Closed wilfonba closed 4 days ago
Ok so the code you wrote throws an error and everything fails everywhere, though this prevents any of the benchmarks from running at all, or at least fails basically immediately. So, we aren't really seeing all of the checks you were hoping to check for (I think).
I would expect all of the test suites to fail on the first bubbles case since I added an MPI_Abort if bubbles = T. I expected the benchmarks to handle this though since it runs the 4 cases independently. Looking at the CI it appears that neither benchmark actually got a node and ran at all. It could be the case that I needed to introduce a seg fault rather than an MPI_Abort, but I'll have to see what happens when the benchmarks actually run to know.
Please sync this branch with master
I'd like to see this check get to the point where it gets to point where it actually prints benchmark results instead of failing during compilation (presumably with the aforementioned N/A
's in the appropriate places).
The only thing I really want to be checked now is if the benchmark cases run on Delta/non-Phoenix computers.
I'll do a run on delta
I just did. Check the issue for the latest at #396
I'm not really sure what's going on.
this can be closed due to #497 thanks @ChrisZYJ
This is a mock PR that checks the benchmarking PR. It does the following: