Ticket #11375 (closed: fixed)
Fix MPI build (for use on fermi)
| Reported by: | Stuart Campbell | Owned by: | Stuart Campbell |
|---|---|---|---|
| Priority: | major | Milestone: | Release 3.4 |
| Component: | Framework | Keywords: | |
| Cc: | Blocked By: | ||
| Blocking: | Tester: | VickieLynch |
Description
The MPI build of mantid that we use on fermi.ornl.gov has been broken for a very long time.
Change History
comment:1 Changed 5 years ago by Stuart Campbell
- Status changed from new to verify
- Resolution set to fixed
comment:2 Changed 5 years ago by Stuart Campbell
To test, you can just look at the status of the job (http://builds.mantidproject.org/job/master_clean_rhel6_mpi_systemtested/) - which at the moment is looking at this branch, so if you merge this, please DO NOT delete this branch until I have changed the job on jenkins.
comment:4 Changed 5 years ago by anonymous
I want to see if this works on my rhel6 machine. I will not merge until you have removed the In Progress label.
comment:5 Changed 5 years ago by Stuart Campbell
There was a little nastiness that I did in order to get the test data in the correct place, you can see the full horror in the job config. It is basically down to the Testing directory in parallel to the Framework (which we are running cmake from and therefore it can't find it)
I also found that I had to use LD_PRELOAD to point at the MPI library.
comment:6 Changed 5 years ago by VickieLynch
- Status changed from verifying to closed
- Tester set to VickieLynch
Merge pull request #605 from mantidproject/11375_fix_mpi_build
Changes to make MPI build work again
Full changeset: 339ddad0ebf99500aab78f7323072300ab323a54
This is being verified as pull request #605.