,1,k),tr3,ti4)
+ PM(CH(i ,2,k),CH(ic ,1,k),tr4,ti3)
+ }
+ }
+
+ static void radf5(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=5;
+ static const double tr11= 0.3090169943749474241, ti11=0.95105651629515357212,
+ tr12=-0.8090169943749474241, ti12=0.58778525229247312917;
+ size_t i, k, ic;
+ double ci2, di2, ci4, ci5, di3, di4, di5, ci3, cr2, cr3, dr2, dr3,
+ dr4, dr5, cr5, cr4, ti2, ti3, ti5, ti4, tr2, tr3, tr4, tr5;
+
+ for (k=0; k<l1; k++)
+ {
+ PM (cr2,ci5,CC(0,k,4),CC(0,k,1))
+ PM (cr3,ci4,CC(0,k,3),CC(0,k,2))
+ CH(0,0,k)=CC(0,k,0)+cr2+cr3;
+ CH(ido-1,1,k)=CC(0,k,0)+tr11*cr2+tr12*cr3;
+ CH(0,2,k)=ti11*ci5+ti12*ci4;
+ CH(ido-1,3,k)=CC(0,k,0)+tr12*cr2+tr11*cr3;
+ CH(0,4,k)=ti12*ci5-ti11*ci4;
+ }
+ if (ido==1) return;
+ for (k=0; k<l1;++k)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ MULPM (dr2,di2,WA(0,i-2),WA(0,i-1),CC(i-1,k,1),CC(i,k,1))
+ MULPM (dr3,di3,WA(1,i-2),WA(1,i-1),CC(i-1,k,2),CC(i,k,2))
+ MULPM (dr4,di4,WA(2,i-2),WA(2,i-1),CC(i-1,k,3),CC(i,k,3))
+ MULPM (dr5,di5,WA(3,i-2),WA(3,i-1),CC(i-1,k,4),CC(i,k,4))
+ PM(cr2,ci5,dr5,dr2)
+ PM(ci2,cr5,di2,di5)
+ PM(cr3,ci4,dr4,dr3)
+ PM(ci3,cr4,di3,di4)
+ CH(i-1,0,k)=CC(i-1,k,0)+cr2+cr3;
+ CH(i ,0,k)=CC(i ,k,0)+ci2+ci3;
+ tr2=CC(i-1,k,0)+tr11*cr2+tr12*cr3;
+ ti2=CC(i ,k,0)+tr11*ci2+tr12*ci3;
+ tr3=CC(i-1,k,0)+tr12*cr2+tr11*cr3;
+ ti3=CC(i ,k,0)+tr12*ci2+tr11*ci3;
+ MULPM(tr5,tr4,cr5,cr4,ti11,ti12)
+ MULPM(ti5,ti4,ci5,ci4,ti11,ti12)
+ PM(CH(i-1,2,k),CH(ic-1,1,k),tr2,tr5)
+ PM(CH(i ,2,k),CH(ic ,1,k),ti5,ti2)
+ PM(CH(i-1,4,k),CH(ic-1,3,k),tr3,tr4)
+ PM(CH(i ,4,k),CH(ic ,3,k),ti4,ti3)
+ }
+ }
#undef CH
#undef CC
@@ -232,378 +275,378 @@ extern "C" {
#define C1(a,b,c) cc[(a)+ido*((b)+l1*(c))]
#define C2(a,b) cc[(a)+idl1*(b)]
#define CH2(a,b) ch[(a)+idl1*(b)]
- static void radfg(size_t ido, size_t ip, size_t l1, size_t idl1,
- double *cc, double *ch, const double *wa)
- {
- const size_t cdim=ip;
- static const double twopi=6.28318530717958647692;
- size_t idij, ipph, i, j, k, l, j2, ic, jc, lc, ik;
- double ai1, ai2, ar1, ar2, arg;
- double *csarr;
- size_t aidx;
-
- ipph=(ip+1)/ 2;
- if(ido!=1)
- {
- memcpy(ch,cc,idl1*sizeof(double));
-
- for(j=1; j<ip; j++)
- for(k=0; k<l1; k++)
- {
- CH(0,k,j)=C1(0,k,j);
- idij=(j-1)*ido+1;
- for(i=2; i<ido; i+=2,idij+=2)
- MULPM(CH(i-1,k,j),CH(i,k,j),wa[idij-1],wa[idij],C1(i-1,k,j),C1(i,k,j))
- }
-
- for(j=1,jc=ip-1; j<ipph; j++,jc--)
- for(k=0; k<l1; k++)
- for(i=2; i<ido; i+=2)
- {
- PM(C1(i-1,k,j),C1(i ,k,jc),CH(i-1,k,jc),CH(i-1,k,j ))
- PM(C1(i ,k,j),C1(i-1,k,jc),CH(i ,k,j ),CH(i ,k,jc))
- }
- }
- else
- memcpy(cc,ch,idl1*sizeof(double));
-
- for(j=1,jc=ip-1; j<ipph; j++,jc--)
- for(k=0; k<l1; k++)
- PM(C1(0,k,j),C1(0,k,jc),CH(0,k,jc),CH(0,k,j))
-
- csarr=RALLOC(double,2*ip);
- arg=twopi / ip;
- csarr[0]=1.;
- csarr[1]=0.;
- csarr[2]=csarr[2*ip-2]=cos(arg);
- csarr[3]=sin(arg); csarr[2*ip-1]=-csarr[3];
- for (i=2; i<=ip/2; ++i)
- {
- csarr[2*i]=csarr[2*ip-2*i]=cos(i*arg);
- csarr[2*i+1]=sin(i*arg);
- csarr[2*ip-2*i+1]=-csarr[2*i+1];
- }
- for(l=1,lc=ip-1; l<ipph; l++,lc--)
- {
- ar1=csarr[2*l];
- ai1=csarr[2*l+1];
- for(ik=0; ik<idl1; ik++)
- {
- CH2(ik,l)=C2(ik,0)+ar1*C2(ik,1);
- CH2(ik,lc)=ai1*C2(ik,ip-1);
- }
- aidx=2*l;
- for(j=2,jc=ip-2; j<ipph; j++,jc--)
- {
- aidx+=2*l;
- if (aidx>=2*ip) aidx-=2*ip;
- ar2=csarr[aidx];
- ai2=csarr[aidx+1];
- for(ik=0; ik<idl1; ik++)
- {
- CH2(ik,l )+=ar2*C2(ik,j );
- CH2(ik,lc)+=ai2*C2(ik,jc);
- }
- }
- }
- DEALLOC(csarr);
-
- for(j=1; j<ipph; j++)
- for(ik=0; ik<idl1; ik++)
- CH2(ik,0)+=C2(ik,j);
-
- for(k=0; k<l1; k++)
- memcpy(&CC(0,0,k),&CH(0,k,0),ido*sizeof(double));
- for(j=1; j<ipph; j++)
- {
- jc=ip-j;
- j2=2*j;
- for(k=0; k<l1; k++)
- {
- CC(ido-1,j2-1,k) = CH(0,k,j );
- CC(0 ,j2 ,k) = CH(0,k,jc);
- }
- }
- if(ido==1) return;
-
- for(j=1; j<ipph; j++)
- {
- jc=ip-j;
- j2=2*j;
- for(k=0; k<l1; k++)
- for(i=2; i<ido; i+=2)
- {
- ic=ido-i;
- PM (CC(i-1,j2,k),CC(ic-1,j2-1,k),CH(i-1,k,j ),CH(i-1,k,jc))
- PM (CC(i ,j2,k),CC(ic ,j2-1,k),CH(i ,k,jc),CH(i ,k,j ))
- }
- }
- }
+ static void radfg(size_t ido, size_t ip, size_t l1, size_t idl1,
+ double *cc, double *ch, const double *wa)
+ {
+ const size_t cdim=ip;
+ static const double twopi=6.28318530717958647692;
+ size_t idij, ipph, i, j, k, l, j2, ic, jc, lc, ik;
+ double ai1, ai2, ar1, ar2, arg;
+ double *csarr;
+ size_t aidx;
+
+ ipph=(ip+1)/ 2;
+ if(ido!=1)
+ {
+ memcpy(ch,cc,idl1*sizeof(double));
+
+ for(j=1; j<ip; j++)
+ for(k=0; k<l1; k++)
+ {
+ CH(0,k,j)=C1(0,k,j);
+ idij=(j-1)*ido+1;
+ for(i=2; i<ido; i+=2,idij+=2)
+ MULPM(CH(i-1,k,j),CH(i,k,j),wa[idij-1],wa[idij],C1(i-1,k,j),C1(i,k,j))
+ }
+
+ for(j=1,jc=ip-1; j<ipph; j++,jc--)
+ for(k=0; k<l1; k++)
+ for(i=2; i<ido; i+=2)
+ {
+ PM(C1(i-1,k,j),C1(i ,k,jc),CH(i-1,k,jc),CH(i-1,k,j ))
+ PM(C1(i ,k,j),C1(i-1,k,jc),CH(i ,k,j ),CH(i ,k,jc))
+ }
+ }
+ else
+ memcpy(cc,ch,idl1*sizeof(double));
+
+ for(j=1,jc=ip-1; j<ipph; j++,jc--)
+ for(k=0; k<l1; k++)
+ PM(C1(0,k,j),C1(0,k,jc),CH(0,k,jc),CH(0,k,j))
+
+ csarr=RALLOC(double,2*ip);
+ arg=twopi / ip;
+ csarr[0]=1.;
+ csarr[1]=0.;
+ csarr[2]=csarr[2*ip-2]=cos(arg);
+ csarr[3]=sin(arg); csarr[2*ip-1]=-csarr[3];
+ for (i=2; i<=ip/2; ++i)
+ {
+ csarr[2*i]=csarr[2*ip-2*i]=cos(i*arg);
+ csarr[2*i+1]=sin(i*arg);
+ csarr[2*ip-2*i+1]=-csarr[2*i+1];
+ }
+ for(l=1,lc=ip-1; l<ipph; l++,lc--)
+ {
+ ar1=csarr[2*l];
+ ai1=csarr[2*l+1];
+ for(ik=0; ik<idl1; ik++)
+ {
+ CH2(ik,l)=C2(ik,0)+ar1*C2(ik,1);
+ CH2(ik,lc)=ai1*C2(ik,ip-1);
+ }
+ aidx=2*l;
+ for(j=2,jc=ip-2; j<ipph; j++,jc--)
+ {
+ aidx+=2*l;
+ if (aidx>=2*ip) aidx-=2*ip;
+ ar2=csarr[aidx];
+ ai2=csarr[aidx+1];
+ for(ik=0; ik<idl1; ik++)
+ {
+ CH2(ik,l )+=ar2*C2(ik,j );
+ CH2(ik,lc)+=ai2*C2(ik,jc);
+ }
+ }
+ }
+ DEALLOC(csarr);
+
+ for(j=1; j<ipph; j++)
+ for(ik=0; ik<idl1; ik++)
+ CH2(ik,0)+=C2(ik,j);
+
+ for(k=0; k<l1; k++)
+ memcpy(&CC(0,0,k),&CH(0,k,0),ido*sizeof(double));
+ for(j=1; j<ipph; j++)
+ {
+ jc=ip-j;
+ j2=2*j;
+ for(k=0; k<l1; k++)
+ {
+ CC(ido-1,j2-1,k) = CH(0,k,j );
+ CC(0 ,j2 ,k) = CH(0,k,jc);
+ }
+ }
+ if(ido==1) return;
+
+ for(j=1; j<ipph; j++)
+ {
+ jc=ip-j;
+ j2=2*j;
+ for(k=0; k<l1; k++)
+ for(i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ PM (CC(i-1,j2,k),CC(ic-1,j2-1,k),CH(i-1,k,j ),CH(i-1,k,jc))
+ PM (CC(i ,j2,k),CC(ic ,j2-1,k),CH(i ,k,jc),CH(i ,k,j ))
+ }
+ }
+ }
#undef CC
#undef CH
#define CH(a,b,c) ch[(a)+ido*((b)+l1*(c))]
#define CC(a,b,c) cc[(a)+ido*((b)+cdim*(c))]
- static void radb2(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=2;
- size_t i, k, ic;
- double ti2, tr2;
-
- for (k=0; k<l1; k++)
- PM (CH(0,k,0),CH(0,k,1),CC(0,0,k),CC(ido-1,1,k))
- if ((ido&1)==0)
- for (k=0; k<l1; k++)
- {
- CH(ido-1,k,0) = 2*CC(ido-1,0,k);
- CH(ido-1,k,1) = -2*CC(0 ,1,k);
- }
- if (ido<=2) return;
- for (k=0; k<l1;++k)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- PM (CH(i-1,k,0),tr2,CC(i-1,0,k),CC(ic-1,1,k))
- PM (ti2,CH(i ,k,0),CC(i ,0,k),CC(ic ,1,k))
- MULPM (CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),ti2,tr2)
- }
- }
-
- static void radb3(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=3;
- static const double taur=-0.5, taui=0.86602540378443864676;
- size_t i, k, ic;
- double ci2, ci3, di2, di3, cr2, cr3, dr2, dr3, ti2, tr2;
-
- for (k=0; k<l1; k++)
- {
- tr2=2*CC(ido-1,1,k);
- cr2=CC(0,0,k)+taur*tr2;
- CH(0,k,0)=CC(0,0,k)+tr2;
- ci3=2*taui*CC(0,2,k);
- PM (CH(0,k,2),CH(0,k,1),cr2,ci3);
- }
- if (ido==1) return;
- for (k=0; k<l1; k++)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- tr2=CC(i-1,2,k)+CC(ic-1,1,k);
- ti2=CC(i ,2,k)-CC(ic ,1,k);
- cr2=CC(i-1,0,k)+taur*tr2;
- ci2=CC(i ,0,k)+taur*ti2;
- CH(i-1,k,0)=CC(i-1,0,k)+tr2;
- CH(i ,k,0)=CC(i ,0,k)+ti2;
- cr3=taui*(CC(i-1,2,k)-CC(ic-1,1,k));
- ci3=taui*(CC(i ,2,k)+CC(ic ,1,k));
- PM(dr3,dr2,cr2,ci3)
- PM(di2,di3,ci2,cr3)
- MULPM(CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),di2,dr2)
- MULPM(CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),di3,dr3)
- }
- }
-
- static void radb4(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=4;
- static const double sqrt2=1.41421356237309504880;
- size_t i, k, ic;
- double ci2, ci3, ci4, cr2, cr3, cr4, ti1, ti2, ti3, ti4, tr1, tr2, tr3, tr4;
-
- for (k=0; k<l1; k++)
- {
- PM (tr2,tr1,CC(0,0,k),CC(ido-1,3,k))
- tr3=2*CC(ido-1,1,k);
- tr4=2*CC(0,2,k);
- PM (CH(0,k,0),CH(0,k,2),tr2,tr3)
- PM (CH(0,k,3),CH(0,k,1),tr1,tr4)
- }
- if ((ido&1)==0)
- for (k=0; k<l1; k++)
- {
- PM (ti1,ti2,CC(0 ,3,k),CC(0 ,1,k))
- PM (tr2,tr1,CC(ido-1,0,k),CC(ido-1,2,k))
- CH(ido-1,k,0)=tr2+tr2;
- CH(ido-1,k,1)=sqrt2*(tr1-ti1);
- CH(ido-1,k,2)=ti2+ti2;
- CH(ido-1,k,3)=-sqrt2*(tr1+ti1);
- }
- if (ido<=2) return;
- for (k=0; k<l1;++k)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- PM (tr2,tr1,CC(i-1,0,k),CC(ic-1,3,k))
- PM (ti1,ti2,CC(i ,0,k),CC(ic ,3,k))
- PM (tr4,ti3,CC(i ,2,k),CC(ic ,1,k))
- PM (tr3,ti4,CC(i-1,2,k),CC(ic-1,1,k))
- PM (CH(i-1,k,0),cr3,tr2,tr3)
- PM (CH(i ,k,0),ci3,ti2,ti3)
- PM (cr4,cr2,tr1,tr4)
- PM (ci2,ci4,ti1,ti4)
- MULPM (CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),ci2,cr2)
- MULPM (CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),ci3,cr3)
- MULPM (CH(i,k,3),CH(i-1,k,3),WA(2,i-2),WA(2,i-1),ci4,cr4)
- }
- }
-
- static void radb5(size_t ido, size_t l1, const double *cc, double *ch,
- const double *wa)
- {
- const size_t cdim=5;
- static const double tr11= 0.3090169943749474241, ti11=0.95105651629515357212,
- tr12=-0.8090169943749474241, ti12=0.58778525229247312917;
- size_t i, k, ic;
- double ci2, ci3, ci4, ci5, di3, di4, di5, di2, cr2, cr3, cr5, cr4,
- ti2, ti3, ti4, ti5, dr3, dr4, dr5, dr2, tr2, tr3, tr4, tr5;
-
- for (k=0; k<l1; k++)
- {
- ti5=2*CC(0,2,k);
- ti4=2*CC(0,4,k);
- tr2=2*CC(ido-1,1,k);
- tr3=2*CC(ido-1,3,k);
- CH(0,k,0)=CC(0,0,k)+tr2+tr3;
- cr2=CC(0,0,k)+tr11*tr2+tr12*tr3;
- cr3=CC(0,0,k)+tr12*tr2+tr11*tr3;
- MULPM(ci5,ci4,ti5,ti4,ti11,ti12)
- PM(CH(0,k,4),CH(0,k,1),cr2,ci5)
- PM(CH(0,k,3),CH(0,k,2),cr3,ci4)
- }
- if (ido==1) return;
- for (k=0; k<l1;++k)
- for (i=2; i<ido; i+=2)
- {
- ic=ido-i;
- PM(tr2,tr5,CC(i-1,2,k),CC(ic-1,1,k))
- PM(ti5,ti2,CC(i ,2,k),CC(ic ,1,k))
- PM(tr3,tr4,CC(i-1,4,k),CC(ic-1,3,k))
- PM(ti4,ti3,CC(i ,4,k),CC(ic ,3,k))
- CH(i-1,k,0)=CC(i-1,0,k)+tr2+tr3;
- CH(i ,k,0)=CC(i ,0,k)+ti2+ti3;
- cr2=CC(i-1,0,k)+tr11*tr2+tr12*tr3;
- ci2=CC(i ,0,k)+tr11*ti2+tr12*ti3;
- cr3=CC(i-1,0,k)+tr12*tr2+tr11*tr3;
- ci3=CC(i ,0,k)+tr12*ti2+tr11*ti3;
- MULPM(cr5,cr4,tr5,tr4,ti11,ti12)
- MULPM(ci5,ci4,ti5,ti4,ti11,ti12)
- PM(dr4,dr3,cr3,ci4)
- PM(di3,di4,ci3,cr4)
- PM(dr5,dr2,cr2,ci5)
- PM(di2,di5,ci2,cr5)
- MULPM(CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),di2,dr2)
- MULPM(CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),di3,dr3)
- MULPM(CH(i,k,3),CH(i-1,k,3),WA(2,i-2),WA(2,i-1),di4,dr4)
- MULPM(CH(i,k,4),CH(i-1,k,4),WA(3,i-2),WA(3,i-1),di5,dr5)
- }
- }
-
- static void radbg(size_t ido, size_t ip, size_t l1, size_t idl1,
- double *cc, double *ch, const double *wa)
- {
- const size_t cdim=ip;
- static const double twopi=6.28318530717958647692;
- size_t idij, ipph, i, j, k, l, j2, ic, jc, lc, ik;
- double ai1, ai2, ar1, ar2, arg;
- double *csarr;
- size_t aidx;
-
- ipph=(ip+1)/ 2;
- for(k=0; k<l1; k++)
- memcpy(&CH(0,k,0),&CC(0,0,k),ido*sizeof(double));
- for(j=1; j<ipph; j++)
- {
- jc=ip-j;
- j2=2*j;
- for(k=0; k<l1; k++)
- {
- CH(0,k,j )=2*CC(ido-1,j2-1,k);
- CH(0,k,jc)=2*CC(0 ,j2 ,k);
- }
- }
-
- if(ido!=1)
- for(j=1,jc=ip-1; j<ipph; j++,jc--)
- for(k=0; k<l1; k++)
- for(i=2; i<ido; i+=2)
- {
- ic=ido-i;
- PM (CH(i-1,k,j ),CH(i-1,k,jc),CC(i-1,2*j,k),CC(ic-1,2*j-1,k))
- PM (CH(i ,k,jc),CH(i ,k,j ),CC(i ,2*j,k),CC(ic ,2*j-1,k))
- }
-
- csarr=RALLOC(double,2*ip);
- arg=twopi/ip;
- csarr[0]=1.;
- csarr[1]=0.;
- csarr[2]=csarr[2*ip-2]=cos(arg);
- csarr[3]=sin(arg); csarr[2*ip-1]=-csarr[3];
- for (i=2; i<=ip/2; ++i)
- {
- csarr[2*i]=csarr[2*ip-2*i]=cos(i*arg);
- csarr[2*i+1]=sin(i*arg);
- csarr[2*ip-2*i+1]=-csarr[2*i+1];
- }
- for(l=1; l<ipph; l++)
- {
- lc=ip-l;
- ar1=csarr[2*l];
- ai1=csarr[2*l+1];
- for(ik=0; ik<idl1; ik++)
- {
- C2(ik,l)=CH2(ik,0)+ar1*CH2(ik,1);
- C2(ik,lc)=ai1*CH2(ik,ip-1);
- }
- aidx=2*l;
- for(j=2; j<ipph; j++)
- {
- jc=ip-j;
- aidx+=2*l;
- if (aidx>=2*ip) aidx-=2*ip;
- ar2=csarr[aidx];
- ai2=csarr[aidx+1];
- for(ik=0; ik<idl1; ik++)
- {
- C2(ik,l )+=ar2*CH2(ik,j );
- C2(ik,lc)+=ai2*CH2(ik,jc);
- }
- }
- }
- DEALLOC(csarr);
-
- for(j=1; j<ipph; j++)
- for(ik=0; ik<idl1; ik++)
- CH2(ik,0)+=CH2(ik,j);
-
- for(j=1,jc=ip-1; j<ipph; j++,jc--)
- for(k=0; k<l1; k++)
- PM (CH(0,k,jc),CH(0,k,j),C1(0,k,j),C1(0,k,jc))
-
- if(ido==1)
- return;
- for(j=1,jc=ip-1; j<ipph; j++,jc--)
- for(k=0; k<l1; k++)
- for(i=2; i<ido; i+=2)
- {
- PM (CH(i-1,k,jc),CH(i-1,k,j ),C1(i-1,k,j),C1(i ,k,jc))
- PM (CH(i ,k,j ),CH(i ,k,jc),C1(i ,k,j),C1(i-1,k,jc))
- }
- memcpy(cc,ch,idl1*sizeof(double));
-
- for(j=1; j<ip; j++)
- for(k=0; k<l1; k++)
- {
- C1(0,k,j)=CH(0,k,j);
- idij=(j-1)*ido+1;
- for(i=2; i<ido; i+=2,idij+=2)
- MULPM (C1(i,k,j),C1(i-1,k,j),wa[idij-1],wa[idij],CH(i,k,j),CH(i-1,k,j))
- }
- }
+ static void radb2(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=2;
+ size_t i, k, ic;
+ double ti2, tr2;
+
+ for (k=0; k<l1; k++)
+ PM (CH(0,k,0),CH(0,k,1),CC(0,0,k),CC(ido-1,1,k))
+ if ((ido&1)==0)
+ for (k=0; k<l1; k++)
+ {
+ CH(ido-1,k,0) = 2*CC(ido-1,0,k);
+ CH(ido-1,k,1) = -2*CC(0 ,1,k);
+ }
+ if (ido<=2) return;
+ for (k=0; k<l1;++k)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ PM (CH(i-1,k,0),tr2,CC(i-1,0,k),CC(ic-1,1,k))
+ PM (ti2,CH(i ,k,0),CC(i ,0,k),CC(ic ,1,k))
+ MULPM (CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),ti2,tr2)
+ }
+ }
+
+ static void radb3(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=3;
+ static const double taur=-0.5, taui=0.86602540378443864676;
+ size_t i, k, ic;
+ double ci2, ci3, di2, di3, cr2, cr3, dr2, dr3, ti2, tr2;
+
+ for (k=0; k<l1; k++)
+ {
+ tr2=2*CC(ido-1,1,k);
+ cr2=CC(0,0,k)+taur*tr2;
+ CH(0,k,0)=CC(0,0,k)+tr2;
+ ci3=2*taui*CC(0,2,k);
+ PM (CH(0,k,2),CH(0,k,1),cr2,ci3);
+ }
+ if (ido==1) return;
+ for (k=0; k<l1; k++)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ tr2=CC(i-1,2,k)+CC(ic-1,1,k);
+ ti2=CC(i ,2,k)-CC(ic ,1,k);
+ cr2=CC(i-1,0,k)+taur*tr2;
+ ci2=CC(i ,0,k)+taur*ti2;
+ CH(i-1,k,0)=CC(i-1,0,k)+tr2;
+ CH(i ,k,0)=CC(i ,0,k)+ti2;
+ cr3=taui*(CC(i-1,2,k)-CC(ic-1,1,k));
+ ci3=taui*(CC(i ,2,k)+CC(ic ,1,k));
+ PM(dr3,dr2,cr2,ci3)
+ PM(di2,di3,ci2,cr3)
+ MULPM(CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),di2,dr2)
+ MULPM(CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),di3,dr3)
+ }
+ }
+
+ static void radb4(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=4;
+ static const double sqrt2=1.41421356237309504880;
+ size_t i, k, ic;
+ double ci2, ci3, ci4, cr2, cr3, cr4, ti1, ti2, ti3, ti4, tr1, tr2, tr3, tr4;
+
+ for (k=0; k<l1; k++)
+ {
+ PM (tr2,tr1,CC(0,0,k),CC(ido-1,3,k))
+ tr3=2*CC(ido-1,1,k);
+ tr4=2*CC(0,2,k);
+ PM (CH(0,k,0),CH(0,k,2),tr2,tr3)
+ PM (CH(0,k,3),CH(0,k,1),tr1,tr4)
+ }
+ if ((ido&1)==0)
+ for (k=0; k<l1; k++)
+ {
+ PM (ti1,ti2,CC(0 ,3,k),CC(0 ,1,k))
+ PM (tr2,tr1,CC(ido-1,0,k),CC(ido-1,2,k))
+ CH(ido-1,k,0)=tr2+tr2;
+ CH(ido-1,k,1)=sqrt2*(tr1-ti1);
+ CH(ido-1,k,2)=ti2+ti2;
+ CH(ido-1,k,3)=-sqrt2*(tr1+ti1);
+ }
+ if (ido<=2) return;
+ for (k=0; k<l1;++k)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ PM (tr2,tr1,CC(i-1,0,k),CC(ic-1,3,k))
+ PM (ti1,ti2,CC(i ,0,k),CC(ic ,3,k))
+ PM (tr4,ti3,CC(i ,2,k),CC(ic ,1,k))
+ PM (tr3,ti4,CC(i-1,2,k),CC(ic-1,1,k))
+ PM (CH(i-1,k,0),cr3,tr2,tr3)
+ PM (CH(i ,k,0),ci3,ti2,ti3)
+ PM (cr4,cr2,tr1,tr4)
+ PM (ci2,ci4,ti1,ti4)
+ MULPM (CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),ci2,cr2)
+ MULPM (CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),ci3,cr3)
+ MULPM (CH(i,k,3),CH(i-1,k,3),WA(2,i-2),WA(2,i-1),ci4,cr4)
+ }
+ }
+
+ static void radb5(size_t ido, size_t l1, const double *cc, double *ch,
+ const double *wa)
+ {
+ const size_t cdim=5;
+ static const double tr11= 0.3090169943749474241, ti11=0.95105651629515357212,
+ tr12=-0.8090169943749474241, ti12=0.58778525229247312917;
+ size_t i, k, ic;
+ double ci2, ci3, ci4, ci5, di3, di4, di5, di2, cr2, cr3, cr5, cr4,
+ ti2, ti3, ti4, ti5, dr3, dr4, dr5, dr2, tr2, tr3, tr4, tr5;
+
+ for (k=0; k<l1; k++)
+ {
+ ti5=2*CC(0,2,k);
+ ti4=2*CC(0,4,k);
+ tr2=2*CC(ido-1,1,k);
+ tr3=2*CC(ido-1,3,k);
+ CH(0,k,0)=CC(0,0,k)+tr2+tr3;
+ cr2=CC(0,0,k)+tr11*tr2+tr12*tr3;
+ cr3=CC(0,0,k)+tr12*tr2+tr11*tr3;
+ MULPM(ci5,ci4,ti5,ti4,ti11,ti12)
+ PM(CH(0,k,4),CH(0,k,1),cr2,ci5)
+ PM(CH(0,k,3),CH(0,k,2),cr3,ci4)
+ }
+ if (ido==1) return;
+ for (k=0; k<l1;++k)
+ for (i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ PM(tr2,tr5,CC(i-1,2,k),CC(ic-1,1,k))
+ PM(ti5,ti2,CC(i ,2,k),CC(ic ,1,k))
+ PM(tr3,tr4,CC(i-1,4,k),CC(ic-1,3,k))
+ PM(ti4,ti3,CC(i ,4,k),CC(ic ,3,k))
+ CH(i-1,k,0)=CC(i-1,0,k)+tr2+tr3;
+ CH(i ,k,0)=CC(i ,0,k)+ti2+ti3;
+ cr2=CC(i-1,0,k)+tr11*tr2+tr12*tr3;
+ ci2=CC(i ,0,k)+tr11*ti2+tr12*ti3;
+ cr3=CC(i-1,0,k)+tr12*tr2+tr11*tr3;
+ ci3=CC(i ,0,k)+tr12*ti2+tr11*ti3;
+ MULPM(cr5,cr4,tr5,tr4,ti11,ti12)
+ MULPM(ci5,ci4,ti5,ti4,ti11,ti12)
+ PM(dr4,dr3,cr3,ci4)
+ PM(di3,di4,ci3,cr4)
+ PM(dr5,dr2,cr2,ci5)
+ PM(di2,di5,ci2,cr5)
+ MULPM(CH(i,k,1),CH(i-1,k,1),WA(0,i-2),WA(0,i-1),di2,dr2)
+ MULPM(CH(i,k,2),CH(i-1,k,2),WA(1,i-2),WA(1,i-1),di3,dr3)
+ MULPM(CH(i,k,3),CH(i-1,k,3),WA(2,i-2),WA(2,i-1),di4,dr4)
+ MULPM(CH(i,k,4),CH(i-1,k,4),WA(3,i-2),WA(3,i-1),di5,dr5)
+ }
+ }
+
+ static void radbg(size_t ido, size_t ip, size_t l1, size_t idl1,
+ double *cc, double *ch, const double *wa)
+ {
+ const size_t cdim=ip;
+ static const double twopi=6.28318530717958647692;
+ size_t idij, ipph, i, j, k, l, j2, ic, jc, lc, ik;
+ double ai1, ai2, ar1, ar2, arg;
+ double *csarr;
+ size_t aidx;
+
+ ipph=(ip+1)/ 2;
+ for(k=0; k<l1; k++)
+ memcpy(&CH(0,k,0),&CC(0,0,k),ido*sizeof(double));
+ for(j=1; j<ipph; j++)
+ {
+ jc=ip-j;
+ j2=2*j;
+ for(k=0; k<l1; k++)
+ {
+ CH(0,k,j )=2*CC(ido-1,j2-1,k);
+ CH(0,k,jc)=2*CC(0 ,j2 ,k);
+ }
+ }
+
+ if(ido!=1)
+ for(j=1,jc=ip-1; j<ipph; j++,jc--)
+ for(k=0; k<l1; k++)
+ for(i=2; i<ido; i+=2)
+ {
+ ic=ido-i;
+ PM (CH(i-1,k,j ),CH(i-1,k,jc),CC(i-1,2*j,k),CC(ic-1,2*j-1,k))
+ PM (CH(i ,k,jc),CH(i ,k,j ),CC(i ,2*j,k),CC(ic ,2*j-1,k))
+ }
+
+ csarr=RALLOC(double,2*ip);
+ arg=twopi/ip;
+ csarr[0]=1.;
+ csarr[1]=0.;
+ csarr[2]=csarr[2*ip-2]=cos(arg);
+ csarr[3]=sin(arg); csarr[2*ip-1]=-csarr[3];
+ for (i=2; i<=ip/2; ++i)
+ {
+ csarr[2*i]=csarr[2*ip-2*i]=cos(i*arg);
+ csarr[2*i+1]=sin(i*arg);
+ csarr[2*ip-2*i+1]=-csarr[2*i+1];
+ }
+ for(l=1; l<ipph; l++)
+ {
+ lc=ip-l;
+ ar1=csarr[2*l];
+ ai1=csarr[2*l+1];
+ for(ik=0; ik<idl1; ik++)
+ {
+ C2(ik,l)=CH2(ik,0)+ar1*CH2(ik,1);
+ C2(ik,lc)=ai1*CH2(ik,ip-1);
+ }
+ aidx=2*l;
+ for(j=2; j<ipph; j++)
+ {
+ jc=ip-j;
+ aidx+=2*l;
+ if (aidx>=2*ip) aidx-=2*ip;
+ ar2=csarr[aidx];
+ ai2=csarr[aidx+1];
+ for(ik=0; ik<idl1; ik++)
+ {
+ C2(ik,l )+=ar2*CH2(ik,j );
+ C2(ik,lc)+=ai2*CH2(ik,jc);
+ }
+ }
+ }
+ DEALLOC(csarr);
+
+ for(j=1; j<ipph; j++)
+ for(ik=0; ik<idl1; ik++)
+ CH2(ik,0)+=CH2(ik,j);
+
+ for(j=1,jc=ip-1; j<ipph; j++,jc--)
+ for(k=0; k<l1; k++)
+ PM (CH(0,k,jc),CH(0,k,j),C1(0,k,j),C1(0,k,jc))
+
+ if(ido==1)
+ return;
+ for(j=1,jc=ip-1; j<ipph; j++,jc--)
+ for(k=0; k<l1; k++)
+ for(i=2; i<ido; i+=2)
+ {
+ PM (CH(i-1,k,jc),CH(i-1,k,j ),C1(i-1,k,j),C1(i ,k,jc))
+ PM (CH(i ,k,j ),CH(i ,k,jc),C1(i ,k,j),C1(i-1,k,jc))
+ }
+ memcpy(cc,ch,idl1*sizeof(double));
+
+ for(j=1; j<ip; j++)
+ for(k=0; k<l1; k++)
+ {
+ C1(0,k,j)=CH(0,k,j);
+ idij=(j-1)*ido+1;
+ for(i=2; i<ido; i+=2,idij+=2)
+ MULPM (C1(i,k,j),C1(i-1,k,j),wa[idij-1],wa[idij],CH(i,k,j),CH(i-1,k,j))
+ }
+ }
#undef CC
#undef CH
@@ -615,318 +658,324 @@ extern "C" {
cfftf1, cfftb1, cfftf, cfftb, cffti1, cffti. Complex FFTs.
----------------------------------------------------------------------*/
- static void cfft1(size_t n, cmplx c[], cmplx ch[], const cmplx wa[],
- const size_t ifac[], int isign)
- {
- size_t k1, l1=1, nf=ifac[1], iw=0;
- cmplx *p1=c, *p2=ch;
-
- for(k1=0; k1<nf; k1++)
- {
- size_t ip=ifac[k1+2];
- size_t l2=ip*l1;
- size_t ido = n/l2;
- if(ip==4)
- (isign>0) ? passb4(ido, l1, p1, p2, wa+iw)
- : passf4(ido, l1, p1, p2, wa+iw);
- else if(ip==2)
- (isign>0) ? passb2(ido, l1, p1, p2, wa+iw)
- : passf2(ido, l1, p1, p2, wa+iw);
- else if(ip==3)
- (isign>0) ? passb3(ido, l1, p1, p2, wa+iw)
- : passf3(ido, l1, p1, p2, wa+iw);
- else if(ip==5)
- (isign>0) ? passb5(ido, l1, p1, p2, wa+iw)
- : passf5(ido, l1, p1, p2, wa+iw);
- else if(ip==6)
- (isign>0) ? passb6(ido, l1, p1, p2, wa+iw)
- : passf6(ido, l1, p1, p2, wa+iw);
- else
- (isign>0) ? passbg(ido, ip, l1, p1, p2, wa+iw)
- : passfg(ido, ip, l1, p1, p2, wa+iw);
- SWAP(p1,p2,cmplx *);
- l1=l2;
- iw+=(ip-1)*ido;
- }
- if (p1!=c)
- memcpy (c,p1,n*sizeof(cmplx));
- }
-
- void cfftf(size_t n, double c[], double wsave[])
- {
- if (n!=1)
- cfft1(n, (cmplx*)c, (cmplx*)wsave, (cmplx*)(wsave+2*n),
- (size_t*)(wsave+4*n),-1);
- }
-
- void cfftb(size_t n, double c[], double wsave[])
- {
- if (n!=1)
- cfft1(n, (cmplx*)c, (cmplx*)wsave, (cmplx*)(wsave+2*n),
- (size_t*)(wsave+4*n),+1);
- }
-
- static void factorize (size_t n, const size_t *pf, size_t npf, size_t *ifac)
- {
- size_t nl=n, nf=0, ntry=0, j=0, i;
-
- startloop:
- j++;
- ntry = (j<=npf) ? pf[j-1] : ntry+2;
- do
- {
- size_t nq=nl / ntry;
- size_t nr=nl-ntry*nq;
- if (nr!=0)
- goto startloop;
- nf++;
- ifac[nf+1]=ntry;
- nl=nq;
- if ((ntry==2) && (nf!=1))
- {
- for (i=nf+1; i>2; --i)
- ifac[i]=ifac[i-1];
- ifac[2]=2;
- }
- }
- while(nl!=1);
- ifac[0]=n;
- ifac[1]=nf;
- }
-
- static void cffti1(size_t n, double wa[], size_t ifac[])
- {
- static const size_t ntryh[5]={4,6,3,2,5};
- static const double twopi=6.28318530717958647692;
- size_t j, k, fi;
-
- double argh=twopi/n;
- size_t i=0, l1=1;
- factorize (n,ntryh,5,ifac);
- for(k=1; k<=ifac[1]; k++)
- {
- size_t ip=ifac[k+1];
- size_t ido=n/(l1*ip);
- for(j=1; j<ip; j++)
- {
- size_t is = i;
- double argld=j*l1*argh;
- wa[i ]=1;
- wa[i+1]=0;
- for(fi=1; fi<=ido; fi++)
- {
- double arg=fi*argld;
- i+=2;
- wa[i ]=cos(arg);
- wa[i+1]=sin(arg);
- }
- if(ip>6)
- {
- wa[is ]=wa[i ];
- wa[is+1]=wa[i+1];
- }
- }
- l1*=ip;
- }
- }
-
- void cffti(size_t n, double wsave[])
- { if (n!=1) cffti1(n, wsave+2*n,(size_t*)(wsave+4*n)); }
+ static void cfft1(size_t n, cmplx c[], cmplx ch[], const cmplx wa[],
+ const size_t ifac[], int isign)
+ {
+ size_t k1, l1=1, nf=ifac[1], iw=0;
+ cmplx *p1=c, *p2=ch;
+
+ for(k1=0; k1<nf; k1++)
+ {
+ size_t ip=ifac[k1+2];
+ size_t l2=ip*l1;
+ size_t ido = n/l2;
+ if(ip==4)
+ (isign>0) ? passb4(ido, l1, p1, p2, wa+iw)
+ : passf4(ido, l1, p1, p2, wa+iw);
+ else if(ip==2)
+ (isign>0) ? passb2(ido, l1, p1, p2, wa+iw)
+ : passf2(ido, l1, p1, p2, wa+iw);
+ else if(ip==3)
+ (isign>0) ? passb3(ido, l1, p1, p2, wa+iw)
+ : passf3(ido, l1, p1, p2, wa+iw);
+ else if(ip==5)
+ (isign>0) ? passb5(ido, l1, p1, p2, wa+iw)
+ : passf5(ido, l1, p1, p2, wa+iw);
+ else if(ip==6)
+ (isign>0) ? passb6(ido, l1, p1, p2, wa+iw)
+ : passf6(ido, l1, p1, p2, wa+iw);
+ else
+ (isign>0) ? passbg(ido, ip, l1, p1, p2, wa+iw)
+ : passfg(ido, ip, l1, p1, p2, wa+iw);
+ SWAP(p1,p2,cmplx *);
+ l1=l2;
+ iw+=(ip-1)*ido;
+ }
+ if (p1!=c)
+ memcpy (c,p1,n*sizeof(cmplx));
+ }
+
+ void cfftf(size_t n, double c[], double wsave[])
+ {
+ if (n!=1)
+ cfft1(n, (cmplx*)c, (cmplx*)wsave, (cmplx*)(wsave+2*n),
+ (size_t*)(wsave+4*n),-1);
+ }
+
+ void cfftb(size_t n, double c[], double wsave[])
+ {
+ if (n!=1)
+ cfft1(n, (cmplx*)c, (cmplx*)wsave, (cmplx*)(wsave+2*n),
+ (size_t*)(wsave+4*n),+1);
+ }
+
+ static void factorize (size_t n, const size_t *pf, size_t npf, size_t *ifac)
+ {
+ size_t nl=n, nf=0, ntry=0, j=0, i;
+
+ startloop:
+ j++;
+ ntry = (j<=npf) ? pf[j-1] : ntry+2;
+ do
+ {
+ size_t nq=nl / ntry;
+ size_t nr=nl-ntry*nq;
+ if (nr!=0)
+ goto startloop;
+ nf++;
+ ifac[nf+1]=ntry;
+ nl=nq;
+ if ((ntry==2) && (nf!=1))
+ {
+ for (i=nf+1; i>2; --i)
+ ifac[i]=ifac[i-1];
+ ifac[2]=2;
+ }
+ }
+ while(nl!=1);
+ ifac[0]=n;
+ ifac[1]=nf;
+ }
+
+ static void cffti1(size_t n, double wa[], size_t ifac[])
+ {
+ static const size_t ntryh[5]={4,6,3,2,5};
+ static const double twopi=6.28318530717958647692;
+ size_t j, k, fi;
+
+ double argh=twopi/n;
+ size_t i=0, l1=1;
+ factorize (n,ntryh,5,ifac);
+ for(k=1; k<=ifac[1]; k++)
+ {
+ size_t ip=ifac[k+1];
+ size_t ido=n/(l1*ip);
+ for(j=1; j<ip; j++)
+ {
+ size_t is = i;
+ double argld=j*l1*argh;
+ wa[i ]=1;
+ wa[i+1]=0;
+ for(fi=1; fi<=ido; fi++)
+ {
+ double arg=fi*argld;
+ i+=2;
+ wa[i ]=cos(arg);
+ wa[i+1]=sin(arg);
+ }
+ if(ip>6)
+ {
+ wa[is ]=wa[i ];
+ wa[is+1]=wa[i+1];
+ }
+ }
+ l1*=ip;
+ }
+ }
+
+ void cffti(size_t n, double wsave[])
+ { if (n!=1) cffti1(n, wsave+2*n,(size_t*)(wsave+4*n)); }
/*----------------------------------------------------------------------
rfftf1, rfftb1, rfftf, rfftb, rffti1, rffti. Real FFTs.
----------------------------------------------------------------------*/
- static void rfftf1(
- size_t n,
- double c[],
- double ch[],
- const double wa[],
- const size_t ifac[]
- ) {
- size_t k1, l1=n, nf=ifac[1], iw=n-1;
- double *p1=ch, *p2=c;
-
- for(k1=1; k1<=nf;++k1)
- {
- size_t ip=ifac[nf-k1+2];
- size_t ido=n / l1;
- l1 /= ip;
- iw-=(ip-1)*ido;
- SWAP (p1,p2,double *);
- if(ip==4)
- radf4(ido, l1, p1, p2, wa+iw);
- else if(ip==2)
- radf2(ido, l1, p1, p2, wa+iw);
- else if(ip==3)
- radf3(ido, l1, p1, p2, wa+iw);
- else if(ip==5)
- radf5(ido, l1, p1, p2, wa+iw);
- else
- {
- if (ido==1)
- SWAP (p1,p2,double *);
- radfg(ido, ip, l1, ido*l1, p1, p2, wa+iw);
- SWAP (p1,p2,double *);
- }
- }
- if (p1==c)
- memcpy (c,ch,n*sizeof(double));
- }
-
- static void rfftb1(size_t n, double c[], double ch[], const double wa[],
- const size_t ifac[])
- {
- size_t k1, l1=1, nf=ifac[1], iw=0;
- double *p1=c, *p2=ch;
-
- for(k1=1; k1<=nf; k1++)
- {
- size_t ip = ifac[k1+1],
- ido= n/(ip*l1);
- if(ip==4)
- radb4(ido, l1, p1, p2, wa+iw);
- else if(ip==2)
- radb2(ido, l1, p1, p2, wa+iw);
- else if(ip==3)
- radb3(ido, l1, p1, p2, wa+iw);
- else if(ip==5)
- radb5(ido, l1, p1, p2, wa+iw);
- else
- {
- radbg(ido, ip, l1, ido*l1, p1, p2, wa+iw);
- if (ido!=1)
- SWAP (p1,p2,double *);
- }
- SWAP (p1,p2,double *);
- l1*=ip;
- iw+=(ip-1)*ido;
- }
- if (p1!=c)
- memcpy (c,ch,n*sizeof(double));
- }
-
- void rfftf(size_t n, double r[], double wsave[])
- { if(n!=1) rfftf1(n, r, wsave, wsave+n,(size_t*)(wsave+2*n)); }
-
- void rfftb(size_t n, double r[], double wsave[])
- { if(n!=1) rfftb1(n, r, wsave, wsave+n,(size_t*)(wsave+2*n)); }
-
- static void rffti1(size_t n, double wa[], size_t ifac[])
- {
- static const size_t ntryh[4]={4,2,3,5};
- static const double twopi=6.28318530717958647692;
- size_t i, j, k, fi;
-
- double argh=twopi/n;
- size_t is=0, l1=1;
- factorize (n,ntryh,4,ifac);
- for (k=1; k<ifac[1]; k++)
- {
- size_t ip=ifac[k+1],
- ido=n/(l1*ip);
- for (j=1; j<ip; ++j)
- {
- double argld=j*l1*argh;
- for(i=is,fi=1; i<=ido+is-3; i+=2,++fi)
- {
- double arg=fi*argld;
- wa[i ]=cos(arg);
- wa[i+1]=sin(arg);
- }
- is+=ido;
- }
- l1*=ip;
- }
- }
-
- void rffti(size_t n, double wsave[])
- {
- if (n!=1) {
- rffti1(n, wsave+n,(size_t*)(wsave+2*n));
- }
- }
-
- const double pi = 3.14159265358979;
- void sinti (size_t n, double wsave[])
- {
- if (n <= 1) return;
- size_t ns2 = n / 2;
- double dt = pi / double (n+1);
- for (size_t k = 0; k < ns2; k++) {
- wsave[k] = 2.0 * sin ((k+1) * dt);
- }
- rffti (n+1, wsave+ns2);
- }
-
- /*
- with n = 20
- +0 -> was wsave[0..9]
- +10 -> xh wsave[10..30]
- +31 -> x wsave[31..51]
- +52 -> xxifac wsave[52..]
- */
-
- void sint1 (size_t n, double war[], double was[], double xh[], double x[], double xxifac[])
- {
- double sqrt3 = 1.73205080756888;
-
- for (size_t i = 0; i < n; i++) {
- xh[i] = war[i];
- war[i] = x[i];
- }
-
- if (n < 2) {
- xh[0] *= 2;
- } else if (n == 2) {
- double xhold = sqrt3*(xh[0]+xh[1]);
- xh[1] = sqrt3*(xh[0]-xh[1]);
- xh[0] = xhold;
- } else {
- size_t np1 = n + 1;
- size_t ns2 = n / 2;
- x[0] = 0.0;
- for (size_t k = 0; k < ns2; k++) {
- size_t kc = np1 - k;
- double t1 = xh[k] - xh[kc];
- double t2 = was[k] * (xh[k] + xh[kc]);
- x[k+1] = t1 + t2;
- x[kc+1] = t2 - t1;
- }
- size_t modn = n % 2;
- if (modn != 0) {
- x[ns2+2] = 4.0 * xh[ns2+1];
- }
- rfftf1 (np1, x, xh, war, (size_t*)xxifac);
- xh[0] *= 0.5;
- for (size_t i = 2; i < n; i += 2) {
- xh[i-1] = -x[i];
- xh[i] = xh[i-2] + x[i-1];
- }
- if (modn == 0) {
- xh[n-1] = -x[n];
- }
- }
- for (size_t i = 0; i < n; i++) {
- x[i] = war[i];
- war[i] = xh[i];
- }
- }
-
- void sint (size_t n, double x[], double wsave[])
- {
- sint1 (
- n,
- x,
- wsave,
- wsave + n/2,
- wsave + n/2 + n + 1,
- wsave + n/2 + 2*n + 2
- );
- }
+ static void rfftf1(
+ size_t n,
+ double c[],
+ double ch[],
+ const double wa[],
+ const size_t ifac[]
+ ) {
+ size_t k1, l1=n, nf=ifac[1], iw=n-1;
+ double *p1=ch, *p2=c;
+
+ for(k1=1; k1<=nf;++k1)
+ {
+ size_t ip=ifac[nf-k1+2];
+ size_t ido=n / l1;
+ l1 /= ip;
+ iw-=(ip-1)*ido;
+ SWAP (p1,p2,double *);
+ if(ip==4)
+ radf4(ido, l1, p1, p2, wa+iw);
+ else if(ip==2)
+ radf2(ido, l1, p1, p2, wa+iw);
+ else if(ip==3)
+ radf3(ido, l1, p1, p2, wa+iw);
+ else if(ip==5)
+ radf5(ido, l1, p1, p2, wa+iw);
+ else
+ {
+ if (ido==1)
+ SWAP (p1,p2,double *);
+ radfg(ido, ip, l1, ido*l1, p1, p2, wa+iw);
+ SWAP (p1,p2,double *);
+ }
+ }
+ if (p1==c)
+ memcpy (c,ch,n*sizeof(double));
+ }
+
+ static void rfftb1(size_t n, double c[], double ch[], const double wa[],
+ const size_t ifac[])
+ {
+ size_t k1, l1=1, nf=ifac[1], iw=0;
+ double *p1=c, *p2=ch;
+
+ for(k1=1; k1<=nf; k1++)
+ {
+ size_t ip = ifac[k1+1],
+ ido= n/(ip*l1);
+ if(ip==4)
+ radb4(ido, l1, p1, p2, wa+iw);
+ else if(ip==2)
+ radb2(ido, l1, p1, p2, wa+iw);
+ else if(ip==3)
+ radb3(ido, l1, p1, p2, wa+iw);
+ else if(ip==5)
+ radb5(ido, l1, p1, p2, wa+iw);
+ else
+ {
+ radbg(ido, ip, l1, ido*l1, p1, p2, wa+iw);
+ if (ido!=1)
+ SWAP (p1,p2,double *);
+ }
+ SWAP (p1,p2,double *);
+ l1*=ip;
+ iw+=(ip-1)*ido;
+ }
+ if (p1!=c)
+ memcpy (c,ch,n*sizeof(double));
+ }
+
+ void rfftf(size_t n, double r[], double wsave[])
+ { if(n!=1) rfftf1(n, r, wsave, wsave+n,(size_t*)(wsave+2*n)); }
+
+ void rfftb(size_t n, double r[], double wsave[])
+ { if(n!=1) rfftb1(n, r, wsave, wsave+n,(size_t*)(wsave+2*n)); }
+
+ static void rffti1(size_t n, double wa[], size_t ifac[])
+ {
+ static const size_t ntryh[4]={4,2,3,5};
+ static const double twopi=6.28318530717958647692;
+ size_t i, j, k, fi;
+
+ double argh=twopi/n;
+ size_t is=0, l1=1;
+ factorize (n,ntryh,4,ifac);
+ for (k=1; k<ifac[1]; k++)
+ {
+ size_t ip=ifac[k+1],
+ ido=n/(l1*ip);
+ for (j=1; j<ip; ++j)
+ {
+ double argld=j*l1*argh;
+ for(i=is,fi=1; i<=ido+is-3; i+=2,++fi)
+ {
+ double arg=fi*argld;
+ wa[i ]=cos(arg);
+ wa[i+1]=sin(arg);
+ }
+ is+=ido;
+ }
+ l1*=ip;
+ }
+ }
+
+ void rffti(size_t n, double wsave[])
+ {
+ if (n!=1) {
+ rffti1(n, wsave+n,(size_t*)(wsave+2*n));
+ }
+ }
+
+ const double pi = 3.14159265358979;
+ void sinti (size_t n, double wsave[])
+ {
+ if (n <= 1) return;
+ size_t ns2 = n / 2;
+ double dt = pi / double (n+1);
+ for (size_t k = 0; k < ns2; k++) {
+ wsave[k] = 2.0 * sin ((k+1) * dt);
+ }
+ rffti (n+1, wsave+ns2);
+ }
+
+ /*
+ with n = 20
+ +0 -> was wsave[0..9]
+ +10 -> xh wsave[10..30]
+ +31 -> x wsave[31..51]
+ +52 -> xxifac wsave[52..]
+ */
+
+ void sint1 (size_t n, double war[], double was[], double xh[], double x[], double xxifac[])
+ {
+ double sqrt3 = 1.73205080756888;
+
+ for (size_t i = 0; i < n; i++) {
+ xh[i] = war[i];
+ war[i] = x[i];
+ }
+
+ if (n < 2) {
+ xh[0] *= 2;
+ } else if (n == 2) {
+ double xhold = sqrt3*(xh[0]+xh[1]);
+ xh[1] = sqrt3*(xh[0]-xh[1]);
+ xh[0] = xhold;
+ } else {
+ size_t np1 = n + 1;
+ size_t ns2 = n / 2;
+ x[0] = 0.0;
+ for (size_t k = 0; k < ns2; k++) {
+ size_t kc = np1 - k;
+ double t1 = xh[k] - xh[kc];
+ double t2 = was[k] * (xh[k] + xh[kc]);
+ x[k+1] = t1 + t2;
+ x[kc+1] = t2 - t1;
+ }
+ size_t modn = n % 2;
+ if (modn != 0) {
+ x[ns2+2] = 4.0 * xh[ns2+1];
+ }
+ rfftf1 (np1, x, xh, war, (size_t*)xxifac);
+ xh[0] *= 0.5;
+ for (size_t i = 2; i < n; i += 2) {
+ xh[i-1] = -x[i];
+ xh[i] = xh[i-2] + x[i-1];
+ }
+ if (modn == 0) {
+ xh[n-1] = -x[n];
+ }
+ }
+ for (size_t i = 0; i < n; i++) {
+ x[i] = war[i];
+ war[i] = xh[i];
+ }
+ }
+
+ void sint (size_t n, double x[], double wsave[])
+ {
+ sint1 (
+ n,
+ x,
+ wsave,
+ wsave + n/2,
+ wsave + n/2 + n + 1,
+ wsave + n/2 + 2*n + 2
+ );
+ }
#ifdef __cplusplus
}
#endif
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/fftpack.h b/ippl/src/FFT/fftpack.h
index 0d6b2de9b58fc3939cb36e57d782bd9b131d7a09..9cfccd053259873dc8f59a592e2a1cca0403f40d 100644
--- a/ippl/src/FFT/fftpack.h
+++ b/ippl/src/FFT/fftpack.h
@@ -1,3 +1,13 @@
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
/*
* This file is part of libfftpack.
*
@@ -32,33 +42,38 @@
(reformatted by joerg arndt)
reformatted and slightly enhanced by Martin Reinecke (2004)
- */
+*/
#ifndef PLANCK_FFTPACK_H
#define PLANCK_FFTPACK_H
-/* #include "c_utils.h" */
-
#ifdef __cplusplus
extern "C" {
#endif
/*! forward complex transform */
-void cfftf(size_t N, double complex_data[], double wrk[]);
+ void cfftf(size_t N, double complex_data[], double wrk[]);
/*! backward complex transform */
-void cfftb(size_t N, double complex_data[], double wrk[]);
+ void cfftb(size_t N, double complex_data[], double wrk[]);
/*! initializer for complex transforms */
-void cffti(size_t N, double wrk[]);
+ void cffti(size_t N, double wrk[]);
/*! forward real transform */
-void rfftf(size_t N, double data[], double wrk[]);
+ void rfftf(size_t N, double data[], double wrk[]);
/*! backward real transform */
-void rfftb(size_t N, double data[], double wrk[]);
+ void rfftb(size_t N, double data[], double wrk[]);
/*! initializer for real transforms */
-void rffti(size_t N, double wrk[]);
+ void rffti(size_t N, double wrk[]);
#ifdef __cplusplus
}
#endif
#endif
+
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/fftpack_FFT.h b/ippl/src/FFT/fftpack_FFT.h
index ccfc72d14449eb69b225c881be45e5550e63583d..2e289fc76bd16adbd2006a224ac0a40ad4167884 100644
--- a/ippl/src/FFT/fftpack_FFT.h
+++ b/ippl/src/FFT/fftpack_FFT.h
@@ -1,12 +1,19 @@
-// -*- C++ -*-
-/***************************************************************************
- *
- * The IPPL Framework
- *
- *
- * Visit http://people.web.psi.ch/adelmann/ for more details
- *
- ***************************************************************************/
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
+/*
+ Prototypes for accessing Fortran 1D FFT routines from
+ Netlib, and definitions for templated class FFTPACK, which acts as an
+ FFT engine for the FFT class, providing storage for trigonometric
+ information and performing the 1D FFTs as needed.
+*/
#ifndef IPPL_FFT_FFTPACK_FFT_H
#define IPPL_FFT_FFTPACK_FFT_H
@@ -14,39 +21,30 @@
#include "Utility/PAssert.h"
#include "Utility/IpplInfo.h"
-
-/**************************************************************************
- * fftpack_FFT.h: Prototypes for accessing Fortran 1D FFT routines from
- * Netlib, and definitions for templated class FFTPACK, which acts as an
- * FFT engine for the FFT class, providing storage for trigonometric
- * information and performing the 1D FFTs as needed.
- **************************************************************************/
-
-
// FFTPACK function prototypes for Fortran routines
extern "C" {
- // double-precision CC FFT
- void cffti (size_t n, double& wsave);
- void cfftf (size_t n, double& r, double& wsave);
- void cfftb (size_t n, double& r, double& wsave);
- // double-precision RC FFT
- void rffti (size_t n, double& wsave);
- void rfftf (size_t n, double& r, double& wsave);
- void rfftb (size_t n, double& r, double& wsave);
- // double-precision sine transform
- void sinti (size_t n, double& wsave);
- void sint (size_t n, double& r, double& wsave);
- // single-precision CC FFT
- void fcffti (size_t n, float& wsave);
- void fcfftf (size_t n, float& r, float& wsave);
- void fcfftb (size_t n, float& r, float& wsave);
- // single-precision RC FFT
- void frffti (size_t n, float& wsave);
- void frfftf (size_t n, float& r, float& wsave);
- void frfftb (size_t n, float& r, float& wsave);
- // single-precision sine transform
- void fsinti (size_t n, float& wsave);
- void fsint (size_t n, float& r, float& wsave);
+ // double-precision CC FFT
+ void cffti (size_t n, double& wsave);
+ void cfftf (size_t n, double& r, double& wsave);
+ void cfftb (size_t n, double& r, double& wsave);
+ // double-precision RC FFT
+ void rffti (size_t n, double& wsave);
+ void rfftf (size_t n, double& r, double& wsave);
+ void rfftb (size_t n, double& r, double& wsave);
+ // double-precision sine transform
+ void sinti (size_t n, double& wsave);
+ void sint (size_t n, double& r, double& wsave);
+ // single-precision CC FFT
+ void fcffti (size_t n, float& wsave);
+ void fcfftf (size_t n, float& r, float& wsave);
+ void fcfftb (size_t n, float& r, float& wsave);
+ // single-precision RC FFT
+ void frffti (size_t n, float& wsave);
+ void frfftf (size_t n, float& r, float& wsave);
+ void frfftb (size_t n, float& r, float& wsave);
+ // single-precision sine transform
+ void fsinti (size_t n, float& wsave);
+ void fsint (size_t n, float& r, float& wsave);
}
@@ -60,21 +58,20 @@ template <>
class FFTPACK_wrap<float> {
public:
- // interface functions used by class FFTPACK
-
- // initialization functions for CC FFT, RC FFT, and sine transform
- static void ccffti(size_t n, float* wsave) { fcffti (n, *wsave); }
- static void rcffti(size_t n, float* wsave) { frffti (n, *wsave); }
- static void rrffti(size_t n, float* wsave) { fsinti (n, *wsave); }
- // forward and backward CC FFT
- static void ccfftf(size_t n, float* r, float* wsave) { fcfftf (n, *r, *wsave); }
- static void ccfftb(size_t n, float* r, float* wsave) { fcfftb (n, *r, *wsave); }
- // forward and backward RC FFT
- static void rcfftf(size_t n, float* r, float* wsave) { frfftf (n, *r, *wsave); }
- static void rcfftb(size_t n, float* r, float* wsave) { frfftb (n, *r, *wsave); }
- // sine transform
- static void rrfft(size_t n, float* r, float* wsave) { fsint (n, *r, *wsave); }
-
+ // interface functions used by class FFTPACK
+
+ // initialization functions for CC FFT, RC FFT, and sine transform
+ static void ccffti(size_t n, float* wsave) { fcffti (n, *wsave); }
+ static void rcffti(size_t n, float* wsave) { frffti (n, *wsave); }
+ static void rrffti(size_t n, float* wsave) { fsinti (n, *wsave); }
+ // forward and backward CC FFT
+ static void ccfftf(size_t n, float* r, float* wsave) { fcfftf (n, *r, *wsave); }
+ static void ccfftb(size_t n, float* r, float* wsave) { fcfftb (n, *r, *wsave); }
+ // forward and backward RC FFT
+ static void rcfftf(size_t n, float* r, float* wsave) { frfftf (n, *r, *wsave); }
+ static void rcfftb(size_t n, float* r, float* wsave) { frfftb (n, *r, *wsave); }
+ // sine transform
+ static void rrfft(size_t n, float* r, float* wsave) { fsint (n, *r, *wsave); }
};
// Specialization for double
@@ -82,20 +79,20 @@ template <>
class FFTPACK_wrap<double> {
public:
- // interface functions used by class FFTPACK
-
- // initialization functions for CC FFT, RC FFT, and sine transform
- static void ccffti(size_t n, double* wsave) { cffti (n, *wsave); }
- static void rcffti(size_t n, double* wsave) { rffti (n, *wsave); }
- static void rrffti(size_t n, double* wsave) { sinti (n, *wsave); }
- // forward and backward CC FFT
- static void ccfftf(size_t n, double* r, double* wsave) {cfftf (n, *r, *wsave);}
- static void ccfftb(size_t n, double* r, double* wsave) {cfftb (n, *r, *wsave);}
- // forward and backward RC FFT
- static void rcfftf(size_t n, double* r, double* wsave) {rfftf (n, *r, *wsave);}
- static void rcfftb(size_t n, double* r, double* wsave) {rfftb (n, *r, *wsave);}
- // sine transform
- static void rrfft(size_t n, double* r, double* wsave) { sint (n, *r, *wsave); }
+ // interface functions used by class FFTPACK
+
+ // initialization functions for CC FFT, RC FFT, and sine transform
+ static void ccffti(size_t n, double* wsave) { cffti (n, *wsave); }
+ static void rcffti(size_t n, double* wsave) { rffti (n, *wsave); }
+ static void rrffti(size_t n, double* wsave) { sinti (n, *wsave); }
+ // forward and backward CC FFT
+ static void ccfftf(size_t n, double* r, double* wsave) {cfftf (n, *r, *wsave);}
+ static void ccfftb(size_t n, double* r, double* wsave) {cfftb (n, *r, *wsave);}
+ // forward and backward RC FFT
+ static void rcfftf(size_t n, double* r, double* wsave) {rfftf (n, *r, *wsave);}
+ static void rcfftb(size_t n, double* r, double* wsave) {rfftb (n, *r, *wsave);}
+ // sine transform
+ static void rrfft(size_t n, double* r, double* wsave) { sint (n, *r, *wsave); }
};
@@ -105,37 +102,33 @@ class FFTPACK {
public:
- // definition of complex type
-#ifdef IPPL_HAS_TEMPLATED_COMPLEX
- typedef std::complex<T> Complex_t;
-#else
- typedef complex Complex_t;
-#endif
+ // definition of complex type
+ typedef std::complex<T> Complex_t;
- // Trivial constructor. Do the real work in setup function.
- FFTPACK(void) {}
+ // Trivial constructor. Do the real work in setup function.
+ FFTPACK(void) {}
- // destructor
- ~FFTPACK(void);
+ // destructor
+ ~FFTPACK(void);
- // setup internal storage and prepare to perform FFTs
- // inputs are number of dimensions to transform, the transform types,
- // and the lengths of these dimensions.
- void setup(unsigned numTransformDims, const int* transformTypes,
- const int* axisLengths);
+ // setup internal storage and prepare to perform FFTs
+ // inputs are number of dimensions to transform, the transform types,
+ // and the lengths of these dimensions.
+ void setup(unsigned numTransformDims, const int* transformTypes,
+ const int* axisLengths);
- // invoke FFT on complex data for given dimension and direction
- void callFFT(unsigned transformDim, int direction, Complex_t* data);
+ // invoke FFT on complex data for given dimension and direction
+ void callFFT(unsigned transformDim, int direction, Complex_t* data);
- // invoke FFT on real data for given dimension and direction
- void callFFT(unsigned transformDim, int direction, T* data);
+ // invoke FFT on real data for given dimension and direction
+ void callFFT(unsigned transformDim, int direction, T* data);
private:
- unsigned numTransformDims_m; // number of dimensions to transform
- int* transformType_m; // transform type for each dimension
- int* axisLength_m; // length of each transform dimension
- T** trig_m; // trigonometric tables
+ unsigned numTransformDims_m; // number of dimensions to transform
+ int* transformType_m; // transform type for each dimension
+ int* axisLength_m; // length of each transform dimension
+ T** trig_m; // trigonometric tables
};
@@ -146,12 +139,12 @@ private:
template <class T>
inline
FFTPACK<T>::~FFTPACK(void) {
- // delete storage
- for (unsigned d=0; d<numTransformDims_m; ++d)
- delete [] trig_m[d];
- delete [] trig_m;
- delete [] transformType_m;
- delete [] axisLength_m;
+ // delete storage
+ for (unsigned d=0; d<numTransformDims_m; ++d)
+ delete [] trig_m[d];
+ delete [] trig_m;
+ delete [] transformType_m;
+ delete [] axisLength_m;
}
// setup internal storage to prepare for FFTs
@@ -160,39 +153,39 @@ inline void
FFTPACK<T>::setup(unsigned numTransformDims, const int* transformTypes,
const int* axisLengths) {
- // store transform types and lengths for each transform dim
- numTransformDims_m = numTransformDims;
- transformType_m = new int[numTransformDims_m];
- axisLength_m = new int[numTransformDims_m];
- unsigned d;
- for (d=0; d<numTransformDims_m; ++d) {
- transformType_m[d] = transformTypes[d];
- axisLength_m[d] = axisLengths[d];
- }
-
- // allocate and initialize trig table
- trig_m = new T*[numTransformDims_m];
- for (d=0; d<numTransformDims_m; ++d) {
- switch (transformType_m[d]) {
- case 0: // CC FFT
- trig_m[d] = new T[4 * axisLength_m[d] + 15];
- FFTPACK_wrap<T>::ccffti(axisLength_m[d], trig_m[d]);
- break;
- case 1: // RC FFT
- trig_m[d] = new T[2 * axisLength_m[d] + 15];
- FFTPACK_wrap<T>::rcffti(axisLength_m[d], trig_m[d]);
- break;
- case 2: // Sine transform
- trig_m[d] = new T[static_cast<int>(2.5 * axisLength_m[d] + 0.5) + 15];
- FFTPACK_wrap<T>::rrffti(axisLength_m[d], trig_m[d]);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
+ // store transform types and lengths for each transform dim
+ numTransformDims_m = numTransformDims;
+ transformType_m = new int[numTransformDims_m];
+ axisLength_m = new int[numTransformDims_m];
+ unsigned d;
+ for (d=0; d<numTransformDims_m; ++d) {
+ transformType_m[d] = transformTypes[d];
+ axisLength_m[d] = axisLengths[d];
+ }
+
+ // allocate and initialize trig table
+ trig_m = new T*[numTransformDims_m];
+ for (d=0; d<numTransformDims_m; ++d) {
+ switch (transformType_m[d]) {
+ case 0: // CC FFT
+ trig_m[d] = new T[4 * axisLength_m[d] + 15];
+ FFTPACK_wrap<T>::ccffti(axisLength_m[d], trig_m[d]);
+ break;
+ case 1: // RC FFT
+ trig_m[d] = new T[2 * axisLength_m[d] + 15];
+ FFTPACK_wrap<T>::rcffti(axisLength_m[d], trig_m[d]);
+ break;
+ case 2: // Sine transform
+ trig_m[d] = new T[static_cast<int>(2.5 * axisLength_m[d] + 0.5) + 15];
+ FFTPACK_wrap<T>::rrffti(axisLength_m[d], trig_m[d]);
+ break;
+ default:
+ ERRORMSG("Unknown transform type requested!!" << endl);
+ break;
+ }
}
- }
- return;
+ return;
}
// invoke FFT on complex data for given dimension and direction
@@ -201,59 +194,59 @@ inline void
FFTPACK<T>::callFFT(unsigned transformDim, int direction,
FFTPACK<T>::Complex_t* data) {
- // check transform dimension and direction arguments
- PAssert_LT(transformDim, numTransformDims_m);
- PAssert_EQ(std::abs(direction), 1);
+ // check transform dimension and direction arguments
+ PAssert_LT(transformDim, numTransformDims_m);
+ PAssert_EQ(std::abs(direction), 1);
- // cast complex number pointer to T* for calling Fortran routines
- T* rdata = reinterpret_cast<T*>(data);
+ // cast complex number pointer to T* for calling Fortran routines
+ T* rdata = reinterpret_cast<T*>(data);
- // branch on transform type for this dimension
- switch (transformType_m[transformDim]) {
- case 0: // CC FFT
- if (direction == +1) {
- // call forward complex-to-complex FFT
- FFTPACK_wrap<T>::ccfftf(axisLength_m[transformDim], rdata,
- trig_m[transformDim]);
- }
- else {
- // call backward complex-to-complex FFT
- FFTPACK_wrap<T>::ccfftb(axisLength_m[transformDim], rdata,
- trig_m[transformDim]);
- }
- break;
- case 1: // RC FFT
- if (direction == +1) {
- // call forward real-to-complex FFT
- FFTPACK_wrap<T>::rcfftf(axisLength_m[transformDim], rdata,
- trig_m[transformDim]);
- // rearrange output to conform with SGI format for complex result
- int clen = axisLength_m[transformDim]/2 + 1;
- data[clen-1] = Complex_t(imag(data[clen-2]),0.0);
- for (int i = clen-2; i > 0; --i)
- data[i] = Complex_t(imag(data[i-1]),real(data[i]));
- data[0] = Complex_t(real(data[0]),0.0);
- }
- else {
- // rearrange input to conform with Netlib format for complex modes
- int clen = axisLength_m[transformDim]/2 + 1;
- data[0] = Complex_t(real(data[0]),real(data[1]));
- for (int i = 1; i < clen-1; ++i)
- data[i] = Complex_t(imag(data[i]),real(data[i+1]));
- // call backward complex-to-real FFT
- FFTPACK_wrap<T>::rcfftb(axisLength_m[transformDim], rdata,
- trig_m[transformDim]);
+ // branch on transform type for this dimension
+ switch (transformType_m[transformDim]) {
+ case 0: // CC FFT
+ if (direction == +1) {
+ // call forward complex-to-complex FFT
+ FFTPACK_wrap<T>::ccfftf(axisLength_m[transformDim], rdata,
+ trig_m[transformDim]);
+ }
+ else {
+ // call backward complex-to-complex FFT
+ FFTPACK_wrap<T>::ccfftb(axisLength_m[transformDim], rdata,
+ trig_m[transformDim]);
+ }
+ break;
+ case 1: // RC FFT
+ if (direction == +1) {
+ // call forward real-to-complex FFT
+ FFTPACK_wrap<T>::rcfftf(axisLength_m[transformDim], rdata,
+ trig_m[transformDim]);
+ // rearrange output to conform with SGI format for complex result
+ int clen = axisLength_m[transformDim]/2 + 1;
+ data[clen-1] = Complex_t(imag(data[clen-2]),0.0);
+ for (int i = clen-2; i > 0; --i)
+ data[i] = Complex_t(imag(data[i-1]),real(data[i]));
+ data[0] = Complex_t(real(data[0]),0.0);
+ }
+ else {
+ // rearrange input to conform with Netlib format for complex modes
+ int clen = axisLength_m[transformDim]/2 + 1;
+ data[0] = Complex_t(real(data[0]),real(data[1]));
+ for (int i = 1; i < clen-1; ++i)
+ data[i] = Complex_t(imag(data[i]),real(data[i+1]));
+ // call backward complex-to-real FFT
+ FFTPACK_wrap<T>::rcfftb(axisLength_m[transformDim], rdata,
+ trig_m[transformDim]);
+ }
+ break;
+ case 2: // Sine transform
+ ERRORMSG("Input for real-to-real FFT should be real!!" << endl);
+ break;
+ default:
+ ERRORMSG("Unknown transform type requested!!" << endl);
+ break;
}
- break;
- case 2: // Sine transform
- ERRORMSG("Input for real-to-real FFT should be real!!" << endl);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
- }
-
- return;
+
+ return;
}
// invoke FFT on real data for given dimension and direction
@@ -261,37 +254,36 @@ template <class T>
inline void
FFTPACK<T>::callFFT(unsigned transformDim, int direction, T* data) {
- // check transform dimension and direction arguments
- PAssert_LT(transformDim, numTransformDims_m);
- PAssert_EQ(std::abs(direction), 1);
-
- // branch on transform type for this dimension
- switch (transformType_m[transformDim]) {
- case 0: // CC FFT
- ERRORMSG("Input for complex-to-complex FFT should be complex!!" << endl);
- break;
- case 1: // RC FFT
- ERRORMSG("real-to-complex FFT uses complex input!!" << endl);
- break;
- case 2: // Sine transform
- // invoke the real-to-real transform on the data
- FFTPACK_wrap<T>::rrfft(axisLength_m[transformDim], data,
- trig_m[transformDim]);
- break;
- default:
- ERRORMSG("Unknown transform type requested!!" << endl);
- break;
- }
-
- return;
-}
+ // check transform dimension and direction arguments
+ PAssert_LT(transformDim, numTransformDims_m);
+ PAssert_EQ(std::abs(direction), 1);
+ // branch on transform type for this dimension
+ switch (transformType_m[transformDim]) {
+ case 0: // CC FFT
+ ERRORMSG("Input for complex-to-complex FFT should be complex!!" << endl);
+ break;
+ case 1: // RC FFT
+ ERRORMSG("real-to-complex FFT uses complex input!!" << endl);
+ break;
+ case 2: // Sine transform
+ // invoke the real-to-real transform on the data
+ FFTPACK_wrap<T>::rrfft(axisLength_m[transformDim], data,
+ trig_m[transformDim]);
+ break;
+ default:
+ ERRORMSG("Unknown transform type requested!!" << endl);
+ break;
+ }
+ return;
+}
#endif // IPPL_FFT_FFTPACK_FFT_H
-/***************************************************************************
- * $RCSfile: fftpack_FFT.h,v $ $Author: adelmann $
- * $Revision: 1.1.1.1 $ $Date: 2003/01/23 07:40:26 $
- * IPPL_VERSION_ID: $Id: fftpack_FFT.h,v 1.1.1.1 2003/01/23 07:40:26 adelmann Exp $
- ***************************************************************************/
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/FFT/fftpack_inc.c b/ippl/src/FFT/fftpack_inc.c
index 55d0ac579d7199a5eb4c1a85ad804dabdf906c5c..d17960d5a09d7b37247d724856ecc5537957fe8f 100644
--- a/ippl/src/FFT/fftpack_inc.c
+++ b/ippl/src/FFT/fftpack_inc.c
@@ -1,3 +1,13 @@
+//
+// IPPL FFT
+//
+// Copyright (c) 2008-2018
+// Paul Scherrer Institut, Villigen PSI, Switzerland
+// All rights reserved.
+//
+// OPAL is licensed under GNU GPL version 3.
+//
+
/*
* This file is part of libfftpack.
*
@@ -28,7 +38,7 @@
(Version 4, 1985).
C port by Martin Reinecke (2010)
- */
+*/
#ifdef BACKWARD
#define PSIGN +
@@ -43,264 +53,271 @@
#endif
static void X(2) (size_t ido, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=2;
- size_t k,i;
- cmplx t;
- if (ido==1)
- for (k=0;k<l1;++k)
- PMC (CH(0,k,0),CH(0,k,1),CC(0,0,k),CC(0,1,k))
- else
- for (k=0;k<l1;++k)
- for (i=0;i<ido;++i)
- {
- PMC (CH(i,k,0),t,CC(i,0,k),CC(i,1,k))
- MULPMSIGNC (CH(i,k,1),WA(0,i),t)
- }
- }
+ const cmplx *wa)
+{
+ const size_t cdim=2;
+ size_t k,i;
+ cmplx t;
+ if (ido==1)
+ for (k=0;k<l1;++k)
+ PMC (CH(0,k,0),CH(0,k,1),CC(0,0,k),CC(0,1,k))
+ else
+ for (k=0;k<l1;++k)
+ for (i=0;i<ido;++i)
+ {
+ PMC (CH(i,k,0),t,CC(i,0,k),CC(i,1,k))
+ MULPMSIGNC (CH(i,k,1),WA(0,i),t)
+ }
+}
static void X(3)(size_t ido, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=3;
- static const double taur=-0.5, taui= PSIGN 0.86602540378443864676;
- size_t i, k;
- cmplx c2, c3, d2, d3, t2;
+ const cmplx *wa)
+{
+ const size_t cdim=3;
+ static const double taur=-0.5, taui= PSIGN 0.86602540378443864676;
+ size_t i, k;
+ cmplx c2, c3, d2, d3, t2;
- if (ido==1)
- for (k=0; k<l1; ++k)
- {
- PMC (t2,c3,CC(0,1,k),CC(0,2,k))
- ADDC (CH(0,k,0),t2,CC(0,0,k))
- SCALEC(t2,taur)
- ADDC(c2,CC(0,0,k),t2)
- SCALEC(c3,taui)
- CONJFLIPC(c3)
- PMC(CH(0,k,1),CH(0,k,2),c2,c3)
- }
- else
- for (k=0; k<l1; ++k)
- for (i=0; i<ido; ++i)
+ if (ido==1)
+ for (k=0; k<l1; ++k)
{
- PMC (t2,c3,CC(i,1,k),CC(i,2,k))
- ADDC (CH(i,k,0),t2,CC(i,0,k))
- SCALEC(t2,taur)
- ADDC(c2,CC(i,0,k),t2)
- SCALEC(c3,taui)
- CONJFLIPC(c3)
- PMC(d2,d3,c2,c3)
- MULPMSIGNC(CH(i,k,1),WA(0,i),d2)
- MULPMSIGNC(CH(i,k,2),WA(1,i),d3)
- }
- }
+ PMC (t2,c3,CC(0,1,k),CC(0,2,k))
+ ADDC (CH(0,k,0),t2,CC(0,0,k))
+ SCALEC(t2,taur)
+ ADDC(c2,CC(0,0,k),t2)
+ SCALEC(c3,taui)
+ CONJFLIPC(c3)
+ PMC(CH(0,k,1),CH(0,k,2),c2,c3)
+ }
+ else
+ for (k=0; k<l1; ++k)
+ for (i=0; i<ido; ++i)
+ {
+ PMC (t2,c3,CC(i,1,k),CC(i,2,k))
+ ADDC (CH(i,k,0),t2,CC(i,0,k))
+ SCALEC(t2,taur)
+ ADDC(c2,CC(i,0,k),t2)
+ SCALEC(c3,taui)
+ CONJFLIPC(c3)
+ PMC(d2,d3,c2,c3)
+ MULPMSIGNC(CH(i,k,1),WA(0,i),d2)
+ MULPMSIGNC(CH(i,k,2),WA(1,i),d3)
+ }
+}
static void X(4)(size_t ido, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=4;
- size_t i, k;
- cmplx c2, c3, c4, t1, t2, t3, t4;
+ const cmplx *wa)
+{
+ const size_t cdim=4;
+ size_t i, k;
+ cmplx c2, c3, c4, t1, t2, t3, t4;
- if (ido==1)
- for (k=0; k<l1; ++k)
- {
- PMC(t2,t1,CC(0,0,k),CC(0,2,k))
- PMC(t3,t4,CC(0,1,k),CC(0,3,k))
- CONJFLIPC(t4)
- PMC(CH(0,k,0),CH(0,k,2),t2,t3)
- PMSIGNC (CH(0,k,1),CH(0,k,3),t1,t4)
- }
- else
- for (k=0; k<l1; ++k)
- for (i=0; i<ido; ++i)
+ if (ido==1)
+ for (k=0; k<l1; ++k)
{
- PMC(t2,t1,CC(i,0,k),CC(i,2,k))
- PMC(t3,t4,CC(i,1,k),CC(i,3,k))
- CONJFLIPC(t4)
- PMC(CH(i,k,0),c3,t2,t3)
- PMSIGNC (c2,c4,t1,t4)
- MULPMSIGNC (CH(i,k,1),WA(0,i),c2)
- MULPMSIGNC (CH(i,k,2),WA(1,i),c3)
- MULPMSIGNC (CH(i,k,3),WA(2,i),c4)
- }
- }
+ PMC(t2,t1,CC(0,0,k),CC(0,2,k))
+ PMC(t3,t4,CC(0,1,k),CC(0,3,k))
+ CONJFLIPC(t4)
+ PMC(CH(0,k,0),CH(0,k,2),t2,t3)
+ PMSIGNC (CH(0,k,1),CH(0,k,3),t1,t4)
+ }
+ else
+ for (k=0; k<l1; ++k)
+ for (i=0; i<ido; ++i)
+ {
+ PMC(t2,t1,CC(i,0,k),CC(i,2,k))
+ PMC(t3,t4,CC(i,1,k),CC(i,3,k))
+ CONJFLIPC(t4)
+ PMC(CH(i,k,0),c3,t2,t3)
+ PMSIGNC (c2,c4,t1,t4)
+ MULPMSIGNC (CH(i,k,1),WA(0,i),c2)
+ MULPMSIGNC (CH(i,k,2),WA(1,i),c3)
+ MULPMSIGNC (CH(i,k,3),WA(2,i),c4)
+ }
+}
static void X(5)(size_t ido, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=5;
- static const double tr11= 0.3090169943749474241,
- ti11= PSIGN 0.95105651629515357212,
- tr12=-0.8090169943749474241,
- ti12= PSIGN 0.58778525229247312917;
- size_t i, k;
- cmplx c2, c3, c4, c5, d2, d3, d4, d5, t2, t3, t4, t5;
+ const cmplx *wa)
+{
+ const size_t cdim=5;
+ static const double tr11= 0.3090169943749474241,
+ ti11= PSIGN 0.95105651629515357212,
+ tr12=-0.8090169943749474241,
+ ti12= PSIGN 0.58778525229247312917;
+ size_t i, k;
+ cmplx c2, c3, c4, c5, d2, d3, d4, d5, t2, t3, t4, t5;
- if (ido==1)
- for (k=0; k<l1; ++k)
- {
- PMC (t2,t5,CC(0,1,k),CC(0,4,k))
- PMC (t3,t4,CC(0,2,k),CC(0,3,k))
- CH(0,k,0).r=CC(0,0,k).r+t2.r+t3.r;
- CH(0,k,0).i=CC(0,0,k).i+t2.i+t3.i;
- c2.r=CC(0,0,k).r+tr11*t2.r+tr12*t3.r;
- c2.i=CC(0,0,k).i+tr11*t2.i+tr12*t3.i;
- c3.r=CC(0,0,k).r+tr12*t2.r+tr11*t3.r;
- c3.i=CC(0,0,k).i+tr12*t2.i+tr11*t3.i;
- c5.r=ti11*t5.r+ti12*t4.r;
- c5.i=ti11*t5.i+ti12*t4.i;
- c4.r=ti12*t5.r-ti11*t4.r;
- c4.i=ti12*t5.i-ti11*t4.i;
- CONJFLIPC(c5)
- PMC(CH(0,k,1),CH(0,k,4),c2,c5)
- CONJFLIPC(c4)
- PMC(CH(0,k,2),CH(0,k,3),c3,c4)
- }
- else
- for (k=0; k<l1; ++k)
- for (i=0; i<ido; ++i)
+ if (ido==1)
+ for (k=0; k<l1; ++k)
{
- PMC (t2,t5,CC(i,1,k),CC(i,4,k))
- PMC (t3,t4,CC(i,2,k),CC(i,3,k))
- CH(i,k,0).r=CC(i,0,k).r+t2.r+t3.r;
- CH(i,k,0).i=CC(i,0,k).i+t2.i+t3.i;
- c2.r=CC(i,0,k).r+tr11*t2.r+tr12*t3.r;
- c2.i=CC(i,0,k).i+tr11*t2.i+tr12*t3.i;
- c3.r=CC(i,0,k).r+tr12*t2.r+tr11*t3.r;
- c3.i=CC(i,0,k).i+tr12*t2.i+tr11*t3.i;
- c5.r=ti11*t5.r+ti12*t4.r;
- c5.i=ti11*t5.i+ti12*t4.i;
- c4.r=ti12*t5.r-ti11*t4.r;
- c4.i=ti12*t5.i-ti11*t4.i;
- CONJFLIPC(c5)
- PMC(d2,d5,c2,c5)
- CONJFLIPC(c4)
- PMC(d3,d4,c3,c4)
- MULPMSIGNC (CH(i,k,1),WA(0,i),d2)
- MULPMSIGNC (CH(i,k,2),WA(1,i),d3)
- MULPMSIGNC (CH(i,k,3),WA(2,i),d4)
- MULPMSIGNC (CH(i,k,4),WA(3,i),d5)
- }
- }
+ PMC (t2,t5,CC(0,1,k),CC(0,4,k))
+ PMC (t3,t4,CC(0,2,k),CC(0,3,k))
+ CH(0,k,0).r=CC(0,0,k).r+t2.r+t3.r;
+ CH(0,k,0).i=CC(0,0,k).i+t2.i+t3.i;
+ c2.r=CC(0,0,k).r+tr11*t2.r+tr12*t3.r;
+ c2.i=CC(0,0,k).i+tr11*t2.i+tr12*t3.i;
+ c3.r=CC(0,0,k).r+tr12*t2.r+tr11*t3.r;
+ c3.i=CC(0,0,k).i+tr12*t2.i+tr11*t3.i;
+ c5.r=ti11*t5.r+ti12*t4.r;
+ c5.i=ti11*t5.i+ti12*t4.i;
+ c4.r=ti12*t5.r-ti11*t4.r;
+ c4.i=ti12*t5.i-ti11*t4.i;
+ CONJFLIPC(c5)
+ PMC(CH(0,k,1),CH(0,k,4),c2,c5)
+ CONJFLIPC(c4)
+ PMC(CH(0,k,2),CH(0,k,3),c3,c4)
+ }
+ else
+ for (k=0; k<l1; ++k)
+ for (i=0; i<ido; ++i)
+ {
+ PMC (t2,t5,CC(i,1,k),CC(i,4,k))
+ PMC (t3,t4,CC(i,2,k),CC(i,3,k))
+ CH(i,k,0).r=CC(i,0,k).r+t2.r+t3.r;
+ CH(i,k,0).i=CC(i,0,k).i+t2.i+t3.i;
+ c2.r=CC(i,0,k).r+tr11*t2.r+tr12*t3.r;
+ c2.i=CC(i,0,k).i+tr11*t2.i+tr12*t3.i;
+ c3.r=CC(i,0,k).r+tr12*t2.r+tr11*t3.r;
+ c3.i=CC(i,0,k).i+tr12*t2.i+tr11*t3.i;
+ c5.r=ti11*t5.r+ti12*t4.r;
+ c5.i=ti11*t5.i+ti12*t4.i;
+ c4.r=ti12*t5.r-ti11*t4.r;
+ c4.i=ti12*t5.i-ti11*t4.i;
+ CONJFLIPC(c5)
+ PMC(d2,d5,c2,c5)
+ CONJFLIPC(c4)
+ PMC(d3,d4,c3,c4)
+ MULPMSIGNC (CH(i,k,1),WA(0,i),d2)
+ MULPMSIGNC (CH(i,k,2),WA(1,i),d3)
+ MULPMSIGNC (CH(i,k,3),WA(2,i),d4)
+ MULPMSIGNC (CH(i,k,4),WA(3,i),d5)
+ }
+}
static void X(6)(size_t ido, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=6;
- static const double taui= PSIGN 0.86602540378443864676;
- cmplx ta1,ta2,ta3,a0,a1,a2,tb1,tb2,tb3,b0,b1,b2,d1,d2,d3,d4,d5;
- size_t i, k;
+ const cmplx *wa)
+{
+ const size_t cdim=6;
+ static const double taui= PSIGN 0.86602540378443864676;
+ cmplx ta1,ta2,ta3,a0,a1,a2,tb1,tb2,tb3,b0,b1,b2,d1,d2,d3,d4,d5;
+ size_t i, k;
- if (ido==1)
- for (k=0; k<l1; ++k)
- {
- PMC(ta1,ta3,CC(0,2,k),CC(0,4,k))
- ta2.r = CC(0,0,k).r - .5*ta1.r;
- ta2.i = CC(0,0,k).i - .5*ta1.i;
- SCALEC(ta3,taui)
- ADDC(a0,CC(0,0,k),ta1)
- CONJFLIPC(ta3)
- PMC(a1,a2,ta2,ta3)
- PMC(tb1,tb3,CC(0,5,k),CC(0,1,k))
- tb2.r = CC(0,3,k).r - .5*tb1.r;
- tb2.i = CC(0,3,k).i - .5*tb1.i;
- SCALEC(tb3,taui)
- ADDC(b0,CC(0,3,k),tb1)
- CONJFLIPC(tb3)
- PMC(b1,b2,tb2,tb3)
- PMC(CH(0,k,0),CH(0,k,3),a0,b0)
- PMC(CH(0,k,4),CH(0,k,1),a1,b1)
- PMC(CH(0,k,2),CH(0,k,5),a2,b2)
- }
- else
- for (k=0; k<l1; ++k)
- for (i=0; i<ido; ++i)
+ if (ido==1)
+ for (k=0; k<l1; ++k)
{
- PMC(ta1,ta3,CC(i,2,k),CC(i,4,k))
- ta2.r = CC(i,0,k).r - .5*ta1.r;
- ta2.i = CC(i,0,k).i - .5*ta1.i;
- SCALEC(ta3,taui)
- ADDC(a0,CC(i,0,k),ta1)
- CONJFLIPC(ta3)
- PMC(a1,a2,ta2,ta3)
- PMC(tb1,tb3,CC(i,5,k),CC(i,1,k))
- tb2.r = CC(i,3,k).r - .5*tb1.r;
- tb2.i = CC(i,3,k).i - .5*tb1.i;
- SCALEC(tb3,taui)
- ADDC(b0,CC(i,3,k),tb1)
- CONJFLIPC(tb3)
- PMC(b1,b2,tb2,tb3)
- PMC(CH(i,k,0),d3,a0,b0)
- PMC(d4,d1,a1,b1)
- PMC(d2,d5,a2,b2)
- MULPMSIGNC (CH(i,k,1),WA(0,i),d1)
- MULPMSIGNC (CH(i,k,2),WA(1,i),d2)
- MULPMSIGNC (CH(i,k,3),WA(2,i),d3)
- MULPMSIGNC (CH(i,k,4),WA(3,i),d4)
- MULPMSIGNC (CH(i,k,5),WA(4,i),d5)
- }
- }
+ PMC(ta1,ta3,CC(0,2,k),CC(0,4,k))
+ ta2.r = CC(0,0,k).r - .5*ta1.r;
+ ta2.i = CC(0,0,k).i - .5*ta1.i;
+ SCALEC(ta3,taui)
+ ADDC(a0,CC(0,0,k),ta1)
+ CONJFLIPC(ta3)
+ PMC(a1,a2,ta2,ta3)
+ PMC(tb1,tb3,CC(0,5,k),CC(0,1,k))
+ tb2.r = CC(0,3,k).r - .5*tb1.r;
+ tb2.i = CC(0,3,k).i - .5*tb1.i;
+ SCALEC(tb3,taui)
+ ADDC(b0,CC(0,3,k),tb1)
+ CONJFLIPC(tb3)
+ PMC(b1,b2,tb2,tb3)
+ PMC(CH(0,k,0),CH(0,k,3),a0,b0)
+ PMC(CH(0,k,4),CH(0,k,1),a1,b1)
+ PMC(CH(0,k,2),CH(0,k,5),a2,b2)
+ }
+ else
+ for (k=0; k<l1; ++k)
+ for (i=0; i<ido; ++i)
+ {
+ PMC(ta1,ta3,CC(i,2,k),CC(i,4,k))
+ ta2.r = CC(i,0,k).r - .5*ta1.r;
+ ta2.i = CC(i,0,k).i - .5*ta1.i;
+ SCALEC(ta3,taui)
+ ADDC(a0,CC(i,0,k),ta1)
+ CONJFLIPC(ta3)
+ PMC(a1,a2,ta2,ta3)
+ PMC(tb1,tb3,CC(i,5,k),CC(i,1,k))
+ tb2.r = CC(i,3,k).r - .5*tb1.r;
+ tb2.i = CC(i,3,k).i - .5*tb1.i;
+ SCALEC(tb3,taui)
+ ADDC(b0,CC(i,3,k),tb1)
+ CONJFLIPC(tb3)
+ PMC(b1,b2,tb2,tb3)
+ PMC(CH(i,k,0),d3,a0,b0)
+ PMC(d4,d1,a1,b1)
+ PMC(d2,d5,a2,b2)
+ MULPMSIGNC (CH(i,k,1),WA(0,i),d1)
+ MULPMSIGNC (CH(i,k,2),WA(1,i),d2)
+ MULPMSIGNC (CH(i,k,3),WA(2,i),d3)
+ MULPMSIGNC (CH(i,k,4),WA(3,i),d4)
+ MULPMSIGNC (CH(i,k,5),WA(4,i),d5)
+ }
+}
static void X(g)(size_t ido, size_t ip, size_t l1, const cmplx *cc, cmplx *ch,
- const cmplx *wa)
- {
- const size_t cdim=ip;
- cmplx *tarr=RALLOC(cmplx,2*ip);
- cmplx *ccl=tarr, *wal=tarr+ip;
- size_t i,j,k,l,jc,lc;
- size_t ipph = (ip+1)/2;
+ const cmplx *wa)
+{
+ const size_t cdim=ip;
+ cmplx *tarr=RALLOC(cmplx,2*ip);
+ cmplx *ccl=tarr, *wal=tarr+ip;
+ size_t i,j,k,l,jc,lc;
+ size_t ipph = (ip+1)/2;
- for (i=1; i<ip; ++i)
- wal[i]=wa[ido*(i-1)];
- for (k=0; k<l1; ++k)
- for (i=0; i<ido; ++i)
- {
- cmplx s=CC(i,0,k);
- ccl[0] = CC(i,0,k);
- for(j=1,jc=ip-1; j<ipph; ++j,--jc)
- {
- PMC (ccl[j],ccl[jc],CC(i,j,k),CC(i,jc,k))
- ADDC (s,s,ccl[j])
- }
- CH(i,k,0) = s;
- for (j=1, jc=ip-1; j<=ipph; ++j,--jc)
+ for (i=1; i<ip; ++i)
+ wal[i]=wa[ido*(i-1)];
+ for (k=0; k<l1; ++k)
+ for (i=0; i<ido; ++i)
{
- cmplx abr=ccl[0], abi={0.,0.};
- size_t iang=0;
- for (l=1,lc=ip-1; l<ipph; ++l,--lc)
- {
- iang+=j;
- if (iang>ip) iang-=ip;
- abr.r += ccl[l ].r*wal[iang].r;
- abr.i += ccl[l ].i*wal[iang].r;
- abi.r += ccl[lc].r*wal[iang].i;
- abi.i += ccl[lc].i*wal[iang].i;
- }
+ cmplx s=CC(i,0,k);
+ ccl[0] = CC(i,0,k);
+ for(j=1,jc=ip-1; j<ipph; ++j,--jc)
+ {
+ PMC (ccl[j],ccl[jc],CC(i,j,k),CC(i,jc,k))
+ ADDC (s,s,ccl[j])
+ }
+ CH(i,k,0) = s;
+ for (j=1, jc=ip-1; j<=ipph; ++j,--jc)
+ {
+ cmplx abr=ccl[0], abi={0.,0.};
+ size_t iang=0;
+ for (l=1,lc=ip-1; l<ipph; ++l,--lc)
+ {
+ iang+=j;
+ if (iang>ip) iang-=ip;
+ abr.r += ccl[l ].r*wal[iang].r;
+ abr.i += ccl[l ].i*wal[iang].r;
+ abi.r += ccl[lc].r*wal[iang].i;
+ abi.i += ccl[lc].i*wal[iang].i;
+ }
#ifndef BACKWARD
- { abi.i=-abi.i; abi.r=-abi.r; }
+ { abi.i=-abi.i; abi.r=-abi.r; }
#endif
- CONJFLIPC(abi)
- PMC(CH(i,k,j),CH(i,k,jc),abr,abi)
+ CONJFLIPC(abi)
+ PMC(CH(i,k,j),CH(i,k,jc),abr,abi)
+ }
}
- }
- DEALLOC(tarr);
+ DEALLOC(tarr);
- if (ido==1) return;
+ if (ido==1) return;
- for (j=1; j<ip; ++j)
- for (k=0; k<l1; ++k)
- {
- size_t idij=(j-1)*ido+1;
- for(i=1; i<ido; ++i, ++idij)
+ for (j=1; j<ip; ++j)
+ for (k=0; k<l1; ++k)
{
- cmplx t=CH(i,k,j);
- MULPMSIGNC (CH(i,k,j),wa[idij],t)
+ size_t idij=(j-1)*ido+1;
+ for(i=1; i<ido; ++i, ++idij)
+ {
+ cmplx t=CH(i,k,j);
+ MULPMSIGNC (CH(i,k,j),wa[idij],t)
+ }
}
- }
- }
+}
#undef PSIGN
#undef PMSIGNC
#undef MULPMSIGNC
+
+// vi: set et ts=4 sw=4 sts=4:
+// Local Variables:
+// mode:c
+// c-basic-offset: 4
+// indent-tabs-mode:nil
+// End:
diff --git a/ippl/src/Message/Message.h b/ippl/src/Message/Message.h
index b8d81355ae1bd5fb80580725c278b584d5cfb28e..392540111fe8b07fde4c4554ae7bdcdcf543e41f 100644
--- a/ippl/src/Message/Message.h
+++ b/ippl/src/Message/Message.h
@@ -97,7 +97,6 @@ DEFINE_ALL_BUILTIN_TRAIT_CLASS(long long)
DEFINE_ALL_BUILTIN_TRAIT_CLASS(float)
DEFINE_ALL_BUILTIN_TRAIT_CLASS(double)
DEFINE_ALL_BUILTIN_TRAIT_CLASS(dcomplex)
-DEFINE_ALL_BUILTIN_TRAIT_CLASS(fComplex)
/////////////////////////////////////////////////////////////////////
// a class to put single items into a Message, which can be specialized
diff --git a/ippl/test/6dtrack/GTDistribution/makeranlib b/ippl/test/6dtrack/GTDistribution/makeranlib
index 8b4db33534b7e513a9ad1109e17cc5352ab073c3..d8259e58e793c86e3e870dcb3ff673099a51e21f 100644
--- a/ippl/test/6dtrack/GTDistribution/makeranlib
+++ b/ippl/test/6dtrack/GTDistribution/makeranlib
@@ -1,2 +1,2 @@
-CC -fpermissive -Wno-deprecated -ftemplate-depth-80 -funroll-loops -fstrict-aliasing -g -I/users/adelmann/svnwork/ippl/src/expde -I/users/adelmann/svnwork/ippl/src -I/ufs/packages/hdf5/include -DTTProf -DOLDLAYOUT -DMAPS -DTTRACK -D__cplusplus -I. -I/users/adelmann/svnwork/classic/3.3/src -I/users/adelmann/svnwork/gsl/include -I/users/adelmann/svnwork/gte/GTConst -I/users/adelmann/svnwork/gte/GTUtilities -I./GTElements -I./GTDistribution -I./GTConfigure -I/users/adelmann/svnwork/gte/GTIntegrators -I/users/adelmann/svnwork/gte/GTUtilities -I./GTChargedParticles -I/users/adelmann/svnwork/gte/GTDataSink -I/users/adelmann/svnwork/gte/GTQRSolver -I/users/adelmann/svnwork/ippl/src/hdf5/H5ecloud -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX -DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG -DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM -Drestrict=__restrict__ -DNOCTAssert -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DEBUG -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX -DGTHDF5 -DPARALLEL_IO -I. -c ranlib.cc
+CC -fpermissive -Wno-deprecated -ftemplate-depth-80 -funroll-loops -fstrict-aliasing -g -I/users/adelmann/svnwork/ippl/src/expde -I/users/adelmann/svnwork/ippl/src -I/ufs/packages/hdf5/include -DTTProf -DOLDLAYOUT -DMAPS -DTTRACK -D__cplusplus -I. -I/users/adelmann/svnwork/classic/3.3/src -I/users/adelmann/svnwork/gsl/include -I/users/adelmann/svnwork/gte/GTConst -I/users/adelmann/svnwork/gte/GTUtilities -I./GTElements -I./GTDistribution -I./GTConfigure -I/users/adelmann/svnwork/gte/GTIntegrators -I/users/adelmann/svnwork/gte/GTUtilities -I./GTChargedParticles -I/users/adelmann/svnwork/gte/GTDataSink -I/users/adelmann/svnwork/gte/GTQRSolver -I/users/adelmann/svnwork/ippl/src/hdf5/H5ecloud -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG -DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM -Drestrict=__restrict__ -DNOCTAssert -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DEBUG -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX -DGTHDF5 -DPARALLEL_IO -I. -c ranlib.cc
diff --git a/ippl/test/6dtrack/Makefile b/ippl/test/6dtrack/Makefile
deleted file mode 100644
index 210d375c1ecc85163dd877a7febe4036b94a0cbe..0000000000000000000000000000000000000000
--- a/ippl/test/6dtrack/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/6dtrack/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/6dtrack/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/6dtrack/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/6dtrack/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/6dtrack/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/6dtrack/Makefile.def b/ippl/test/6dtrack/Makefile.def
deleted file mode 100644
index 7c573aa385d96b2ca696263d0e45166ba57d365f..0000000000000000000000000000000000000000
--- a/ippl/test/6dtrack/Makefile.def
+++ /dev/null
@@ -1,77 +0,0 @@
-#############################################################
-# Automatically generated Makefile.def
-# It is generally not a good idea to modify this by hand.
-#############################################################
-#
-# Defines that characterize the architecture.
-#
-
-USER_INCLUDES = -I~roberto/iobuf
-INCLUDES = $(USER_INCLUDES) -I$(IPPL_ROOT)/src/hdf5/H5ecloud
-USER_DEFINES = -DGTHDF5 -DPARALLEL_IO
-DEFINES = -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DEBUG -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX $(USER_DEFINES)
-USER_CXXINCLUDES =
-CXXINCLUDES = -I$(IPPL_ROOT)/src/expde -I$(IPPL_ROOT)/src -I$(H5HOME)/include $(USER_CXXINCLUDES) $(INCLUDES)
-USER_CINCLUDES =
-CINCLUDES = $(USER_CINCLUDES) $(INCLUDES)
-USER_F77INCLUDES =
-F77INCLUDES = $(USER_F77INCLUDES) $(INCLUDES)
-USER_CXXDEFINES =
-CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX\
- -DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG\
- -DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM\
- -Drestrict=__restrict__ -DNOCTAssert $(USER_CXXDEFINES) $(DEFINES)
-USER_CDEFINES =
-CDEFINES = $(USER_CDEFINES) $(DEFINES)
-USER_F77DEFINES =
-F77DEFINES = $(USER_F77DEFINES) $(DEFINES)
-USER_CXXOPT =
-CXXOPT = -g $(USER_CXXOPT)
-USER_COPT =
-COPT = -g $(USER_COPT)
-USER_F77OPT =
-F77OPT = $(USER_F77OPT)
-
-CXXFLAGS = -fpermissive -Wno-deprecated -ftemplate-depth-80 -funroll-loops -fstrict-aliasing $(CXXOPT) $(CXXINCLUDES) $(CXXDEFINES)
-
-USER_CXXFLAGSLIB =
-USER_CXXFLAGSAPP =
-USER_LDFLAGS =
-CXXFLAGSLIB = $(CXXFLAGS) $(USER_CXXFLAGSLIB)
-CXXFLAGSAPP = $(CXXFLAGS) $(USER_CXXFLAGSAPP)
-CFLAGS = -Drestrict=__restrict__ $(COPT) $(CINCLUDES) $(CDEFINES)
-F77FLAGS = -w $(F77OPT) $(F77INCLUDES) $(F77DEFINES)
-F77FLAGSLIB = $(F77FLAGS) $(USER_F77FLAGSLIB)
-F77FLAGSAPP = $(F77FLAGS) $(USER_F77FLAGSAPP)
-AR = ar rc
-CXX_APP_COMPILE_LINE = $(CXX) $(CXXFLAGSAPP) -c
-CXX_LIB_COMPILE_LINE = $(CXX) $(CXXFLAGSLIB) -c
-CC_LIB_COMPILE_LINE = $(CC) $(CXXFLAGSLIB) -c
-CC_COMPILE_LINE = $(CC) $(CFLAGS) -c
-F77_COMPILE_LINE = $(F77) $(F77FLAGS) -c
-F77AR = ar rc
-LDFLAGS = $(CXXFLAGSAPP) $(USER_LDFLAGS)
-
-############################################################
-# The part below is for linking
-# to include the flags and dependencies for the library.
-############################################################
-
-LINKER = $(CXX) -v
-USER_LFLAGS =
-LFLAGS = -L$(IPPL_ROOT)/lib/$(IPPL_ARCH) $(USER_LFLAGS)
-USER_LIBS = -lsz -lhwpc -lpapi ~roberto/iobuf/iobuf.o
-LIBS = -L$(H5HOME)/lib -lhdf5 -lz -lippl -lm -lfft -lstdc++ $(USER_LIBS)
-
-.cpp.o:
- $(CXX) $(CXXFLAGSAPP) -I. -c $<
-
-.F.o:
- $(F77) $(F77FLAGSAPP) -I. -c $<
-
-.cpp:
- $(CXX) $(CXXFLAGSAPP) -I. -c $<
- $(LINKER) $(LDFLAGS) -o $@ $(<:.cpp=.o) $(LFLAGS) $(LIBS)
-
-.o:
- $(LINKER) $(LDFLAGS) -o $@ $< $(LFLAGS) $(LIBS)
diff --git a/ippl/test/Cartesian/Makefile b/ippl/test/Cartesian/Makefile
deleted file mode 100644
index 702c43409ad50da89482ea72dce5b342e4482b12..0000000000000000000000000000000000000000
--- a/ippl/test/Cartesian/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/Cartesian/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Cartesian/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Cartesian/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Cartesian/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Cartesian/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/CavityTrack/Makefile b/ippl/test/CavityTrack/Makefile
deleted file mode 100644
index 388d834aef222e6db9b2d5ea4578de8e05c0d5f9..0000000000000000000000000000000000000000
--- a/ippl/test/CavityTrack/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/CavityTrack/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/CavityTrack/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/CavityTrack/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/CavityTrack/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/CavityTrack/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/FFT/Makefile b/ippl/test/FFT/Makefile
deleted file mode 100644
index ef2d55267805a185c533cf15796e541df99a15b3..0000000000000000000000000000000000000000
--- a/ippl/test/FFT/Makefile
+++ /dev/null
@@ -1,254 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/FFT/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Convenience name for target.
-test/FFT/CMakeFiles/TestFFT.dir/rule:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/CMakeFiles/TestFFT.dir/rule
-.PHONY : test/FFT/CMakeFiles/TestFFT.dir/rule
-
-# Convenience name for target.
-TestFFT: test/FFT/CMakeFiles/TestFFT.dir/rule
-.PHONY : TestFFT
-
-# fast build rule for target.
-TestFFT/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestFFT.dir/build.make test/FFT/CMakeFiles/TestFFT.dir/build
-.PHONY : TestFFT/fast
-
-# Convenience name for target.
-test/FFT/CMakeFiles/TestRC.dir/rule:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/FFT/CMakeFiles/TestRC.dir/rule
-.PHONY : test/FFT/CMakeFiles/TestRC.dir/rule
-
-# Convenience name for target.
-TestRC: test/FFT/CMakeFiles/TestRC.dir/rule
-.PHONY : TestRC
-
-# fast build rule for target.
-TestRC/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestRC.dir/build.make test/FFT/CMakeFiles/TestRC.dir/build
-.PHONY : TestRC/fast
-
-TestFFT.o: TestFFT.cpp.o
-.PHONY : TestFFT.o
-
-# target to build an object file
-TestFFT.cpp.o:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestFFT.dir/build.make test/FFT/CMakeFiles/TestFFT.dir/TestFFT.cpp.o
-.PHONY : TestFFT.cpp.o
-
-TestFFT.i: TestFFT.cpp.i
-.PHONY : TestFFT.i
-
-# target to preprocess a source file
-TestFFT.cpp.i:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestFFT.dir/build.make test/FFT/CMakeFiles/TestFFT.dir/TestFFT.cpp.i
-.PHONY : TestFFT.cpp.i
-
-TestFFT.s: TestFFT.cpp.s
-.PHONY : TestFFT.s
-
-# target to generate assembly for a file
-TestFFT.cpp.s:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestFFT.dir/build.make test/FFT/CMakeFiles/TestFFT.dir/TestFFT.cpp.s
-.PHONY : TestFFT.cpp.s
-
-TestRC.o: TestRC.cpp.o
-.PHONY : TestRC.o
-
-# target to build an object file
-TestRC.cpp.o:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestRC.dir/build.make test/FFT/CMakeFiles/TestRC.dir/TestRC.cpp.o
-.PHONY : TestRC.cpp.o
-
-TestRC.i: TestRC.cpp.i
-.PHONY : TestRC.i
-
-# target to preprocess a source file
-TestRC.cpp.i:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestRC.dir/build.make test/FFT/CMakeFiles/TestRC.dir/TestRC.cpp.i
-.PHONY : TestRC.cpp.i
-
-TestRC.s: TestRC.cpp.s
-.PHONY : TestRC.s
-
-# target to generate assembly for a file
-TestRC.cpp.s:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/FFT/CMakeFiles/TestRC.dir/build.make test/FFT/CMakeFiles/TestRC.dir/TestRC.cpp.s
-.PHONY : TestRC.cpp.s
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... TestFFT"
- @echo "... TestRC"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
- @echo "... TestFFT.o"
- @echo "... TestFFT.i"
- @echo "... TestFFT.s"
- @echo "... TestRC.o"
- @echo "... TestRC.i"
- @echo "... TestRC.s"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/FFT/SeaborgRes/TestRC.cpp b/ippl/test/FFT/SeaborgRes/TestRC.cpp
index 08619d8f33cf2bdc377629ce86e103dfe3a0a213..4ce2342ebc86053f0f463b90b25bd0725300ad4b 100644
--- a/ippl/test/FFT/SeaborgRes/TestRC.cpp
+++ b/ippl/test/FFT/SeaborgRes/TestRC.cpp
@@ -137,7 +137,6 @@ int main(int argc, char *argv[])
// zeroth axis serial, zeroth axis half-size domain, permuted axis order
FieldLayout<D> layoutSPPerm0h(ndiPermuted0h,serialParallel,vnodes);
#endif
-#ifndef IPPL_USE_SCSL_FFT
// all parallel layout, first axis half-size domain, normal axis order
FieldLayout<D> layoutPPStan1h(ndiStandard1h,allParallel,vnodes);
#ifndef ONED
@@ -145,7 +144,6 @@ int main(int argc, char *argv[])
FieldLayout<D> layoutSPStan1h(ndiStandard1h,serialParallel,vnodes);
// zeroth axis serial, first axis half-size domain, permuted axis order
FieldLayout<D> layoutSPPerm1h(ndiPermuted1h,serialParallel,vnodes);
-#endif
#endif
// create test Fields for complex-to-complex FFT
@@ -170,14 +168,12 @@ int main(int argc, char *argv[])
BareField<dcomplex,D> CFieldSPPerm0h(layoutSPPerm0h);
#endif
-#ifndef IPPL_USE_SCSL_FFT
// create test Fields for sine transform and real-to-complex FFT
BareField<dcomplex,D> CFieldPPStan1h(layoutPPStan1h);
#ifndef ONED
BareField<dcomplex,D> CFieldSPStan1h(layoutSPStan1h);
BareField<dcomplex,D> CFieldSPPerm1h(layoutSPPerm1h);
#endif
-#endif
// For calling FieldDebug functions from debugger, set up output format:
diff --git a/ippl/test/FFT/TestFFT.cpp.org b/ippl/test/FFT/TestFFT.cpp.org
index 024fe971b59f87d71792c3dca522362fa76d9543..15656bec173c58ce467c6a02615defcccf7231fb 100644
--- a/ippl/test/FFT/TestFFT.cpp.org
+++ b/ippl/test/FFT/TestFFT.cpp.org
@@ -162,7 +162,6 @@ int main(int argc, char *argv[])
for (d=1; d<D; d++) ndiPermuted0h[d] = ndiStandard0h[d-1];
#endif
-#ifndef IPPL_USE_SCSL_FFT
// create half-size domain for sine transform along zeroth axis
// and RC transform along first axis
NDIndex<D> ndiStandard1h = ndiStandard;
@@ -172,7 +171,6 @@ int main(int argc, char *argv[])
NDIndex<D> ndiPermuted1h;
ndiPermuted1h[0] = ndiStandard1h[D-1];
for (d=1; d<D; d++) ndiPermuted1h[d] = ndiStandard1h[d-1];
-#endif
#endif
// all parallel layout, standard domain, normal axis order
@@ -191,7 +189,6 @@ int main(int argc, char *argv[])
// zeroth axis serial, zeroth axis half-size domain, permuted axis order
FieldLayout<D> layoutSPPerm0h(ndiPermuted0h,serialParallel,vnodes);
#endif
-#ifndef IPPL_USE_SCSL_FFT
// all parallel layout, first axis half-size domain, normal axis order
FieldLayout<D> layoutPPStan1h(ndiStandard1h,allParallel,vnodes);
#ifndef ONED
@@ -199,7 +196,6 @@ int main(int argc, char *argv[])
FieldLayout<D> layoutSPStan1h(ndiStandard1h,serialParallel,vnodes);
// zeroth axis serial, first axis half-size domain, permuted axis order
FieldLayout<D> layoutSPPerm1h(ndiPermuted1h,serialParallel,vnodes);
-#endif
#endif
// create test Fields for complex-to-complex FFT
@@ -224,13 +220,11 @@ int main(int argc, char *argv[])
BareField<dcomplex,D> CFieldSPPerm0h(layoutSPPerm0h);
#endif
-#ifndef IPPL_USE_SCSL_FFT
// create test Fields for sine transform and real-to-complex FFT
BareField<dcomplex,D> CFieldPPStan1h(layoutPPStan1h);
#ifndef ONED
BareField<dcomplex,D> CFieldSPStan1h(layoutSPStan1h);
BareField<dcomplex,D> CFieldSPPerm1h(layoutSPPerm1h);
-#endif
#endif
// For calling FieldDebug functions from debugger, set up output format:
@@ -581,8 +575,6 @@ int main(int argc, char *argv[])
//-------------------------------------------------------------------------
#endif
-#ifndef IPPL_USE_SCSL_FFT
-
// define zeroth axis to be sine transform
bool sineTransformDims[D];
sineTransformDims[0] = true;
@@ -874,8 +866,6 @@ int main(int argc, char *argv[])
//-------------------------------------------------------------------------
#endif
-#endif // IPPL_USE_SCSL_FFT
-
// Report if any errors have happened:
testmsg << ( (correct) ? "PASSED" : "FAILED" ) << endl;
diff --git a/ippl/test/Hal/Makefile b/ippl/test/Hal/Makefile
deleted file mode 100644
index 935a3a3a7f9e78dcffbce639951ac5dde1d73c03..0000000000000000000000000000000000000000
--- a/ippl/test/Hal/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/Hal/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Hal/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Hal/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Hal/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/Hal/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/Makefile.def b/ippl/test/Makefile.def
index ce88af2f7b0a94d44f81564a4cb6ff8087a48af5..dbe4e8489c42b88c5bd0ee875226a4367b3d63fc 100644
--- a/ippl/test/Makefile.def
+++ b/ippl/test/Makefile.def
@@ -17,7 +17,7 @@ CINCLUDES = $(USER_CINCLUDES) $(INCLUDES)
USER_F77INCLUDES =
F77INCLUDES = $(USER_F77INCLUDES) $(INCLUDES)
USER_CXXDEFINES =
-CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX\
+CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY \
-DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG\
-DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM\
-Drestrict=__restrict__ -DNOCTAssert $(USER_CXXDEFINES) $(DEFINES)
diff --git a/ippl/test/PwrSpec/Makefile b/ippl/test/PwrSpec/Makefile
deleted file mode 100644
index f30e94c1a52a27700fbaf69f96424c39ac097df8..0000000000000000000000000000000000000000
--- a/ippl/test/PwrSpec/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/PwrSpec/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/PwrSpec/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/PwrSpec/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/PwrSpec/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/PwrSpec/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/adaTests/Makefile b/ippl/test/adaTests/Makefile
deleted file mode 100644
index a9264a474de039314b314f2fbadd39ca31abb2e7..0000000000000000000000000000000000000000
--- a/ippl/test/adaTests/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/adaTests/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/adaTests/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/adaTests/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/adaTests/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/adaTests/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/hdf5/Makefile b/ippl/test/hdf5/Makefile
deleted file mode 100644
index 10f1a9377d5721d3b88188d67562dc6f22c81693..0000000000000000000000000000000000000000
--- a/ippl/test/hdf5/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/hdf5/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/hdf5/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/hdf5/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/hdf5/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/hdf5/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/parallel/Makefile b/ippl/test/parallel/Makefile
deleted file mode 100644
index 77e359d1aac0f6add10cf27e66c2ec39339c8acb..0000000000000000000000000000000000000000
--- a/ippl/test/parallel/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/parallel/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/parallel/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/parallel/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/parallel/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/parallel/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/particle/Makefile b/ippl/test/particle/Makefile
deleted file mode 100644
index 1b4667942c7f6fd3ff6e124553ae74c3abf38d0f..0000000000000000000000000000000000000000
--- a/ippl/test/particle/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/particle/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/particle/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/particle/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/particle/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/particle/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/poissonsolvers/Makefile b/ippl/test/poissonsolvers/Makefile
deleted file mode 100644
index 3855a0b007289c651f0d4889d7b2ac0c11271ce7..0000000000000000000000000000000000000000
--- a/ippl/test/poissonsolvers/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/poissonsolvers/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/poissonsolvers/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/poissonsolvers/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/poissonsolvers/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/poissonsolvers/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/simple-new/Makefile b/ippl/test/simple-new/Makefile
deleted file mode 100644
index 6aa0793c2d0e56bc93b2a162b2db2cb61a81cc52..0000000000000000000000000000000000000000
--- a/ippl/test/simple-new/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/simple-new/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple-new/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple-new/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple-new/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple-new/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/simple/Makefile b/ippl/test/simple/Makefile
deleted file mode 100644
index 1b096726064ba66b3aef643a2bf6e5d05a361dcc..0000000000000000000000000000000000000000
--- a/ippl/test/simple/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/simple/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/simple/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/sindex/Makefile b/ippl/test/sindex/Makefile
deleted file mode 100644
index 6e8f00c8e9662e2a7d0e751228be8c1807dc9334..0000000000000000000000000000000000000000
--- a/ippl/test/sindex/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/sindex/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/sindex/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/sindex/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/sindex/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/sindex/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/speed/Makefile b/ippl/test/speed/Makefile
deleted file mode 100644
index 3add0de55b5c14a6c355570878d4f63074a804b4..0000000000000000000000000000000000000000
--- a/ippl/test/speed/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/speed/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/speed/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/speed/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/speed/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/speed/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/speed/make.out b/ippl/test/speed/make.out
deleted file mode 100644
index bb13fa2d257b29c3d0e0c41ab00fbda0bda90b66..0000000000000000000000000000000000000000
--- a/ippl/test/speed/make.out
+++ /dev/null
@@ -1,3 +0,0 @@
-g++3.2 -ftemplate-depth-60 -g -I/home/adelmann/r1/src -I/usr/local/mpich/include -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX -DIPPL_USE_MEMBER_TEMPLATES -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_USE_STANDARD_HEADERS -DIPPL_STDSTL -DIPPL_LONGLONG -Drestrict=__restrict__ -DNOCTAssert -DIPPL_DEBUG -DIPPL_MPI -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX -I. -c speed1.cpp
-g++3.2 -ftemplate-depth-60 -g -I/home/adelmann/r1/src -I/usr/local/mpich/include -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX -DIPPL_USE_MEMBER_TEMPLATES -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_USE_STANDARD_HEADERS -DIPPL_STDSTL -DIPPL_LONGLONG -Drestrict=__restrict__ -DNOCTAssert -DIPPL_DEBUG -DIPPL_MPI -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX -o speed1 speed1.o -L/usr/local/mpich/lib -L/home/adelmann/r1/lib/LINUX -static -lippl -lfft -lmpich -lm
-g++3.2 -ftemplate-depth-60 -g -I/home/adelmann/r1/src -I/usr/local/mpich/include -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX -DIPPL_USE_MEMBER_TEMPLATES -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_USE_STANDARD_HEADERS -DIPPL_STDSTL -DIPPL_LONGLONG -Drestrict=__restrict__ -DNOCTAssert -DIPPL_DEBUG -DIPPL_MPI -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX -I. -c stencil1.cpp
diff --git a/ippl/test/toyfdtf/Makefile b/ippl/test/toyfdtf/Makefile
deleted file mode 100644
index bdee6a85216e92facb77e2402100e7f0a27c9c69..0000000000000000000000000000000000000000
--- a/ippl/test/toyfdtf/Makefile
+++ /dev/null
@@ -1,239 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/toyfdtf/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/toyfdtf/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/toyfdtf/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/toyfdtf/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/toyfdtf/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Convenience name for target.
-test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/rule:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/rule
-.PHONY : test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/rule
-
-# Convenience name for target.
-ipplToyFDTD: test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/rule
-.PHONY : ipplToyFDTD
-
-# fast build rule for target.
-ipplToyFDTD/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build
-.PHONY : ipplToyFDTD/fast
-
-ToyFDTD-DataSink/DataSink.o: ToyFDTD-DataSink/DataSink.cpp.o
-.PHONY : ToyFDTD-DataSink/DataSink.o
-
-# target to build an object file
-ToyFDTD-DataSink/DataSink.cpp.o:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/DataSink.cpp.o
-.PHONY : ToyFDTD-DataSink/DataSink.cpp.o
-
-ToyFDTD-DataSink/DataSink.i: ToyFDTD-DataSink/DataSink.cpp.i
-.PHONY : ToyFDTD-DataSink/DataSink.i
-
-# target to preprocess a source file
-ToyFDTD-DataSink/DataSink.cpp.i:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/DataSink.cpp.i
-.PHONY : ToyFDTD-DataSink/DataSink.cpp.i
-
-ToyFDTD-DataSink/DataSink.s: ToyFDTD-DataSink/DataSink.cpp.s
-.PHONY : ToyFDTD-DataSink/DataSink.s
-
-# target to generate assembly for a file
-ToyFDTD-DataSink/DataSink.cpp.s:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/DataSink.cpp.s
-.PHONY : ToyFDTD-DataSink/DataSink.cpp.s
-
-ToyFDTD-DataSink/ipplToyFDTD2.o: ToyFDTD-DataSink/ipplToyFDTD2.cpp.o
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.o
-
-# target to build an object file
-ToyFDTD-DataSink/ipplToyFDTD2.cpp.o:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/ipplToyFDTD2.cpp.o
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.cpp.o
-
-ToyFDTD-DataSink/ipplToyFDTD2.i: ToyFDTD-DataSink/ipplToyFDTD2.cpp.i
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.i
-
-# target to preprocess a source file
-ToyFDTD-DataSink/ipplToyFDTD2.cpp.i:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/ipplToyFDTD2.cpp.i
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.cpp.i
-
-ToyFDTD-DataSink/ipplToyFDTD2.s: ToyFDTD-DataSink/ipplToyFDTD2.cpp.s
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.s
-
-# target to generate assembly for a file
-ToyFDTD-DataSink/ipplToyFDTD2.cpp.s:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/build.make test/toyfdtf/CMakeFiles/ipplToyFDTD.dir/ToyFDTD-DataSink/ipplToyFDTD2.cpp.s
-.PHONY : ToyFDTD-DataSink/ipplToyFDTD2.cpp.s
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... ipplToyFDTD"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
- @echo "... ToyFDTD-DataSink/DataSink.o"
- @echo "... ToyFDTD-DataSink/DataSink.i"
- @echo "... ToyFDTD-DataSink/DataSink.s"
- @echo "... ToyFDTD-DataSink/ipplToyFDTD2.o"
- @echo "... ToyFDTD-DataSink/ipplToyFDTD2.i"
- @echo "... ToyFDTD-DataSink/ipplToyFDTD2.s"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/ttrack/Makefile b/ippl/test/ttrack/Makefile
deleted file mode 100644
index f71a2c7e2b65a0c0da3f46c2f20b14b01dfc5a4c..0000000000000000000000000000000000000000
--- a/ippl/test/ttrack/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/ttrack/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/ttrack/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/ttrack/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/ttrack/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/ttrack/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/ippl/test/ttrack/Makefile.def b/ippl/test/ttrack/Makefile.def
deleted file mode 100644
index 7c573aa385d96b2ca696263d0e45166ba57d365f..0000000000000000000000000000000000000000
--- a/ippl/test/ttrack/Makefile.def
+++ /dev/null
@@ -1,77 +0,0 @@
-#############################################################
-# Automatically generated Makefile.def
-# It is generally not a good idea to modify this by hand.
-#############################################################
-#
-# Defines that characterize the architecture.
-#
-
-USER_INCLUDES = -I~roberto/iobuf
-INCLUDES = $(USER_INCLUDES) -I$(IPPL_ROOT)/src/hdf5/H5ecloud
-USER_DEFINES = -DGTHDF5 -DPARALLEL_IO
-DEFINES = -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DEBUG -DIPPL_GCC -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DIPPL_LINUX $(USER_DEFINES)
-USER_CXXINCLUDES =
-CXXINCLUDES = -I$(IPPL_ROOT)/src/expde -I$(IPPL_ROOT)/src -I$(H5HOME)/include $(USER_CXXINCLUDES) $(INCLUDES)
-USER_CINCLUDES =
-CINCLUDES = $(USER_CINCLUDES) $(INCLUDES)
-USER_F77INCLUDES =
-F77INCLUDES = $(USER_F77INCLUDES) $(INCLUDES)
-USER_CXXDEFINES =
-CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX\
- -DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG\
- -DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM\
- -Drestrict=__restrict__ -DNOCTAssert $(USER_CXXDEFINES) $(DEFINES)
-USER_CDEFINES =
-CDEFINES = $(USER_CDEFINES) $(DEFINES)
-USER_F77DEFINES =
-F77DEFINES = $(USER_F77DEFINES) $(DEFINES)
-USER_CXXOPT =
-CXXOPT = -g $(USER_CXXOPT)
-USER_COPT =
-COPT = -g $(USER_COPT)
-USER_F77OPT =
-F77OPT = $(USER_F77OPT)
-
-CXXFLAGS = -fpermissive -Wno-deprecated -ftemplate-depth-80 -funroll-loops -fstrict-aliasing $(CXXOPT) $(CXXINCLUDES) $(CXXDEFINES)
-
-USER_CXXFLAGSLIB =
-USER_CXXFLAGSAPP =
-USER_LDFLAGS =
-CXXFLAGSLIB = $(CXXFLAGS) $(USER_CXXFLAGSLIB)
-CXXFLAGSAPP = $(CXXFLAGS) $(USER_CXXFLAGSAPP)
-CFLAGS = -Drestrict=__restrict__ $(COPT) $(CINCLUDES) $(CDEFINES)
-F77FLAGS = -w $(F77OPT) $(F77INCLUDES) $(F77DEFINES)
-F77FLAGSLIB = $(F77FLAGS) $(USER_F77FLAGSLIB)
-F77FLAGSAPP = $(F77FLAGS) $(USER_F77FLAGSAPP)
-AR = ar rc
-CXX_APP_COMPILE_LINE = $(CXX) $(CXXFLAGSAPP) -c
-CXX_LIB_COMPILE_LINE = $(CXX) $(CXXFLAGSLIB) -c
-CC_LIB_COMPILE_LINE = $(CC) $(CXXFLAGSLIB) -c
-CC_COMPILE_LINE = $(CC) $(CFLAGS) -c
-F77_COMPILE_LINE = $(F77) $(F77FLAGS) -c
-F77AR = ar rc
-LDFLAGS = $(CXXFLAGSAPP) $(USER_LDFLAGS)
-
-############################################################
-# The part below is for linking
-# to include the flags and dependencies for the library.
-############################################################
-
-LINKER = $(CXX) -v
-USER_LFLAGS =
-LFLAGS = -L$(IPPL_ROOT)/lib/$(IPPL_ARCH) $(USER_LFLAGS)
-USER_LIBS = -lsz -lhwpc -lpapi ~roberto/iobuf/iobuf.o
-LIBS = -L$(H5HOME)/lib -lhdf5 -lz -lippl -lm -lfft -lstdc++ $(USER_LIBS)
-
-.cpp.o:
- $(CXX) $(CXXFLAGSAPP) -I. -c $<
-
-.F.o:
- $(F77) $(F77FLAGSAPP) -I. -c $<
-
-.cpp:
- $(CXX) $(CXXFLAGSAPP) -I. -c $<
- $(LINKER) $(LDFLAGS) -o $@ $(<:.cpp=.o) $(LFLAGS) $(LIBS)
-
-.o:
- $(LINKER) $(LDFLAGS) -o $@ $< $(LFLAGS) $(LIBS)
diff --git a/ippl/test/twostream-1/Makefile b/ippl/test/twostream-1/Makefile
deleted file mode 100644
index 7d6df1b138335dd0fed26ee19c0528ed1de09d50..0000000000000000000000000000000000000000
--- a/ippl/test/twostream-1/Makefile
+++ /dev/null
@@ -1,170 +0,0 @@
-# CMAKE generated file: DO NOT EDIT!
-# Generated by "Unix Makefiles" Generator, CMake Version 2.8
-
-# Default target executed when no arguments are given to make.
-default_target: all
-.PHONY : default_target
-
-#=============================================================================
-# Special targets provided by cmake.
-
-# Disable implicit rules so canonical targets will work.
-.SUFFIXES:
-
-# Remove some rules from gmake that .SUFFIXES does not remove.
-SUFFIXES =
-
-.SUFFIXES: .hpux_make_needs_suffix_list
-
-# Suppress display of executed commands.
-$(VERBOSE).SILENT:
-
-# A target that is always out of date.
-cmake_force:
-.PHONY : cmake_force
-
-#=============================================================================
-# Set environment variables for the build.
-
-# The shell in which to execute make rules.
-SHELL = /bin/sh
-
-# The CMake executable.
-CMAKE_COMMAND = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake
-
-# The command to remove a file.
-RM = /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -E remove -f
-
-# Escaping for special characters.
-EQUALS = =
-
-# The program to use to edit the cache.
-CMAKE_EDIT_COMMAND = /usr/bin/ccmake
-
-# The top-level source directory on which CMake was run.
-CMAKE_SOURCE_DIR = /home/l_locans/programs/IPPL/trunk
-
-# The top-level build directory on which CMake was run.
-CMAKE_BINARY_DIR = /home/l_locans/programs/IPPL/trunk
-
-#=============================================================================
-# Targets provided globally by CMake.
-
-# Special rule for the target edit_cache
-edit_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake cache editor..."
- /usr/bin/ccmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : edit_cache
-
-# Special rule for the target edit_cache
-edit_cache/fast: edit_cache
-.PHONY : edit_cache/fast
-
-# Special rule for the target install
-install: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install
-
-# Special rule for the target install
-install/fast: preinstall/fast
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Install the project..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -P cmake_install.cmake
-.PHONY : install/fast
-
-# Special rule for the target install/local
-install/local: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing only the local directory..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_LOCAL_ONLY=1 -P cmake_install.cmake
-.PHONY : install/local
-
-# Special rule for the target install/local
-install/local/fast: install/local
-.PHONY : install/local/fast
-
-# Special rule for the target install/strip
-install/strip: preinstall
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Installing the project stripped..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -DCMAKE_INSTALL_DO_STRIP=1 -P cmake_install.cmake
-.PHONY : install/strip
-
-# Special rule for the target install/strip
-install/strip/fast: install/strip
-.PHONY : install/strip/fast
-
-# Special rule for the target list_install_components
-list_install_components:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Available install components are: \"Unspecified\""
-.PHONY : list_install_components
-
-# Special rule for the target list_install_components
-list_install_components/fast: list_install_components
-.PHONY : list_install_components/fast
-
-# Special rule for the target rebuild_cache
-rebuild_cache:
- @$(CMAKE_COMMAND) -E cmake_echo_color --switch=$(COLOR) --cyan "Running CMake to regenerate build system..."
- /afs/psi.ch/sys/psi.x86_64_slp6/Programming/cmake/2.8.12.2/bin/cmake -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR)
-.PHONY : rebuild_cache
-
-# Special rule for the target rebuild_cache
-rebuild_cache/fast: rebuild_cache
-.PHONY : rebuild_cache/fast
-
-# The main all target
-all: cmake_check_build_system
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles /home/l_locans/programs/IPPL/trunk/test/twostream-1/CMakeFiles/progress.marks
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/twostream-1/all
- $(CMAKE_COMMAND) -E cmake_progress_start /home/l_locans/programs/IPPL/trunk/CMakeFiles 0
-.PHONY : all
-
-# The main clean target
-clean:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/twostream-1/clean
-.PHONY : clean
-
-# The main clean target
-clean/fast: clean
-.PHONY : clean/fast
-
-# Prepare targets for installation.
-preinstall: all
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/twostream-1/preinstall
-.PHONY : preinstall
-
-# Prepare targets for installation.
-preinstall/fast:
- cd /home/l_locans/programs/IPPL/trunk && $(MAKE) -f CMakeFiles/Makefile2 test/twostream-1/preinstall
-.PHONY : preinstall/fast
-
-# clear depends
-depend:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 1
-.PHONY : depend
-
-# Help Target
-help:
- @echo "The following are some of the valid targets for this Makefile:"
- @echo "... all (the default if no target is provided)"
- @echo "... clean"
- @echo "... depend"
- @echo "... edit_cache"
- @echo "... install"
- @echo "... install/local"
- @echo "... install/strip"
- @echo "... list_install_components"
- @echo "... rebuild_cache"
-.PHONY : help
-
-
-
-#=============================================================================
-# Special targets to cleanup operation of make.
-
-# Special rule to run CMake to check the build system integrity.
-# No rule that depends on this can have commands that come from listfiles
-# because they might be regenerated.
-cmake_check_build_system:
- cd /home/l_locans/programs/IPPL/trunk && $(CMAKE_COMMAND) -H$(CMAKE_SOURCE_DIR) -B$(CMAKE_BINARY_DIR) --check-build-system CMakeFiles/Makefile.cmake 0
-.PHONY : cmake_check_build_system
-
diff --git a/src/Classic/autogen-gele.sh b/src/Classic/autogen-gele.sh
deleted file mode 100644
index 9465bc6cf2f6dbc4cb004ab6eb4b1f788b91547d..0000000000000000000000000000000000000000
--- a/src/Classic/autogen-gele.sh
+++ /dev/null
@@ -1,13 +0,0 @@
-#!/bin/bash
-#
-mkdir -p config
-aclocal
-autoconf
-automake -a -c
-make clean
-CXX=CC ./configure --host=x86_64-unknown-linux-gnu \
- --with-fftw3-includedir=/apps/fftw/fftw-3.1.2_gnu3.3_PE1.4.48/include \
- --with-fftw3-libdir=/apps/fftw/fftw-3.1.2_gnu3.3_PE1.4.48/lib \
- --with-ippl-includedir=$IPPL_ROOT/src
-make -j 10
-
diff --git a/src/Classic/autogen-palu.sh b/src/Classic/autogen-palu.sh
deleted file mode 100644
index 7525e4c6424dae9a9eda4bbd7a910df45154f190..0000000000000000000000000000000000000000
--- a/src/Classic/autogen-palu.sh
+++ /dev/null
@@ -1,16 +0,0 @@
-#!/bin/bash
-#
-mkdir -p config
-aclocal
-autoheader
-autoconf
-automake -a -c
-autoreconf
-make clean
-CXX=CC ./configure --host=x86_64-unknown-linux-gnu CPP="CC" \
- --with-fftw3-includedir=/apps/fftw/fftw-3.1.2_gnu3.3_PE1.5.47/include \
- --with-fftw3-libdir=/apps/fftw/fftw-3.1.2_gnu3.3_PE1.5.47/lib \
- --with-ippl-includedir=$IPPL_ROOT/src \
- --with-hdf5-includedir=$H5HOME/include --with-hdf5-libdir=$H5HOME/lib \
- --with-h5part-includedir=$H5PartHOME/src --with-h5part-libdir=$H5PartHOME/src
-make -j 10
diff --git a/src/Classic/autogen-regression.sh b/src/Classic/autogen-regression.sh
deleted file mode 100644
index 6750c230b8efa11c3af269e3858bcc61619712c7..0000000000000000000000000000000000000000
--- a/src/Classic/autogen-regression.sh
+++ /dev/null
@@ -1,11 +0,0 @@
-#!/bin/bash
-#
-mkdir -p config
-aclocal --force
-libtoolize --force --copy
-automake --force --add-missing --copy
-autoheader --force
-autoconf --force
-autoreconf
-
-./configure --with-ippl-includedir=/home2/ineichen/svnwork/ippl//src --with-h5part-includedir=/home2/ineichen/felsim/H5Part/src --with-h5part-libdir=/home2/ineichen/felsim/H5Part/src --with-hdf5-includedir=/opt/hdf5/hdf5-1.6.10-openmpi-1.2.6-intel-11.1/include --with-hdf5-libdir=/opt/hdf5/hdf5-1.6.10-openmpi-1.2.6-intel-11.1/lib --with-gsl-includedir=/opt/gsl/gsl-1.12/include --with-gsl-libdir=/opt/gsl/gsl-1.12/lib CC=mpicc CXX=mpicxx F77=mpif77
diff --git a/src/Classic/autogen.sh b/src/Classic/autogen.sh
deleted file mode 100644
index 1459f58104f63e855294f9e645d3e0a47d694e21..0000000000000000000000000000000000000000
--- a/src/Classic/autogen.sh
+++ /dev/null
@@ -1,15 +0,0 @@
-#!/bin/bash
-#
-mkdir -p config
-aclocal --force
-libtoolize --force --copy
-automake --force --add-missing --copy
-autoheader --force
-autoconf --force
-autoreconf
-CXX=mpicxx ./configure \
- --with-ippl-includedir=$IPPL_ROOT/src \
- --with-h5part-includedir=$H5Part/src --with-h5part-libdir=$H5Part/src \
- --with-gsl-includedir=$GSL_PREFIX/include --with-gsl-libdir=$GSL_PREFIX/lib \
- --with-hdf5-includedir=$HDF5_INCLUDE_PATH --with-hdf5-libdir=$HDF5_LIBRARY_PATH
-make -j 10
diff --git a/src/Classic/configure.ac b/src/Classic/configure.ac
deleted file mode 100644
index 5457de7a4e75f86ff11f4de3de451b769f84346f..0000000000000000000000000000000000000000
--- a/src/Classic/configure.ac
+++ /dev/null
@@ -1,129 +0,0 @@
-AC_INIT([classic],[5.1.2],[opal@lists.psi.ch])
-
-AC_PROG_CC([mpicc])
-AC_PROG_CXX([mpicxx])
-
-AC_CONFIG_AUX_DIR(config)
-#disable f77 tests
-m4_defun([_LT_AC_LANG_F77_CONFIG], [:])
-AM_INIT_AUTOMAKE
-AC_LIBTOOL_DLOPEN
-AC_PROG_LIBTOOL
-
-AC_CONFIG_HEADER([./config.h:./config.in])
-AC_CONFIG_FILES([Makefile])
-
-AC_LANG(C++)
-
-IPPLDEFS="-DIPPL_MPI\
- -DMPICH_SKIP_MPICXX\
- -DIPPL_DEBUG\
- -DIPPL_DONT_POOL\
- -DIPPL_USE_XDIV_RNG\
- -DIPPL_LINUX\
- -DIPPL_NO_STRINGSTREAM\
- -DPETE_BITWISE_COPY\
- -DIPPL_HAS_TEMPLATED_COMPLEX\
- -DIPPL_USE_STANDARD_HEADERS\
- -DIPPL_USE_PARTIAL_SPECIALIZATION\
- -DIPPL_STDSTL\
- -DIPPL_LONGLONG\
- -Drestrict=__restrict__ -DNOCTAssert -DPARALLEL_IO -w"
-
-OPALSTUFF="$OPAL_ROOT/src"
-
-AC_ARG_WITH(ippl-includedir,
- AC_HELP_STRING(
- [--with-ippl-includedir=dir],
- [ippl include files in dir]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval ${IPPLDEFS} -I$OPALSTUFF"
- ]
-)
-AC_ARG_WITH(gsl,
- AC_HELP_STRING(
- [--with-gsl=dir],
- [gsl prefix]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval/include"
- test X$withvalue != Xno && LDFLAGS="$LDFLAGS -L$withval/lib"
- ]
-)
-AC_ARG_WITH(gsl-includedir,
- AC_HELP_STRING(
- [--with-gsl-includedir=dir],
- [gsl include files in dir]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval"
- ]
-)
-AC_ARG_WITH(gsl-libdir,
- AC_HELP_STRING(
- [--with-gsl-libdir=dir],
- [gsl library libgsl.a in dir]
- ),
- [
- test X$withvalue != Xno && LDFLAGS="$LDFLAGS -L$withval"
- HAVE_GSL_=true
- ]
-)
-AC_ARG_WITH(h5hut,
- AC_HELP_STRING(
- [--with-h5hut=dir],
- [H5hut prefix]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval/include"
- test X$withvalue != Xno && LDFLAGS="$LDFLAGS -L$withval/lib"
- ]
-)
-AC_ARG_WITH(h5hut-includedir,
- AC_HELP_STRING(
- [--with-h5hut-includedir=dir],
- [H5hut include files in dir]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval"
- ]
-)
-AC_ARG_WITH(h5hut-libdir,
- AC_HELP_STRING(
- [--with-h5hut-libdir=dir],
- [h5hut libraries in dir]
- ),
- [
- test X$withvalue != Xno && LDFLAGS="$LDFLAGS -L$withval"
- ]
-)
-AC_ARG_WITH(hdf5,
- AC_HELP_STRING(
- [--with-hdf5-includedir=dir],
- [hdf5 prefix]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval/include"
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval/lib"
- ]
-)
-AC_ARG_WITH(hdf5-includedir,
- AC_HELP_STRING(
- [--with-hdf5-includedir=dir],
- [hdf5 include files in dir]
- ),
- [
- test X$withvalue != Xno && CPPFLAGS="$CPPFLAGS -I$withval"
- ]
-)
-AC_ARG_WITH(hdf5-libdir,
- AC_HELP_STRING(
- [--with-hdf5-libdir=dir],
- [hdf5 library in dir]
- ),
- [
- test X$withvalue != Xno && LDFLAGS="$LDFLAGS -L$withval"
- ]
-)
-AC_OUTPUT
diff --git a/src/Copyright.readme b/src/Copyright.readme
index 4261a0e74144c8150fe01a18408866e63b7ded13..99cce54171958bd744f5450d03945abe54151094 100644
--- a/src/Copyright.readme
+++ b/src/Copyright.readme
@@ -1,38 +1,29 @@
-// -------------------------------------------------------------------------
-//
-// PSI Paul Scherrer Institut
-//
-// Program name: OPAL (Object Oriented Parallel Accelerator Library)
-//
-// Version: 1.0
-//
-// Date: xx.2.2008
-//
-// Authors and contact: Andreas Adelmann
-// andreas.adelmann@psi.ch
-//
-// Address: AMAS (Accelerator Modelling and Advanced Simulations)
-// PSI
-// CH-5232 Villigen PSI
-// SWITZERLAND
-//
-// Copyright PSI, Villigen 2007 - Copyright and any other appropriate legal
-// protection of this computer program and associated documentation reserved
-// in all countries of the world.
-//
-// Organizations collaborating with PSI may receive this program and
-// documentation freely and without charge.
-//
-// PSI undertakes no obligation for the maintenance of this program, nor
-// responsibility for its correctness, and accepts no liability whatsoever
-// resulting from its use.
-//
-// Program and documentation are provided solely for the use of the
-// organisation to which they are distributed.
-//
-// This program may not be copied or otherwise distributed without permission.
-// This message must be retained on this and any other authorized copies.
-//
-// The material cannot be sold. PSI should be given credit in all references.
-//
-// -------------------------------------------------------------------------
+Copyright (c) 2008-2018
+Paul Scherrer Institut, Villigen PSI, Switzerland
+All rights reserved.
+
+OPAL is licensed under the terms of the GNU General Public License
+version 3.
+
+(1) Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+
+(2) Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in the
+documentation and/or other materials provided with the distribution.
+
+(3) Neither the name of Paul Scherrer Institut nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/tools/BandRF/Makefile.def b/tools/BandRF/Makefile.def
index 978cb798d0f8ce4d845adbde1efcf42971a39f77..3193ef2c5afeead06c2f7dd11ca617c4caddb0e6 100644
--- a/tools/BandRF/Makefile.def
+++ b/tools/BandRF/Makefile.def
@@ -17,7 +17,7 @@ CINCLUDES = $(USER_CINCLUDES) $(INCLUDES)
USER_F77INCLUDES =
F77INCLUDES = $(USER_F77INCLUDES) $(INCLUDES)
USER_CXXDEFINES =
-CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX\
+CXXDEFINES = -DIPPL_NO_STRINGSTREAM -DPETE_BITWISE_COPY \
-DIPPL_USE_STANDARD_HEADERS -DIPPL_USE_PARTIAL_SPECIALIZATION -DIPPL_STDSTL -DIPPL_LONGLONG\
-DWITH_BRICK -DnoCOMP_GNUOLD -DIPPL_STRINGSTREAM\
-Drestrict=__restrict__ -DNOCTAssert $(USER_CXXDEFINES) $(DEFINES)
diff --git a/tools/mslang/CMakeLists.txt b/tools/mslang/CMakeLists.txt
index c0dfd7a4c37d3bb05b467d6c3db01f635f1f1a5d..1f0078167c592a3ca35f7a7252b2f4c66ad15b15 100644
--- a/tools/mslang/CMakeLists.txt
+++ b/tools/mslang/CMakeLists.txt
@@ -6,7 +6,7 @@ PROJECT (MSLANG VERSION 0.1)
#SET (MSLANG_VERSION_MINOR 1)
SET (IPPL_CXX_FLAGS
- "-DIPPL_LINUX -DIPPL_STRINGSTREAM -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DPETE_BITWISE_COPY -DIPPL_HAS_TEMPLATED_COMPLEX -DIPPL_USE_PARTIAL_SPECIALIZATION -Drestrict=__restrict__ -DNOCTAssert ${IPPL_CXX_FLAGS}"
+ "-DIPPL_LINUX -DIPPL_STRINGSTREAM -DIPPL_MPI -DMPICH_SKIP_MPICXX -DIPPL_DONT_POOL -DIPPL_USE_XDIV_RNG -DPETE_BITWISE_COPY -DIPPL_USE_PARTIAL_SPECIALIZATION -Drestrict=__restrict__ -DNOCTAssert ${IPPL_CXX_FLAGS}"
)
SET (CMAKE_CXX_FLAGS
@@ -43,4 +43,4 @@ MESSAGE( STATUS "Compiling MSLang")
ADD_EXECUTABLE( mslang mslang.cpp )
TARGET_LINK_LIBRARIES( mslang ${MSLANG_LIBS} )
-INSTALL(TARGETS mslang RUNTIME DESTINATION "${CMAKE_INSTALL_PREFIX}/bin")
\ No newline at end of file
+INSTALL(TARGETS mslang RUNTIME DESTINATION "${CMAKE_INSTALL_PREFIX}/bin")