This is an archive of the discontinued LLVM Phabricator instance.

[clang][ci] Improves buildkite artifacts.
ClosedPublic

Authored by Mordante on May 7 2023, 5:51 AM.

Details

Reviewers
EricWF
ldionne
Group Reviewers
Restricted Project
Commits
rGf271df0e7903: [clang][ci] Improves buildkite artifacts.
Summary

The financial cost of the network I/O for the Clang install artifacts is
quite significant. afd3478f37c8 improved this by creating tarballs. This
commit improves the tarball by using xz compression instead of gzip. This
option is the slowest, but gives the smallest size.

size  time           time
      (compression)  (decompression)

gzip 51 M 7 s 1.2 s
bz2 44 M 17 s 5.8 s
xz 33 M 76 s 3.1 s

Diff Detail

Event Timeline

Mordante created this revision.May 7 2023, 5:51 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 7 2023, 5:51 AM
Herald added a subscriber: arichardson. · View Herald Transcript
Mordante requested review of this revision.May 7 2023, 5:51 AM
Herald added a project: Restricted Project. · View Herald TranscriptMay 7 2023, 5:51 AM
Herald added a reviewer: Restricted Project. · View Herald Transcript
Mordante updated this revision to Diff 520183.May 7 2023, 7:13 AM

CI fixes.

Mordante added inline comments.May 7 2023, 8:37 AM
libcxx/utils/ci/buildkite-pipeline-clang.yml
28

I used this build https://buildkite.com/llvm-project/libcxx-ci/builds/23579 to test the changes with Clang.
This is build https://buildkite.com/llvm-project/libcxx-ci/builds/23567 of Clang before the changes.

Unsurprisingly, here the timings are closer than locally. I haven't checked what the noise on the timings is for Clang builds.

Are you sure that it's worth it? Adding a minute of compression time for a decrease of 20MB seems like it's the wrong call to me. Without this change we already shaved off ~4Gigs/clang pipeline.

Are you sure that it's worth it? Adding a minute of compression time for a decrease of 20MB seems like it's the wrong call to me. Without this change we already shaved off ~4Gigs/clang pipeline.

That was on my system. Looking at the CI the difference is much smaller. It seems our CI has a beefier machine than I do ;-)

ldionne accepted this revision.May 9 2023, 9:12 AM

The difference is like 15 seconds (and that might be due to normal variance in the CI), but we're reducing the size by 40% so IMO this is worth it. Not like a *must*, but interesting savings nonetheless. Let's do it!

This revision is now accepted and ready to land.May 9 2023, 9:12 AM
This revision was automatically updated to reflect the committed changes.