Could you model it with another nonblocking collective operation? I notice, for example, that MPI_Iallreduce is supported.
On 17 October 2014 12:36, Martin Sandve Alnæs <[email protected]> wrote: > Thanks, but ibarrier doesn't seem to be in the stable version of openmpi: > http://www.open-mpi.org/doc/v1.8/ > Otherwise mpi_ibarrier+mpi_test+homemade time/sleep loop would do the > trick. > > Martin > > > On 17 October 2014 12:23, David Ham <[email protected]> wrote: > >> Mpi_ibarrier? >> >> http://www.mpich.org/static/docs/v3.1/www3/MPI_Ibarrier.html >> >> On 17 October 2014 10:58, Martin Sandve Alnæs <[email protected]> wrote: >> >>> How can I achieve an MPI barrier with a timeout? I'd like to start >>> some of the tests with garbage collection and an mpi barrier with a timeout >>> so we can exit semi-gracefully and report a point where things fail. >>> >>> A workaround I can think of is to run MPI.barrier in a python thread >>> and poll that thread until a timeout and then die if it doesn't return. I'm >>> not concerned with the fallout when blowing up here (or should I be?) >>> because this would probably be better than a deadlock and a later kill from >>> the outside anyway. >>> >>> Martin >>> >> >> >> >> -- >> Dr David Ham >> Departments of Mathematics and Computing >> Imperial College London >> >> http://www.imperial.ac.uk/people/david.ham >> > > -- Dr David Ham Departments of Mathematics and Computing Imperial College London http://www.imperial.ac.uk/people/david.ham
_______________________________________________ fenics mailing list [email protected] http://fenicsproject.org/mailman/listinfo/fenics
