CI: Use nproc to count only on-line CPUs for GNUMAKEFLAGS
#14345
+2
−2
Conversation
This file contains hidden or 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
shyouhei
approved these changes
Aug 27, 2025
There was a problem hiding this comment.
At that time I added this code, hyper threading was still a thing. Under such circumstance nproc --all boosted overall CI speed. But that was an old story. Now that datacenter CPUs employ completely different setups. Removing --all just makes sense today. LGTM.
ca37b6a to
1a9b08b
Compare
Use `nproc` rather than `nproc --all`. Because the number by the `nproc` is different from the number by the `nproc --all` on GitHub Actions ppc64le/s390x. This caused the `make` command runs much more jobs than the number of on-line CPUs on these environments. The make command ran 193 jobs in parallel for 4 on-line CPUs in GitHub Actions ppc64le, and ran 9 jobs in parallel for 4 on-line CPUs in GitHub Actions s390x. And this caused the high load average 34.58 on ppc64le, and 6.69 on s390x. These values should be less than 4.0. I believe we should use the `nproc` rather than `nproc --all`. ``` + nproc 4 + nproc --all 192 ``` ``` + nproc 4 + nproc --all 8 ``` See IBM/actionspz#38 for details. Note the `--all` option in the `nproc --all` was originally added to boost CPU in a hyper threading environment where one physical core works like two logical cores. https://www.intel.com/content/www/us/en/gaming/resources/hyper-threading.html
1a9b08b to
1ad67e2
Compare
|
Thanks for your review and info! I updated the commit message adding the context about the hyper threading, and rebasing this PR on the latest master branch. |
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Outdated suggestions cannot be applied.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.
Use
nprocrather thannproc --all. Because the number by thenprocis different from the number by thenproc --allon GitHub Actions ppc64le/s390x. This caused themakecommand runs much more jobs than the number of on-line CPUs. The make command ran 193 jobs in parallel for 4 on-line CPUs in GitHub Actions ppc64le, and ran 9 jobs in parallel for 4 on-line CPUs in GitHub Actions s390x.And this caused the high load average 34.58 on ppc64le, and 6.69 on s390x. These values should be less than 4.0. I believe we should use the
nprocrather thannproc --all.See IBM/actionspz#38 for details.