Dear All, The code I work on experiences significant memory leak, reported as follows by valgrind:
16,384 bytes in 2 blocks are definitely lost in loss record 2,246 of 2,367 ==2110== at 0x10041B3B1: malloc (vg_replace_malloc.c:303) ==2110== by 0x1009367C7: operator new(unsigned long) (in /usr/local/lib/libstdc++.6.dylib) ==2110== by 0x10002DF95: ISPCAlloc (tasksys.cpp:1191) ==2110== by 0x100201D8C: forces_sse4 (forces.ispc:1287) ==2110== by 0x3F8000003F7FFFFF: ??? ==2110== by 0x10364286F: ??? The forces function is called many times (every time step during a simulation) and it launches several taks: here is the reported line 1287 <https://github.com/tkoziara/parmec/blob/master/forces.ispc#L1287>. The ISPC version is 1.9.1 and the tasksys.cpp file used here comes from this version. The line 1191 in tasksys.cpp can be viewed here <https://github.com/ispc/ispc/blob/master/examples/tasksys.cpp#L1191>. I hope that you would be able to comment about why there could be a memory leak like this. The software is indeed eating up memory when the "springs_task" at line 1287 is being called -- and not doing this when this line is commented out. There is not memory allocation within the task function. Best regards, Tomek -- You received this message because you are subscribed to the Google Groups "Intel SPMD Program Compiler Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
