Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add rdma #606

Closed
wants to merge 5 commits into from
Closed

Add rdma #606

wants to merge 5 commits into from

Conversation

aricer123
Copy link
Contributor

Description

Please include a summary of the changes and the related issue(s) if they exist.
Please also include relevant motivation and context.

Fixes #598

Adds flag for rdma for testing and passes to ./mfc sh run
Compiles on macbook and runs tests but needs to be tested on gpu with rdma, may have errors

Copy link
Member

@sbryngelson sbryngelson left a comment

Choose a reason for hiding this comment

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

Adds flag for rdma for testing and passes to ./mfc sh run

This PR indeed does this but does not do anything differently within the ./mfc.sh run environment. It should ensure the RDMA_MPI = 'T' in the generated case.py file.

@henryleberre
Copy link
Member

The changes in the toolchain folder can be undone. See #598 (comment).

To clarify, when you ./mfc.sh test -- -c phoenix -- --rdma this is what happens when we follow the execution of a single test:

  • We generate a case.py file in the tests/<case slug>/ folder.
  • We execute ./mfc.sh run tests/<case slug>/case.py -c phoenix -- --rdma. There are more arguments but you don't need to worry about these.
  • When ./mfc.sh run executes, it basically runs the tests/<case slug>/case.py file and passes the --rdma flag.

So there is no need to have the --rdma file registered with ./mfc.sh test or ./mfc.sh run. The case file that is autogenerated should be the one that parses this argument.

@aricer123
Copy link
Contributor Author

  • We generate a case.py file in the tests/<case slug>/ folder.

Where in the project/program is the case.py file generated? From my understanding each case.py contains the configurations for its respective test/example and is manually written.

I understand the rest of how the test and run features work as well as how the arguments including the case file are passed.

@henryleberre
Copy link
Member

  • We generate a case.py file in the tests/<case slug>/ folder.

Where in the project/program is the case.py file generated? From my understanding each case.py contains the configurations for its respective test/example and is manually written.

I understand the rest of how the test and run features work as well as how the arguments including the case file are passed.

This is where we declare the parser of the case.py files used for testing:

parser.add_argument("dict", type=str, metavar="DICT", help=argparse.SUPPRESS)
. You need to add the --rdma flag and make it set rdma_mpi={'T' if rdma else 'F'} in the case dictionary printed by the case file.

@aricer123
Copy link
Contributor Author

got it thanks, i missed that when i was going through the files

@aricer123 aricer123 closed this Sep 5, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Development

Successfully merging this pull request may close these issues.

Add rdma_mpi to test suite
3 participants