Opal maps
This MR will implement Map tracking for:
- Quadrupole
- Dipole
- Drifts
- Bends
TODO list:
-
Code implementation (@ganz_p): ThickTracker, Hamiltonian, MapAnalyser -
Doxygen documentation (@ganz_p) -
Documentation (@ganz_p) -
Code cleaning (@frey_m) -
Regression tests (@snuverink_j) -
Examples (@ganz_p) -
Jupyter notebook (@ganz_p) -
Use beam line order
Remaining issues:
- Output dumps in local coordinates only
- Dipole until 1st order
- Tune calculation not stable
- Space charge calculations needs to be verified
- Fringe fields for Bends
Merge request reports
Activity
assigned to @snuverink_j
assigned to @ganz_p
added 1 commit
- 40af4740 - OPAL-maps: remove std::ofstream + COSY constants
src/Algorithms/MapAnalyser.cpp 0 → 1006444 5 6 #include "Physics/Physics.h" 7 8 #include <gsl/gsl_math.h> 9 #include <gsl/gsl_eigen.h> 10 #include <gsl/gsl_linalg.h> 11 12 13 MapAnalyser::MapAnalyser() 14 : mapAnalysis_m(IpplTimings::getTimer("mapAnalysis")) 15 , bunchAnalysis_m(IpplTimings::getTimer("bunchAnalysis")) 16 { } 17 18 //Analyzes a TransferMap for the tunes, symplecticity and stability 19 void MapAnalyser::linTAnalyze(const fMatrix_t& tMatrix){ -
@ganz_p some Doxygen description would be nice and helpful
added 1 commit
- 66932071 - added bunch and Hardedge in tests/Maps/FODODipole
marked the checklist item Code cleaning (@frey_m) as completed
added 8 commits
-
2ac63720...971d4db8 - 7 commits from branch
master - 698ea61c - Merge branch 'master' of gitlab.psi.ch:OPAL/src into OPAL-maps
-
2ac63720...971d4db8 - 7 commits from branch
47 48 SELECT, LINE=QUADTEST; 48 49 49 50 50 TRACK, LINE= QUADTEST, BEAM=BEAM1, MAXSTEPS=10000, DT=1.0e-10, ZSTOP=4.0, MAP_ORDER=8; 51 TRACK, LINE= QUADTEST, BEAM=BEAM1, MAXSTEPS=10000, DT=1.0e-10, ZSTOP=4.0, MAP_ORDER=2; -
@ganz_p : was this change in map_order on purpose? Should I also change this in the regression test?
With
MAP_ORDERyou can change the polynomial length of the map (, respective the truncation order).You can change it, if you like to.
The advantages:
If you set it to
MAP_ORDER=2the simulation would speed up. In my opinion the expansion would be sufficient for a regression test, because here the loop of the Lie expansion (ExpMapfunction inFTps) is then done twice (withMAP_ORDER=8its just repeated 6 times more...). Also the map accumulation would be faster, arguing similar as before (overloaded * operator inFVps).-
okay thanks for the info. I have changed the MAP_ORDERs to 2 in regression-tests/8b5f2002
mentioned in issue #227
marked the checklist item Doxygen documentation (@ganz_p) as completed
marked the checklist item Documentation (@ganz_p) as completed
marked the checklist item Documentation (@ganz_p) as incomplete
-
@ganz_p : The output section in the documentation is still empty. Could you add a few words there perhaps?
added 56 commits
-
2aa25b83...711b982f - 55 commits from branch
master - 9e63da24 - Merge branch 'master' into 'OPAL-maps'
-
2aa25b83...711b982f - 55 commits from branch
added 1 commit
- 458a3704 - disable fringe fields for SBends; remove trailing whitespace
marked the checklist item Documentation (@ganz_p) as completed
added 3 commits
-
e237ec11...9945e210 - 2 commits from branch
master - 37fe3016 - Merge branch 'master' into 'OPAL-maps'
-
e237ec11...9945e210 - 2 commits from branch
added 1 commit
- 9a686667 - fix bug: skip fringe fields in checkElementOrder_m; small improvements
marked the checklist item Regression tests (@snuverink_j) as completed
mentioned in merge request regression-tests!2 (merged)
