CI Remove conda environment cache in CUDA CI #31900
Merged
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
|
Let's remove it. The conda cache was added speculatively. The assumption was that setting up conda envs is mostly limited by network speed and that the network speed to the cache might be slightly faster than to the conda package server. Seems like this is not the case and not using the cache makes the CI simpler. So I'd remove it |
betatim
approved these changes
Aug 11, 2025
ogrisel
approved these changes
Aug 11, 2025
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Outdated suggestions cannot be applied.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.
Installing and creating the environment from scratch takes 1 min 30s:
build log
Reusing the cache takes ~2 minutes so it's actually slower:
build log
It seems like this pattern is consistent e.g. build log without cache and build log with cache.
I noticed because the cache is quite big ~6.4GB (out of a 10GB quota in principle), see all our caches.
So basically if you run the CUDA CI in 2 different PRs you will create 2 cache entries (different branches) and you are above the quota limit. I am not sure how strongly this quota is enforced to be honest but I noticed there was a warning the other day that we were above our limit.
cc @betatim in case you remember some details about using caching for the conda environment in #29130.