I wrote the following dummy MPI program, and followed the wiki page to
generate the suppression file. It helps a bit, but still generates a lot
of unwanted ones. For a 8-process run, I got the output file with over
45000 lines.
Is this mainly due to MPI program? I remember using Linux version
Sherry,
I think the memory leak is our mistake. Thanks for checking the
uninitialized copies
Barry
> On Jan 16, 2016, at 2:04 PM, Xiaoye S. Li wrote:
>
> The first leak is related to data structure for distributed sparse CSR matrix
> A (input). How does the petsc interface free A's
Sherry, Satish:
Building MPI from source on Cray machines is not trivial, because Cray MPI
is proprietary.
However building MPICH on Cray is actually possible now as of a month or
two ago thanks to libfabric, but there are approximately two people on
earth who have done it successfully.
If it is
On Sat, 16 Jan 2016, Xiaoye S. Li wrote:
> By the way, I have a question about valgrind. Lately I have been chasing
> a possible memory corruption on Cori at NERSC. Valgrind gave LOTS OF
> internal system-related warnings, such as:
>
> ==39059== Conditional jump or move depends on uninitialis
Sherry,
We think there is a small memory leak in Superlu; could you check it?
> ==11431== 40 bytes in 1 blocks are definitely lost in loss record 1 of 1
> ==11431==at 0x4C2B6CD: malloc (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==11431==by 0x67B5CBD: superlu_malloc_d