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

update INSTALL file #162

Closed
jeffhammond opened this issue Oct 31, 2015 · 2 comments
Closed

update INSTALL file #162

jeffhammond opened this issue Oct 31, 2015 · 2 comments

Comments

@jeffhammond
Copy link
Member

The INSTALL file is a bit old.

Here are some issues we should address:

  • Jaguar and Blue Gene/P are dead. Let's get the right build examples for contemporary machines like IBM Blue Gene/Q, Cray XC30, etc.
  • We do not mention Elemental as an optional capability.
  • PC examples related to Mac etc. reference relatively old compilers.
  • ...

We should not have examples involving OpenMPI, because it is not reliably thread-safe.

@gifornl
Copy link
Contributor

gifornl commented Nov 1, 2015

?Jeff,

Do you have a simple example of openmpi being not thread-safe in the mtm build for 1.8.4?

I have been using this with Intel 15 and gcc 4.8.2 builds. The sys-admins want a simple example that is repeatable...which is not so easy.

Thanks, George


From: Jeff Hammond notifications@github.com
Sent: Saturday, October 31, 2015 3:35 PM
To: m-a-d-n-e-s-s/madness
Subject: [madness] update INSTALL file (#162)

The INSTALL file is a bit old.

Here are some issues we should address:

  • Jaguar and Blue Gene/P are dead. Let's get the right build examples for contemporary machines like IBM Blue Gene/Q, Cray XC30, etc.
  • We do not mention Elemental as an optional capability.
  • PC examples related to Mac etc. reference relatively old compilers.
  • ...

We should not have examples involving OpenMPI, because it is not reliably thread-safe.

Reply to this email directly or view it on GitHubhttps://github.com//issues/162.

@jeffhammond
Copy link
Member Author

@gifornl I do not have a reproducer handy, because there is no dispute that it is broken (open-mpi/ompi#157). However, there appears to be a complete reproducer on the OpenMPI mailing list.

What's worse than OpenMPI's implementation of MPI_THREAD_MULTIPLE being consistently broken is that it seems to be correct, deadlock or crash, depending on the configuration and usage. I have no patience for such low-quality software.

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

No branches or pull requests

3 participants