Skip to content

Conversation

@csegarragonz
Copy link
Collaborator

@csegarragonz csegarragonz commented Nov 12, 2021

In this PR I include two distributed tests for MPI.

  • One runs a broadcast across two hosts and works just fine.
  • The other one runs two long-lasting all-to-all functions and is expected to fail (so it does). I commit it commented out as I think it is a good starting point.

I also bump the code version to re-build the containers as required in faasm/faabric#177.

@csegarragonz csegarragonz self-assigned this Nov 12, 2021
msg.set_mpiworldsize(4);

// Call the functions
sch.callFunctions(req);
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It would be nice to somehow test that these had been executed on different hosts. Could we perhaps check the execution graph? Would be a nice indirect test of that working in a distributed environment too.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, I think this is a very valid point. I add a utility function in faasm/faabric#178 and then check for it here.

@csegarragonz csegarragonz marked this pull request as draft November 22, 2021 16:11
@csegarragonz csegarragonz marked this pull request as ready for review November 22, 2021 20:03
@csegarragonz csegarragonz force-pushed the mpi-dist-test branch 3 times, most recently from 2d7fd71 to 60308b2 Compare November 23, 2021 16:59
@@ -1,4 +1,4 @@
FROM faasm/faabric-base:0.2.0
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was needed to merge the latest conan changes in faabric.

@csegarragonz csegarragonz merged commit d49d974 into master Nov 24, 2021
@csegarragonz csegarragonz deleted the mpi-dist-test branch November 24, 2021 16:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants