On 11/15/2013 5:12 PM, Ralph Castain wrote:
Perhaps if Pasha or somebody else proficient in the OMPI code could help out,
then the iWARP CPC could be moved. W/O help from OMPI developers, its going to
take me a very long time...
I believe we would all be willing to provide advice - we just have no way of
testing.
I have lots of testing facilities. :)
Can't we leave the openib rdma CPC code as is until we can get the rdmacm CPC
moved into OFACM. What is the harm with that exactly? I mean, if no iWARP
devices support these accelerated MPI collectives, then leave the rdmacm CPC in
the openib btl so we can at least support iWARP via the openib BTL...
I see no reason why we can't just push the rdma over to ofacm - I'd prefer that
to leaving the code in the openib btl. Forcing the openib btl to use both cpc's
from ofacm AND its own would be ugly.
If you or someone can do the move and make it compile, I could then
test/debug/resolve all the other issues and make it work. I just balk
at the task of moving the stuff over because of my lack of knowledge of
the OMPI code above the openib/connect code.
Steve.