I don't think it's possible, but just to check: can you specify memory
affinity distinct from core binding somehow with OMPI (i.e. not with
hwloc-bind as a shim under mpirun)?
It seems to be relevant in Knight's Landing "hybrid" mode with separate
MCDRAM NUMA nodes as I assume you still want core
Brice Goglin writes:
> Hello
> Assuming this NVML detection is actually done by hwloc, I guess there's
> nothing in OMPI to disable it. It's not the first time we get such an
> issue with OMPI not having all hwloc's --disable-foo options, but I
> don't think we actually want to propagate all of t
Brice,
An other option is to add a --with-hwloc-flags configure option to Open
MPI, and pass the value to embedded hwloc configure.
We already do that for ROMIO (--with-io-romio-flags)
Cheers,
Gilles
On Tuesday, September 20, 2016, Brice Goglin wrote:
> Hello
> Assuming this NVML detection is
Hello
Assuming this NVML detection is actually done by hwloc, I guess there's
nothing in OMPI to disable it. It's not the first time we get such an
issue with OMPI not having all hwloc's --disable-foo options, but I
don't think we actually want to propagate all of them.
Maybe we should just force s
Fred,
Can you try to configure with
--disable-nvml
This is an option that should be passed to embedded hwloc.
Cheers,
Gilles
On Tuesday, September 20, 2016, Fred Mioux wrote:
> Hello ,
>
>
>
> I compile OpenMPI on a machine which support CUDA and NVML, but I don’t
> want to include this in m
Hello ,
I compile OpenMPI on a machine which support CUDA and NVML, but I don’t
want to include this in my distribution of OpenMPI :
-With cuda : I have the option –without-cuda
-With nvml : I have nothing excecpt "export enable_nvml=no" before
launch the configure.
Is there a
