| Summary: | Memory leak when using the "new" command /grdm/setRadioactiveDecayFile | ||
|---|---|---|---|
| Product: | Geant4 | Reporter: | stuff |
| Component: | processes/hadronic/models/radioactive_decay | Assignee: | desorgher |
| Status: | REOPENED FIXED | ||
| Severity: | critical | CC: | dennis.herbert.wright |
| Priority: | P5 | ||
| Version: | 9.5 | ||
| Hardware: | Apple | ||
| OS: | Mac OS X | ||
| Attachments: |
patch to fix bug
screen dump for massif heap profiler |
||
Dear Joa, Many thanks for the bug report! Indeed for the cases of user defined radioactive decay file., the nuceli were re-loaded indefinitively. Some of the solutions that you proposed was already brought to the code for solving a problem with the life time of the user defined radioactive decay file. But for what concern the stop of reloading of the same file, I did correct the code following your suggestions. Best regards Laurent Created attachment 186 [details]
screen dump for massif heap profiler
Hi all, it seems as if this bug has resurfaced with geant4.9.6.b01. See included screen dump with massif heap profiler I presently do not have the time to fix this. cheers Joa |
Created attachment 174 [details] patch to fix bug Hi, I have a simulation where I create a lot of radioactive nuclei and these decay. If I set my own decay file I noticed that the programmed filled my disk by "swapping". This did however not slow down the program, this because the created leaked object was never used again... Difficult to find since the actually used memory of the code remains the same on any system with swap (i.e all). However, I think I found the problem, and I include a patch. When using a user set decay file the code did not keep track of the fact that it had already loaded decay data to it was done every time I created an ion. cheers Joa