I don't know who exactly is on the bug tracking list, so I thought I'd
update here that I finally ran a variety of OMPI r numbers through MTT
and measured the performance with NetPIPE/sm on a woodcrest-class
x86_64 machine. See my update on #1250:
https://svn.open-mpi.org/trac/ompi/ticket/1250
--
Jeff Squyres
Cisco Systems