[CI] support multiple CUDA versions #1234
Closed
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
Summary: This patch adds multiple CUDA versions for OSS CI (build only). Previously on the OSS side, we tested only 11.3, but this patch expands it to 11.3, 11.5, and 11.6. This CI is useful. For example, it revealed an issue in `cub_namespace_postfix.cuh`: this change was cherry-picked for CUB 1.13.1 (https://github.com/NVIDIA/cub/blob/1.13.X/cub/util_namespace.cuh), so the `CUB_VERSION` condition should be `CUB_VERSION >= 101301`. Q&A - Why don't we test multiple ROCm versions? - PyTorch provides two versions for ROCm at this point: 5.0 and 5.1.1, but FBGEMM does not support ROCm 5.0. - Why don't we test CUDA 11.2 or older? - I am not sure if FBGEMM supports CUDA <= 11.2. If needed, we can add it. - Why don't we test CUDA 11.4? - PyTorch does not have nightly CUDA 11.4 build for some reasons at present. - Why don't we test CUDA 11.7? - It has some compilation issues. Differential Revision: D38359867 fbshipit-source-id: faf6d8321269da72b92cf2af3ee1f940531ff8b8
✅ Deploy Preview for eclectic-stroopwafel-199537 canceled.
|
|
This pull request was exported from Phabricator. Differential Revision: D38359867 |
shintaro-iwasaki
changed the title
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Outdated suggestions cannot be applied.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.
Summary:
This patch adds multiple CUDA versions for OSS CI (build only). Previously on the OSS side, we tested only 11.3, but this patch expands it to 11.3, 11.5, and 11.6.
This CI is useful. For example, it revealed an issue in
cub_namespace_postfix.cuh: this change was cherry-picked for CUB 1.13.1 (https://github.com/NVIDIA/cub/blob/1.13.X/cub/util_namespace.cuh), so theCUB_VERSIONcondition should beCUB_VERSION >= 101301.Q&A
Differential Revision: D38359867