[mvapich-discuss] MPI_Comm_spawn broken in 2.1rc1

[Jian Lin]

Hi, Nicolas,

MVAPICH2 v2.1rc2 is released. The issue on MPI_Comm_spawn has been
resolved. Please try out the new version and let us know whether your
program works or not. Thank you!

On Mon, 9 Mar 2015 18:41:22 +0100
Nicolas Niclausse <nicolas.niclausse at inria.fr> wrote:

> Hari Subramoni a écrit le 09/03/2015 18:32 :
> > Hi Nicolas,
> > 
> > Thanks for confirming that you are using the correct environment
> > variable. Can you please provide us with a reproducer so that we
> > can try it out locally?
> 
> 
> Yes, i was able to reproduce the problem even on a very basic example
> taken from:
> 
> http://mpi.deino.net/mpi_functions/mpi_comm_spawn.html
> 
> 
> #include "mpi.h"
> #include <stdio.h>
> #include <stdlib.h>
> 
> #define NUM_SPAWNS 2
> 
> int main( int argc, char *argv[] )
> {
>     int np = NUM_SPAWNS;
>     int errcodes[NUM_SPAWNS];
>     MPI_Comm parentcomm, intercomm;
> 
>     MPI_Init( &argc, &argv );
>     MPI_Comm_get_parent( &parentcomm );
>     if (parentcomm == MPI_COMM_NULL)
>     {
>         /* Create 2 more processes - this example must be called
> spawn_example.exe for this to work. */
>         MPI_Comm_spawn( "spawn_example", MPI_ARGV_NULL, np,
> MPI_INFO_NULL, 0, MPI_COMM_WORLD, &intercomm, errcodes );
>         printf("I'm the parent.\n");
>     }
>     else
>     {
>         printf("I'm the spawned.\n");
>     }
>     fflush(stdout);
>     MPI_Finalize();
>     return 0;
> }
> 
> 



-- 
Jian Lin
http://linjian.org