On 12 November 2010 11:09, Jed Brown <jed at 59a2.org> wrote: > On Fri, Nov 12, 2010 at 15:04, Jed Brown <jed at 59a2.org> wrote: >> >> The comparison we are more interested is with an equivalent number of MPI >> processes per node. > > Whoops, missed your second attachment. ?Indeed, that is what I expect, > especially at relatively low node counts. ?I think it has been fairly well > demonstrated that straight MPI does alright for weak scaling with decent > sized subdomains. ?The trouble, and the place that threading (OpenMP or > otherwise) offers an advantage, is for the very small subdomains you end up > with when strong scaling.
Jed, Rodrigo's results are strong scaling. Unfortunately, we just had 4 cores I7 and 8 cores Xeon per node to test. -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169
