Hello Roy,
I checked that there is no infinite loop.
Both ranks pass this->send (dest_processor_id, sendvec, type1, req,
send_tag);
and hang on this->receive (source_processor_id, recv, type2, recv_tag);
Both processes are sending zero elements, is this correct?
Could you, please, suggest a simple MPI test to mimic this situation?
Then I can check the MPI implementation.
Thank you,
Michael.
On 09/10/2017 09:07 PM, Roy Stogner wrote:
On Sat, 9 Sep 2017, Michael Povolotskyi wrote:
(gdb) p recv_tag
$2 = (const libMesh::Parallel::MessageTag &) @0x2ac9d6d3d980: {
static invalid_tag = -2147483648, _tagvalue = -1, _comm = 0x0}
(gdb) p recv_tag
$2 = (const libMesh::Parallel::MessageTag &) @0x2b9331482980: {
static invalid_tag = -2147483648, _tagvalue = -1, _comm = 0x0}
So, does it look like a logic error or like a problem with openMPI
installation?
In openmpi MPI_ANY_TAG==-1, so they're both set to the right thing,
not waiting for the a message that will never come.
Could you, after attaching the debugger, try stepping forward in both
processes, and make sure they're both hanging, not infinite looping?
---
Roy
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users
