Fix Geant4 configuration

With the additional parameter specialCuts the energy based Geant3 cuts are automatically converted to the range based Geant4 cuts by VMC. Without the parameter the default Geant4 cuts are used. The change reduced the number of secondaries by a factor of 5 or more for a normal CbmRoot UrQMD simulation and speed up the runtime by factors. The runtime compared to Geant3 went from being slower by a factor of 4 to being slower by 30 %. The change of the maximum number of steps from 1e7 to the correct number of 10000 has no effect. There are no tracks exceeding 10000 steps such that there is no additional speedup.

added BugFix Framework Simulation labels

requested review from @v.friese

assigned to @p.-a.loizeau

added 1 commit

• 62eb14b0 - Apply clang format

Compare with previous version

Dear @f.uhlig, @v.friese, @p.-a.loizeau,

you have been identified as code owner of at least one file which was changed with this merge request.

Please check the changes and approve them or request changes.

added CodeOwners label

I will test these changes, but to be frank, I am a little surprised by this change makes any difference as I thought it was already exactly the Geant4 configuration which we had already.

I had checked these parameters in sim/transport/gconfig/g4Config.C not sim/transport/steer/CbmGeant4Settings.h** and change you make** is already present, i.e.

line 32

TG4RunConfiguration* runConfiguration = new TG4RunConfiguration("geomGeant4", "QGSP_BERT_EMV", "stepLimiter+specialCuts");

and

line 77

geant4->SetMaxNStep(10000);

I would strongly suggest we remove sim/transport/gconfig/g4Config.C if it does nothing or is overwritten by steer/CbmGeant4Settings.h** as it fooled me.**

requested review from @e.clerkin and removed review request for @v.friese

removed review request for @e.clerkin

requested review from @v.friese

This was tested for beam transport with Trajectories turned on, there was a remarkable saving in terms of run-time for Geant4, which was completely unusable for my task before implementation.

The difference in time per event between Geant3 and Geant4 on 20 events was approximately 10%. However it should be stressed that Geant3 and Geant4 where not giving identical beam trajectories, with many beam ion terminating before the end of the beampipe/cave, thus perhaps giving an unfair time comparison.

@v.friese,

is there an issue why you don't approve the MR?

No. Please attend to the comments by Eoin.

resolved all threads

added 27 commits

• 62eb14b0...966663ab - 25 commits from branch computing:master
• f3b808cf - Fix Geant4 configuration
• 8d79c728 - Apply clang-format

Compare with previous version

@v.friese,

all issues were fixed.

approved this merge request