Use ninja for build - Githubissues

conda-forge / scipoptsuite-feedstock

A conda-smithy repository for scipoptsuite.

BSD 3-Clause "New" or "Revised" License

25 stars 13 forks source link

Use ninja for build #55

Closed dg0yt closed 11 months ago

dg0yt commented 11 months ago

Checklist

[x] Used a personal fork of the feedstock to propose changes
[ ] Bumped the build number (if the version is unchanged)
[ ] Reset the build number to 0 (if the version changed)
[ ] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
[ ] Ensured the license file is being packaged.

conda-forge-webservices[bot] commented 11 months ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe:

It looks like the 'papilo' output doesn't have any tests.
License is not an SPDX identifier (or a custom LicenseRef) nor an SPDX license expression.

Documentation on acceptable licenses can be found here.

dg0yt commented 11 months ago

CI build time changes:

Target	Before this PR	`ninja`, this PR	Remark
linux	30 min	42 min	12 min worse - unexpected!
win	55 min	43 min	12 min improvement
osx x64	54 min	45 min	9 min improvement
osx arm64	50 min	42 min	8 min improvement

NB: Build times in CI vary widely, but I only took single data points.

linux-64

make before this PR:

Resource usage statistics from building scipoptsuite:
   Process count: 16
   CPU time: Sys=0:00:57.0, User=0:27:37.2
   Memory: 5.0G
   Disk usage: 381.1K
   Time elapsed: 0:20:18.6

ninja with this PR:

Resource usage statistics from building scipoptsuite:
   Process count: 11
   CPU time: Sys=0:01:32.3, User=0:39:43.2
   Memory: 5.0G
   Disk usage: 318.0K
   Time elapsed: 0:27:36.9

I did expect at least a small improvement, similar to osx. Something strange going on here. Or simply a machine with less cores (cf. Process count).

dg0yt commented 11 months ago

ninja on another CI run:

Resource usage statistics from building scipoptsuite:
   Process count: 11
   CPU time: Sys=0:00:56.0, User=0:27:18.5
   Memory: 5.0G
   Disk usage: 154.6M
   Time elapsed: 0:19:43.0

So everything is okay, only CI machines vary.

h-vetinari commented 11 months ago

So everything is okay, only CI machines vary.

Yeah, the machines we get randomly assigned in Azure Pipelines can be quite different in capabilities. Depending on the recipe, a factor 2x can happen between the fastest and slowest build times.