Added support for Windows thread groups (to use all cores on a 3990X and similar) #188
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
cwoffenden
changed the title
|
This is great, thank you. I'll merge this PR in next. |
|
I'll merge this |
|
Just a FYI - v1.15 is out - so I have time to merge this. |
cwoffenden
force-pushed
the
cpu-affinity
branch
from
2c130b2 to
11d076e
Compare
|
(Rebased onto latest since this was quite out of date) |
|
Thanks- I'll be merging this change into the next milestone's OpenCL release, due end of Dec/early Jan. |
|
NP - I was bringing it up-to-date (partially) for my own testing. (I can make Threadripper available over VPN/RDP for you to test, BTW) |
|
Just a FYI I'm merging this change into v1.16 (the January 2022 release). |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
On Windows with more than 64 cores (and possibly with multiple CPUs with fewer cores, although not tested) only the first 64 is supported by the
job_pool. This addition spreads the threads across all CPUs. A picture speaks more, so here's the before, compiling as-is:And here's after with this PR:
When compressing to UASTC using:
This reduces the best time (from four runs) from 2.823 seconds to 1.766 seconds.
(If I were to make any changes, I'd spend more time making sure the main thread is in the smaller group, but I thought I'd publish this earlier to gather any interest.)