I mentioned uDapl or SDP specifically in terms of the typical MPI usage today - connection oriented.
With both uDapl and SDP the connection requirments were the same so I included them in the same column. I did not intend to indicate Sdp implementation on uDapl.. On 11/7/05, James Lentini <[EMAIL PROTECTED]> wrote: > > > On Mon, 7 Nov 2005, Grant Grundler wrote: > > > On Mon, Nov 07, 2005 at 09:51:48AM -0500, James Lentini wrote: > > ... > > > > Procs per node uDapl/Sdp Rds > > > > 2 19996 9999 > > > > 4 39984 9999 > > > > > > > > Clearly, there is tradeoff in performance as we go from uDapl/Sdp to > > > > Rds. > > ... > > > This isn't an apples to apples comparison. uDAPL is an API and RDS is > > > a protocol. > > > > I thought he was comparing SDP (using uDAPL) vs RDS. > > Did I read that wrong? > > I think so. I'll let Ranjit clarify. > > To the best of my knowledge, SDP has never been implemented on uDAPL. > _______________________________________________ > openib-general mailing list > [email protected] > http://openib.org/mailman/listinfo/openib-general > > To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general > _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
