| Summary: | Infinite tracking loop using Sphere/Orb solids | ||
|---|---|---|---|
| Product: | Geant4 | Reporter: | mkippen |
| Component: | geometry/solids | Assignee: | Vladimir.Grichine |
| Status: | RESOLVED FIXED | ||
| Severity: | critical | CC: | Gabriele.Cosmo, John.Apostolakis |
| Priority: | P1 | ||
| Version: | 6.1 | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| URL: | http://home.lanl.gov/mkippen/g4bugs/testbug.tgz | ||
Thank you for this error report and the very clear and most useful test case. I have identifed now the problem in the geometry that leads to this error. We now also have a few line program that reproduces the problem that leads to this infinite loop. We will be investigating this in the coming week. In addition the Navigator is revised in Geant4 6.2. It has been modified to attempt to recover from getting stuck as in this case or else abandon the relevant event after warning the user. Fix introduced in geom-csg-V06-02-00 |
Problem: infinite tracking loop encountered when tracking (~0.1 - 10 MeV) electrons through a simple geometry. Electrons become trapped at the boundary between two NESTED volumes. Geometry: Master = G4Box (Vacuum) Daughter = G4Orb (Aluminum) Daughter1 = G4Orb (Tantalum) Daughter2 = G4Orb (Vacuum) Daughter3 = G4Tubs (Silicon) Physics: Standard EM All volumes are nested so that there are no two volumes on the same Tree level. Particles become trapped at the boundary of the outermost sphere. The problem occurs rarely (1 out of 10^7 incident electrons) and is independent of the incident electron energy. The problem also occurs replacing G4Orb with G4Sphere, and replacing G4Box with G4Orb or G4Sphere. A simple example code showing the problem is available at the attached web site (compressed tar file).