Artem Polyakov

I think if this is not concise report, below you’ll see the detailed info about all collectives involved.

@ggouaillardet what was the original issue? Can you provide the reference? We have seen issues with Slurm daemons starving of CPUs when spinning in the direct modex case. In this...

@devreal do you mean that master with #6980 fails on 3 tests?

This happened to us when address randomization was enabled on the nodes. @cchambreau If that’s the case, I think that it can be addressed by performing resolutions on each rank...

@cchambreau we were trying to profile some MPI application and noticed that the call stack is broken. IMO it’s one of the greatest features in mpiP so we thought it...

@sb17v patch looks clean, thanks!

@janjust please change the commit message as follows: ``` Co-authored-by: Artem Polyakov Co-authored-by: Sergey Lebedev ``` Per https://docs.github.com/en/free-pro-team@latest/github/committing-changes-to-your-project/creating-a-commit-with-multiple-authors

I tried it out of curiosity and it works as expected: https://github.com/artpol84/xccl/commit/91a6466ba984109480c51fc7125559fdcc0b97d6

Yeah With this I was able to pass through simx device to the L2 OS. Still need to fix mounts forwarding, doesn’t work now so I can’t load Mellanox module...

@jgunthorpe this is not intended for merge. It’s an FYI to indicate what we had to hack to make it work on the particular system. we agreed with @rleon that...