[petsc-users] Memory leak in PetscRandom?

[Sascha Schnepp]

Hello,

when I run ksp/ksp/examples/tutorials/ex2 through valgrind with random exact vector enabled (-random_exact_sol) it shows some lost memory. Patrick Sanan discovered this playing around with random positions of multiple inclusions for ex43 but that is in a fork/branch of his. The part of the valgrind output concerning the memory loss for ex2 with -random_exact_sol is identical.

Cheers,
Sascha

sascha at geop-304 …/ksp/examples/tutorials [master ↓333|?19]
02/25/15 [17:29:26] $ valgrind --leak-check=full --dsymutil=yes  ./ex2 ./ex2 -ksp_monitor_short -m 5 -n 5 -ksp_gmres_cgs_refinement_type refine_always -random_exact_sol==68417== Memcheck, a memory error detector
==68417== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==68417== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
==68417== Command: ./ex2 ./ex2 -ksp_monitor_short -m 5 -n 5 -ksp_gmres_cgs_refinement_type refine_always -random_exact_sol
==68417==
--68417-- run: /usr/bin/dsymutil "./ex2"
--68417-- run: /usr/bin/dsymutil "/Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpifort.12.dylib"
--68417-- run: /usr/bin/dsymutil "/Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpicxx.12.dylib"
--68417-- run: /usr/bin/dsymutil "/Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpi.12.dylib"
--68417-- run: /usr/bin/dsymutil "/Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libpmpi.12.dylib"
  0 KSP Residual norm 2.28401
  1 KSP Residual norm 0.541581
  2 KSP Residual norm 0.114601
  3 KSP Residual norm 0.0109825
  4 KSP Residual norm 0.00112854
  5 KSP Residual norm 8.41066e-05
Norm of error 9.07246e-05 iterations 5
==68417==
==68417== HEAP SUMMARY:
==68417==     in use at exit: 44,377 bytes in 387 blocks
==68417==   total heap usage: 1,999 allocs, 1,612 frees, 328,837 bytes allocated
==68417==
==68417== 1,060 bytes in 1 blocks are possibly lost in loss record 85 of 95
==68417==    at 0x66BB: malloc (vg_replace_malloc.c:300)
==68417==    by 0x234DFC3: __emutls_get_address (in /opt/local/lib/libgcc/libgcc_s.1.dylib)
==68417==
==68417== 2,080 (1,040 direct, 1,040 indirect) bytes in 1 blocks are definitely lost in loss record 92 of 95
==68417==    at 0x66BB: malloc (vg_replace_malloc.c:300)
==68417==    by 0x25175AE: atexit_register (in /usr/lib/system/libsystem_c.dylib)
==68417==    by 0x25176E9: __cxa_atexit (in /usr/lib/system/libsystem_c.dylib)
==68417==    by 0x1E3CC27: _GLOBAL__sub_I_initcxx.cxx (initcxx.cxx:110)
==68417==    by 0x7FFF5FC3D15F: ???
==68417==    by 0x1E3C28F: MPI::Datatype::Get_name(char*, int&) const (in /Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpicxx.12.dylib)
==68417==    by 0x10480D61F: ???
==68417==    by 0x7FFF5FC11C2D: ImageLoaderMachO::doModInitFunctions(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==68417==    by 0x7FFF5FC3D15F: ???
==68417==    by 0x100000011: ??? (in ./ex2)
==68417==    by 0x1E35297: ??? (in /Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpicxx.12.dylib)
==68417==    by 0x1E3555F: ??? (in /Users/sascha/Documents/codes/PETSc/petsc-dev/arch-osx-master-debug/lib/libmpicxx.12.dylib)
==68417==
==68417== LEAK SUMMARY:
==68417==    definitely lost: 1,040 bytes in 1 blocks
==68417==    indirectly lost: 1,040 bytes in 1 blocks
==68417==      possibly lost: 1,060 bytes in 1 blocks
==68417==    still reachable: 5,318 bytes in 15 blocks
==68417==         suppressed: 35,919 bytes in 369 blocks
==68417== Reachable blocks (those to which a pointer was found) are not shown.
==68417== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==68417==
==68417== For counts of detected and suppressed errors, rerun with: -v
==68417== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 17 from 17)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: configure.log
Type: application/octet-stream
Size: 3786789 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150225/318e7a2c/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: make.log
Type: application/octet-stream
Size: 97340 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150225/318e7a2c/attachment-0003.obj>