You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Mar 20, 2023. It is now read-only.
. I see though that for some reason when CoreNeuron runs it prints num_omp_thread=1 even if the OMP_NUM_THREADS is set to 2. Maybe that's because we don't run it with --threading? I need to investigate
@iomaganaris : IMO we need a test with more concurrency (more than ring). e.g. If we are using 36 cores on Jenkis CI, using 6 MPI ranks each with 6 threads. This can help to detect race condition bit. We can discuss this what should/can do.
I think we are currently we running pure MPI tests under CI. We should run MPI+OpenMP tests to catch any threading errors.
The text was updated successfully, but these errors were encountered: