User Details
- User Since
- Apr 20 2015, 1:32 PM (311 w, 5 d)
Mar 5 2021
Hi Johnny,
When I just commit via the instructions on our wiki, I just see this separate e-mail labeled "[Diffusion]" which looks like a separate review in Phabricator. When others push their changes, it appears attached to the original phabricator review, and seems to close it automagically. What am I missing there?
Feb 26 2021
Added masked to cons_text_c.
Feb 18 2021
Jan 29 2021
Jan 28 2021
Addressed Nawrin's comment fixes.
Jan 27 2021
LGTM
Jan 26 2021
LGTM
LGTM
Jan 22 2021
Jan 7 2021
Dec 17 2020
Dec 9 2020
Dec 1 2020
One missed change added.
Nov 30 2020
Build problems associated with target attribute for waitpkg should be fixed now. Please try the new diff.
Fixed build problems associated with target attribute for waitpkg.
Nov 16 2020
LGTM
Nov 10 2020
Oct 23 2019
LGTM
Aug 5 2019
Jul 22 2019
Jul 12 2019
Jul 10 2019
Jun 27 2019
Jun 18 2019
LGTM
Jun 17 2019
May 15 2019
LGTM!
Apr 23 2019
LGTM
Apr 8 2019
LGTM
*ping*
Apr 2 2019
LGTM
Mar 18 2019
Mar 15 2019
*ping*
Mar 5 2019
Feb 27 2019
*ping*
*ping*
Feb 26 2019
Feb 19 2019
Feb 13 2019
LGTM
Feb 12 2019
Feb 11 2019
Fix update for non-CG_root threads in __kmp_free_thread.
Feb 8 2019
I was going for the per-data-environment specification for the ICV, though I agree it is only needed per contention group. Is there any consensus on this? Leave as is, or remove per-data-environment ICV?
Jan 16 2019
Sorry Andrey -- I missed the Windows fix. Added now.
Jan 15 2019
Renamed parameters to match 5.0 spec. Added test.
Jan 2 2019
Dec 21 2018
LGTM
Dec 18 2018
Tests added.
Dec 12 2018
Dec 11 2018
Nov 29 2018
Nov 28 2018
Nov 27 2018
Nov 9 2018
Nov 6 2018
LGTM
Oct 23 2018
Aug 6 2018
Jul 26 2018
LGTM
LGTM
LGTM
LGTM
Apr 11 2018
LGTM
Mar 26 2018
LGTM
Mar 22 2018
LGTM
Mar 19 2018
LGTM
Mar 16 2018
We have a patch coming with libomp-side addition of OMP_TARGET_OFFLOAD envirable plus an internal API so that libomptarget can get the value from libomp. Would it help if we get this patch out faster?
Feb 14 2018
Updated comment to be more general.
Jan 10 2018
Jan 9 2018
Jan 8 2018
Dec 4 2017
Nov 29 2017
LGTM
Nov 27 2017
LGTM
Nov 22 2017
Aug 29 2017
Aug 10 2017
So I think in summary, what you are doing is creating the place partitions with the ceil(P/T)-sized partitions evenly spread amongst the floor(P/T) partitions (where P is num places and T is num_threads). Is that right?
Aug 9 2017
Hi Paul,
Can you be more specific about how this "improves the way threads are spread across cores"? The OMP spec is very specific on exactly how the spread algorithm should work.
Thanks!
Terry
Jul 28 2017
LGTM
Jul 27 2017
Added a test for OMP_THREAD_LIMIT env, api, and general functionality. Note that this test, when running correctly, will emit a Warning and a Hint, because the limit squashes the requested number of threads.