If you care, this mechanism is an artifact of Open MPI's build system: we made
the choice to allow our plugins to be built and loaded dynamically (i.e., each
plugin as its own dynamic shared object) or built statically (i.e., each plugin
is physically part of, for example, libmpi.so or libmpi.a)
Dear all,
Playing around with GPUDirect v1 and Infiniband I noticed that once the payload
is non-contiguous no RDMA is used at all.
Can anybody confirm this?
I'm using Open MPI 1.4.3. If the above is true, has this behavior changed with
later versions of Open MPI?
Thanks a lot.
Best,
Sebasti
Your observations are correct. If the payload is non-contiguous, then RDMA is
not used. The data has to be copied first into an intermediate buffer and then
sent.
This has not changed in later version of Open MPI.
Rolf
>-Original Message-
>From: devel-boun...@open-mpi.org [mailto:de
