I am using v4_9_1_p01" on Linux+2.6 on Fermi Scientific Linux. The /vis/scene/add/trajectories/smooth/rich command does not work on that machine; I shipped a tarball to my Intel Mac, ran it there, and it worked fine. I don't see smooth helices on the Linux box, it's like the command doesn't get turned on. The vis.mac file is below. --thanks, Robert Bernstein # $Id: vis.mac,v 1.6 2007/06/13 10:36:39 maire Exp $ # # Macro file for the initialization phase of "exampleN03.cc" # when running in interactive mode # # Sets some default verbose # /control/verbose 2 /control/saveHistory #/run/verbose 2 # # create empty scene # /vis/scene/create # # Create a scene handler for a specific graphics system # (Edit the next line(s) to choose another graphic system) # # Use this open statement to get an .eps and .prim files # suitable for viewing in DAWN. ###/vis/open DAWNFILE # # Use this open statement instead for OpenGL in immediate mode. /vis/open OGLIX # # Use this open statement instead to get a HepRep version 1 file # suitable for viewing in WIRED. /vis/open HepRepFile # # Use this open statement instead to get a HepRep version 2 file # suitable for viewing in WIRED. #/vis/open HepRepXML # # Output an empty detector #/vis/viewer/flush # # Draw trajectories at end of event, showing trajectory points as # markers of size 2 pixels /tracking/verbose 1 /vis/scene/add/trajectories smooth rich /vis/modeling/trajectories/create/drawByCharge /vis/modeling/trajectories/drawByCharge-0/default/setDrawStepPts true /vis/modeling/trajectories/drawByCharge-0/default/setStepPtsSize 2 /vis/viewer/set/viewpointThetaPhi -20. 20. # (if too many tracks cause core dump => /tracking/storeTrajectory 0) # # To draw gammas only #/vis/filtering/trajectories/create/particleFilter #/vis/filtering/trajectories/particleFilter-0/add gamma # # To draw charged particles only #/vis/filtering/trajectories/particleFilter-0/invert true # # Many other options available with /vis/modeling and /vis/filtering. # For example, select colour by particle ID #/vis/modeling/trajectories/create/drawByParticleID #/vis/modeling/trajectories/drawByParticleID-0/set e- red # /vis/scene/endOfEventAction accumulate # # At end of each run, an automatic flush causes graphical output. /run/beamOn 1 # When you exit Geant4, you will find a file called scene-0.heprep.zip. # Unzipping this will give you three separate HepRep files suitable for # viewing in WIRED. # The first file will contain just detector geometry. # The second file will contain the detector plus one event. # The third file will contain the detector plus ten events.
This has turned out to be an elusive problem to reproduce, but we are now seeing it in some other test applications. We believe it involves an issue of the precision of G4PropagatorInField for areas where magnetic field is set to zero. Only affects visualization, not the actual tracking. We'll update the status when we have a workaround or a fix.
The origin of the problem was the definition of the Magnetic Field. In A01 example the field value was mostly zero and non-zero only for Cylinder. In this case the solution was to use local constant field for Cylinder instead of global field for all World. It solves the problem with trajectory and also makes the tracking much more efficient, because the mechanism for Field Integration is called only for the Cylinder. The problem has been fixed for Geant4 release 9.3.
