OMPI 5.x has no support for the openib BTL, all IB traffic is now going
through the UCX PML. This means that `-mca btl_openib_if_include XXX` is
meaningless, but you can use the UCX_NET_DEVICES to direct UCX to a
specific device.
As the error happens for UD you can switch to a different transport
I checked further using a modification to mpi_hello_world.c (that includes
MPI_Barrier) and a test code that checks connectivity between all processes.
1. On the mpi_hello_world_barrier.c case, openmpi5 failed the same way as
before. mpich-ofi completed without error.
2. On the connectivity_c.c c
UCX1.18.0
On Wednesday, July 2, 2025 at 3:38:57 PM UTC-4 Achilles Vassilicos wrote:
> I checked further using a modification to mpi_hello_world.c (that includes
> MPI_Barrier) and a test code that checks connectivity between all processes.
> 1. On the mpi_hello_world_barrier.c case, openmpi5 fai
A lot to chew on. Thanks.AchillesSent from my iPhoneOn Jul 2, 2025, at 5:40 PM, George Bosilca wrote:OMPI 5.x has no support for the openib BTL, all IB traffic is now going through the UCX PML. This means that `-mca btl_openib_if_include XXX` is meaningless, but you can use the UCX_NET_DEVICES to
UCX 1.8 or UCX 1.18 ?
Your application does not exchange any data so it is possible that MPICH
behavior differs from OMPI (aka not creating connections vs creating them
during MPI_Init). That's why running a slightly different version of the
hello_world with a barrier would clarify the connection'
