Skip to content

Defect: Failed images support not enabled during tests #437

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

Closed
afanfa opened this issue Aug 22, 2017 · 3 comments
Closed

Defect: Failed images support not enabled during tests #437

afanfa opened this issue Aug 22, 2017 · 3 comments
Labels

Comments

@afanfa
Copy link
Contributor

afanfa commented Aug 22, 2017

Sorry for not using the template but I believe this issue should be mostly used for discussion.
Is the --disable-auto-cleanup string passed at runtime for the failed images tests?
Is the --enable-error-checking=all string passed at configure time for mpich?

@zbeekman
Copy link
Collaborator

Interesting...

  • yes -disable-auto-cleanup is passed to mpiexec when running failed images tests. Do you have evidence to the contrary?
  • As for --enable-error-checking=all being passed at configure time for MPICH, I'm not 100% sure. The only circumstance under which we install MPICH as part of the installation process is when using install.sh. I'm guessing, as of right now, that we don't pass --enable-error-checking=all at configure time for MPICH.

Is MPICH being built with --enable-error-checking=all a required condition for failed images to work properly? Do you know how to detect if MPICH was built this way?

Right now, the logic for enabling failed images is:

  1. GCC/GFortran >= 1? Yes: try to enable failed images & proceed to 2. No: disable failed images
  2. Are the required proposed/experimental MPI functions available? Yes: Enable failed images, No: do not enable failed image support

In addition, failed images tests are only run if OPENCOARRAYS_DEVELOPER=TRUE (or 1 or YES or ON)

@zbeekman
Copy link
Collaborator

According to the MPICH wiki --enable-error-checking=all is on by default: https://wiki.mpich.org/mpich/index.php/Fault_Tolerance#Usage .. double checking the output of ./configure for MPICH now, to confirm

@zbeekman
Copy link
Collaborator

  --enable-error-checking=level
      Control the amount of error checking.
        no        - no error checking
        runtime   - error checking controllable at runtime through environment
                    variables
        all       - error checking always enabled (default)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants