-
Notifications
You must be signed in to change notification settings - Fork 3
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
binding may be floating to all places #6
Comments
Test on the Thomas' desktop: 6 cores, 12 HT. I will include the output of the test on my desktop asap, I believe I do not have that problem but there is clearly a bug somewhere. |
The test is running only 2 MPI ranks but the work I did was assuming 1 MPI rank per core and then spread OpenMP threads on the HT of that core. It is still a use case we want to support but it is not yet working. |
Quick follow-up test on BEAKER desktop (lstopo shows: 6 cores, each with 2 hwthreads).
|
The |
Adding this as issue to track status while testing. Using
gv_pmix
branch atda5ea2c
(after PR #4 was merged).I enabled debug prints/instrumentation to
__kmp_affinity_set_init_mask()
to show the binding of OMP threads to places. I see that some threads showall places
instead of a specific placeX
.Here's the command I used on machine with 12 cores (libomp/moc test show NCPUs=12).
NOTE: I lowered the
KA_TRACE()
lines to debug level=5 in code, otherwise useKMP_A_DEBUG=101
.Here's the relevant output sorted for easier reading and with the PID info:
The text was updated successfully, but these errors were encountered: