[MITgcm-support] MPI + shared memory

[chris hill]

Hi Paola,

  I took a quick look at this after Hawaii.
  I got as far as compiling and running and looking at some performance 
tools that were on the ORNL web site. However, the results I got from 
various runs were all somewhat slow in performance, even for very simple 
vector loops. Constantinos was going to take a closer look but I believe 
he got bogged down in some work for Patrick and Carl. I'll check with 
him tomorrow.

Chris
Paola Cessi wrote:
> Is there a MPI compilation option that needs to be activated in genmake 
> in order to utilize shared memory multi-threading, together with 
> distributed memory parallelism in machines that have multiple nodes each 
> with a multi-core (dual or quad) FPU's (i.e. the CRAY X1E or a cluster 
> of dual-core processors)?
> 
> Thanks,
> 
> paola
> 
> -------------------------------------------------------------------------
> Paola Cessi                            Tel: +1 858 534 0622
> Scripps Institution of Oceanography    Fax: +1 858 534 8045
> 9500 Gilman Drive #0213                e-mail: pcessi at ucsd.edu
> La Jolla, CA 92093-0213
> USA                      Web:  http://www-pord.ucsd.edu/~pcessi
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>