src issueshttps://gitlab.psi.ch/OPAL/src/-/issues2018-01-05T08:58:48+01:00https://gitlab.psi.ch/OPAL/src/-/issues/14Particles stored in trackOrbit.dat2018-01-05T08:58:48+01:00Valeria RizzoglioParticles stored in trackOrbit.datAccording to the OPAL manual (pag. 40):
**Multi-particle tracking mode**
The intermediate phase space data of centeral particle (with ID of 0) and an off-centering particle (with ID of 1) are stored in an ASCII file.
Concernin...According to the OPAL manual (pag. 40):
**Multi-particle tracking mode**
The intermediate phase space data of centeral particle (with ID of 0) and an off-centering particle (with ID of 1) are stored in an ASCII file.
Concerning the particle with *ID0*:
* particle position is not updated in case OFFSETY > 0 is set in the distribution definition. The tracking of this particle does not reflect the beam behavior (because of the offset)
* Is the general idea: ID0 particle = reference particle?
Concerning the particle with *ID1*:
* ***Distribution from file:*** the second particle in the distribution file is used as ID1
* ***Generated distribution:*** It seems that a random particle from the distribution is set as ID1
A possible suggestion:
* ***Distribution from file:*** the first and the second particle in the file are used as ID0 and ID1, respectively. The user is completely free to decide which particles track.
* ***Generated distribution (Option 1):*** ID0 is by default assigned to the reference particle (with updated offset). An option can be added to the DISTRIBUTION command where the user can specify which particle uses as ID1 (ie: DISPERSION, CENTROID or USERDEF or NULL = not stored). This means that the first two particles generated by OPAL are replaced with ID0 (reference) and ID1, if option NULL is not specified.
* ***Generated distribution (Option 2):*** An option can be added to the DISTRIBUTION command where the user can specify which particle uses as ID0 and ID1 (ie: DISPERSION, CENTROID or USERDEF or or NULL = not stored)
A possible problem could arise in case of multi-distribution or vector of distributionOPAL 1.5.1adelmannadelmannhttps://gitlab.psi.ch/OPAL/src/-/issues/77use-dks option2017-03-28T13:10:57+02:00Valeria Rizzogliouse-dks optionOn opalrunner, I load the opal-toolschain/1.6 and OPAL/1.5.1-20170217
```
> Currently Loaded Modulefiles:
> 1) gcc/5.4.0 4) hdf5/1.8.18 7) trilinos/12.10.1 10) OpenBLAS/0.2.19 13) opal-toolschain/1.6
>...On opalrunner, I load the opal-toolschain/1.6 and OPAL/1.5.1-20170217
```
> Currently Loaded Modulefiles:
> 1) gcc/5.4.0 4) hdf5/1.8.18 7) trilinos/12.10.1 10) OpenBLAS/0.2.19 13) opal-toolschain/1.6
> 2) openmpi/1.10.4 5) H5hut/2.0.0rc3 8) root/6.08.02 11) cuda/8.0.44 14) OPAL/1.5.1-20170217
> 3) boost/1.62.0 6) gsl/2.2.1 9) cmake/3.6.3 12) dks/1.0.1
```
then run opal with the --use-dks option: `opal --use-dks Input.in`
It seems that the GPU card is not detected (not sure if the IPPL message is correct).
```
Ippl> No GPU device detected! --use-dks flag will have no effect
> ____ _____ ___
> / __ \| __ \ /\ | |
> | | | | |__) / \ | |
> | | | | ___/ /\ \ | |
> | |__| | | / ____ \| |____
> \____/|_| /_/ \_\______|
OPAL>
OPAL> This is OPAL (Object Oriented Parallel Accelerator Library) Version 1.5.1
OPAL>
OPAL> (c) PSI, http://amas.web.psi.ch
OPAL>
OPAL> OPAL compiled with DKS (Dynamic Kernel Scheduler) Version 1.0.0
```OPAL 1.5.1https://gitlab.psi.ch/OPAL/src/-/issues/42APERTURE attribute does not work2019-11-06T10:22:10+01:00Valeria RizzoglioAPERTURE attribute does not workI test the following element without surface physics (as reported on the OPAL manual: **18.5 Example of an Input File**):
`FX5: SLIT, L=0.09, ELEMEDGE=0.01,APERTURE={0.005,0.003};`
This command is not recognized by OPAL and the ...I test the following element without surface physics (as reported on the OPAL manual: **18.5 Example of an Input File**):
`FX5: SLIT, L=0.09, ELEMEDGE=0.01,APERTURE={0.005,0.003};`
This command is not recognized by OPAL and the slit has no aperture:
```
OPAL[3]> * Slit x= 0 Slit y= 0 start= 0.01 fn=
Ippl[3]> FX5 gone live
```
Btw this information is displayed only with `--info 3`.
I could reproduce the same effect with different OPAL versions:
* **OPAL1.4.0 released version:** the beam is lost without any error or warning messages
```
OPAL >
OPAL > * ************** B U N C H *********************************************************
OPAL > * NP = 10000
OPAL > * Qtot = 1.60000e-06 [nC] Qi = 1.60000e-10 [nC]
....
OPAL > * spos = 9.14511e-03 [m]
OPAL > * **********************************************************************************
OPAL >
OPAL >
OPAL > * ************** B U N C H *********************************************************
OPAL > * NP = 1
OPAL > * Qtot = 1.60000e-10 [nC] Qi = 1.60000e-10 [nC]
....
OPAL > * spos = 1.00250e-01 [m]
OPAL > * **********************************************************************************
OPAL >
Ippl> Save FX5.h5
OPAL> done executing ParallelTTracker at 10:43:43
```
* **OPAL/1.5.1-20170217:** the beam is lost but at least there is an error message
```
ParallelTTracker > * Deleted 10000 particles, remaining 0 particles
Warning> reduce: there was nothing to reduce, since the masks were all false.
Warning> reduce: there was nothing to reduce, since the masks were all false.
Warning> reduce: there was nothing to reduce, since the masks were all false.
Warning> reduce: there was nothing to reduce, since the masks were all false.
Warning> reduce: there was nothing to reduce, since the masks were all false.
```
* **Solution and Request**
1- **Don't fix it:** in this case the APERTURE has to be removed as element attribute and delete every reference from the manual
2- **Fix it:** would be the nicer solution since it should allow to create asymmetric apertures in collimators and slits, otherwise not possibleOPAL 1.5.1adelmannadelmannhttps://gitlab.psi.ch/OPAL/src/-/issues/39New OPAL module (OPAL/1.5.1-20170217) and GPU2020-04-02T09:16:08+02:00Valeria RizzoglioNew OPAL module (OPAL/1.5.1-20170217) and GPUOn opalrunner:
* **Module list**
```
Currently Loaded Modulefiles:
1) gcc/5.4.0 3) hdf5/1.8.18 5) OPAL/1.5.1-20170217 7) dks/1.0.1
2) openmpi/1.10.4 4) H5hut/2.0.0rc3 6) cuda/8.0.44
```
...On opalrunner:
* **Module list**
```
Currently Loaded Modulefiles:
1) gcc/5.4.0 3) hdf5/1.8.18 5) OPAL/1.5.1-20170217 7) dks/1.0.1
2) openmpi/1.10.4 4) H5hut/2.0.0rc3 6) cuda/8.0.44
```
* **Run with GPU**
`mpirun -np 8 opal PROSCAN.in --use-dks`
* **Cout**
` Ippl> OPAL compiled without DKS, --use-dks flag has no effect`
Did I miss to load some modules? Did I do somethig wrong?OPAL 1.5.1gsellgsellhttps://gitlab.psi.ch/OPAL/src/-/issues/21Build system puts Boost include directory before OPAL include directories2017-03-16T14:34:54+01:00snuverink_jjochem.snuverink@psi.chBuild system puts Boost include directory before OPAL include directoriesThe build system puts the Boost include directory before the OPAL include directories. When installing OPAL in the same directory as where boost resides (I know perhaps not the best idea), a new build will give preference to the header f...The build system puts the Boost include directory before the OPAL include directories. When installing OPAL in the same directory as where boost resides (I know perhaps not the best idea), a new build will give preference to the header files in the install directory. This gives compilation errors when the files are updated.
I was able to fix this by putting the OPAL include directories explicitly before others (see [attached patch](/uploads/5ec73615603049c56cadb8dd6ec28daa/include.patch)).OPAL 1.5.1gsellgsellhttps://gitlab.psi.ch/OPAL/src/-/issues/15Distribution FROMFILE: different behaviour OPAL-T and OPAL-Cyc2019-10-25T13:47:36+02:00Valeria RizzoglioDistribution FROMFILE: different behaviour OPAL-T and OPAL-CycWhen the distribution "FROMFILE" is used, the number of particles (N) in the distribution file and in the OPAL input file (NPart) should match.
It could happen that two different numbers are used. In this case OPAL-T and OPAL-Cyc have ...When the distribution "FROMFILE" is used, the number of particles (N) in the distribution file and in the OPAL input file (NPart) should match.
It could happen that two different numbers are used. In this case OPAL-T and OPAL-Cyc have different behaviour:
* **OPAL-T**: Warning message, but the simulation still runs
```
OPAL> --------------------------------------------------
OPAL> Warning!! The number of particles in the initial
OPAL> distribution is 100000.
OPAL>
OPAL> This is different from the number of particles
OPAL> defined by the BEAM command: 10
OPAL>
OPAL> This is often happens when using a FROMFILE type
OPAL> distribution and not matching the number of
OPAL> particles in the particles file(s) with the number
OPAL> given in the BEAM command.
OPAL>
OPAL> The number of particles in the initial distribution
OPAL> (100000) will take precedence.
OPAL> ---------------------------------------------------
```
* **OPAL-Cyc**: Error message and the simulation stops
```
OPAL> *** User error detected by function "TrackRun::execute CYCLOTRON_T"
OPAL> *** in line 46 of file "120_deg_CFM.in" at end of statement:
OPAL> RUN,METHOD="CYCLOTRON-T",BEAM=BEAM_4,FIELDSOLVER=FS1,DISTRIBUTION=DIST1;
OPAL> Number of macro particles and NPART on BEAM are not equal
```OPAL 1.5.1adelmannadelmann2017-01-17https://gitlab.psi.ch/OPAL/src/-/issues/11Update toolchain to current versions2019-10-25T13:46:05+02:00gsellUpdate toolchain to current versionsUpdate tool-chain to build OPAL:
| Group | Module | Linux | Darwin |
|-------------|-------------------------|:-----:|:------:|
| Tools | openssl/1.0.2h | x | x |
| Programming | cmake...Update tool-chain to build OPAL:
| Group | Module | Linux | Darwin |
|-------------|-------------------------|:-----:|:------:|
| Tools | openssl/1.0.2h | x | x |
| Programming | cmake/3.6.3 | x | x |
| Programming | cuda/8.0.44 | x | |
| Programming | Tcl/8.6.4 | x | x |
| Programming | Tk/8.6.4 | x | x |
| Programming | Python/2.7.12 | x | x |
| Programming | gcc/5.4.0 | x | x |
| Compiler | boost/1.62 | x | x |
| Compiler | gsl/2.2.1 | x | x |
| Compiler | hdf5_serial/1.8.18 | x | x |
| Compiler | OpenBLAS/0.2.19 | x | x |
| Compiler | openmpi/1.10.4 | x | x |
| Compiler | root/6.08.02 | x | x |
| HDF5_serial | H5hut/2.0.0rc3 | x | x |
| MPI | dks/1.0.1 | x | |
| MPI | hdf5/1.8.18 | x | x |
| MPI | parmetis/4.0.3 | x | x |
| MPI | trilinos/12.10.1 | x | x |
| HDF5 | H5hut/2.0.0rc3 | x | x |
For the time being DKS is not supported on Mac OS X
OPAL 1.5.1gsellgsellhttps://gitlab.psi.ch/OPAL/src/-/issues/7New T7 Field normalisation2017-03-16T14:35:05+01:00adelmannNew T7 Field normalisationT7 Files are normalised differently since 1.5.0. Provide tool to determine conversion factor.T7 Files are normalised differently since 1.5.0. Provide tool to determine conversion factor.OPAL 1.5.1krauskraus