Mathieu,
Your MPI_COMM_UNIVERSE is a inter-communicator, and therefore the MPI_Comm_size
and MPI_Comm_rank return the size, respectively the rank, in the local group.
There is a special accessor for getting the remote group size
(MPI_Comm_remote_size).
Now regarding the previous question (abou
Another question on the same example.
When I ask the size on the inter-communitator (MPI_COMM_UNIVERSE in the
example) between the spaner/parent and the spawned/children processes,
the same number of processes than in MPI_COMM_WORLD is returned. I do
not really understand because I expected m
Steve, thank for the details.
What is the command line that you use to run the benchmark ?
Can you try to add follow mca parameters to your command line:
"--mca btl openib,sm,self --mca btl_openib_max_btls 1"
Thanks,
Pasha
Repsher, Stephen J wrote:
Thanks for keeping on this Hopefully this
Hello,
I am trying to use MPI_Comm_spawn (MPI-2 standard only) and I have an
problem when I use MPI_UNIVERSE_SIZE. Here my code:
int main( int argc, char *argv[] )
{
int wsize=0, wrank=-1 ;
int usize=0, urank=-1 ;
int ier ;
int usize_attr=0, flag=0 ;
MPI_Comm MPI_COMM_UNIVERSE;
ie
Thanks for keeping on this Hopefully this answers all the questions:
The cluster has some blades with XRC, others without. I've tested on both with
the same results. For MVAPICH, a flag is set to turn on XRC; I'm not sure how
OpenMPI handles it but my build is configured --enable-openib-con
Thanks George. I assume we need this in 1.4.2 and 1.5, right?
On Feb 17, 2010, at 6:15 PM, George Bosilca wrote:
> I usually prefer the expanded notation:
>
> unsigned char ret;
> __asm__ __volatile__ (
> "lock; cmpxchgl %3,%4 \n\t"
>
Hey,
I only may to add the XRC and RC have the same latency.
What is the command line that you use to run this benchmark ?
What is the system configuration (one hca, one active port ) ?
Any addition information about system configuration, mpi command line,
etc. will help to analyze your issue.
