vasilis gkanis wrote:
>I had a similar problem with the portland Fortran compiler. I new that this
>was not caused by a network problem ( I run the code on a single node with 4
>CPUs). After I tested pretty much anything, I decided to change the compiler.
>I used the Intel Fortran compiler and everything is running fine.
>It could be a PGI compiler voodoo :)
There were some thoughts on this e-mail thread that the problem could be
related to trac ticket 2043. Note that there has been progress on this
ticket. See https://svn.open-mpi.org/trac/ompi/ticket/2043#comment:18
. The shared-memory (on-node) communications were subject to race
conditions that could be exposed by optimizing compilers. Some signals
could have gotten lost in inter-process communications, quite possibly
leading to hangs.
If you think you got bitten by this bug, please try the revisions
mentioned in the trac ticket and report your success (or, alas, failure)
via the trac ticket or as appropriate.