-
Notifications
You must be signed in to change notification settings - Fork 555
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
Updates to make WW3 thread safe #315
Updates to make WW3 thread safe #315
Conversation
- update from @JamesAbeles-NOAA
machines for future flexibility of having '*.gnu' options
W3SREF which is called in w3srce which is threaded
As a sanity check, I ran the WW3 regression tests with OMP, OMPH tests with a different number of threads and the only regression test that failed is known to fail: ww3_tp2.10/./work_MPI_OMPH -- granted I tried this also with the develop branch and got the same result. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @JessicaMeixner-NOAA
These changes all look very sensible.
I will run some relevant regression tests next week.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @JessicaMeixner-NOAA ,
I've run the regression tests for the hybrid MPI-OMP tests on our system and the results are all identical with those from the develop branch. I'm happy to sign this one off now.
This adds thread private declarations for saved variables (Issue #307) for non-PDLIB cases. There is a long outstanding issue #214 to add regression tests for testing OMP threads and this PR does not address this long standing need. There will be a regression test added for the UFS system for which this bug was originally found. While only the TAUNX/Y in w3srce needed to be added to solve that case, the principal for why that was the fix was applied to the variables I found (except PDLIB). Thanks again to @JamesAbeles-NOAA for finding the solution!
Additionally in this PR two files were removed that should have been removed in PR #284 but were missed when combining a few PRs together.
Lastly, this PR also included porting #306 for WW3 to be used in UFS applications.
Only the usual regtests had differences.
matrixout20200218.tar.gz