8000 [RFC] dropping CUDA 11.8 support in CI/CD · Issue #147383 · pytorch/pytorch · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
8000

[RFC] dropping CUDA 11.8 support in CI/CD #147383

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
atalman opened this issue Feb 18, 2025 · 1 comment
Open

[RFC] dropping CUDA 11.8 support in CI/CD #147383

atalman opened this issue Feb 18, 2025 · 1 comment
Labels
module: build Build system issues module: cuda Related to torch.cuda, and CUDA support in general triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@atalman
Copy link
Contributor
atalman commented Feb 18, 2025 •
edited
Loading

Related to: #145544

Opening this RFC to discuss dropping of CUDA 11.8 possibility and timeline

For PyTorch Release 2.7 we are proceeding with following configuration:
CUDA 11.8, CUDNN 9.1.0.70 - Same as Previous Release 2.6. No changes to CUDA 11.8 - Legacy version
CUDA 12.6 CUDNN 9.x - Version Released to Pypi - Stable version
CUDA 12.8 CUDNN 9.x - New Experimental version

Proposal is to drop it for release 2.8. Hence dropping support of 11.8 in nightlies for Mar 2025-Jun 2025.

cc @malfet @seemethere @ptrblck @msaroufim @eqy @tinglvv @nWEIdia

Versions

2.7-2.8
@atalman atalman changed the title [RFC] CUDA 11.8 support [RFC] dropping CUDA 11.8 support Feb 18, 2025
@colesbury colesbury added module: build Build system issues module: cuda Related to torch.cuda, and CUDA support in general triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module labels Feb 18, 2025
@atalman atalman changed the title [RFC] dropping CUDA 11.8 support [RFC] dropping CUDA 11.8 support in CI/CD Mar 10, 2025
@zpcore
8000 Copy link
Member
zpcore commented Apr 8, 2025

Hi @miladm , @ysiraichi , CUDA 11.8 is already legacy status. I think we can drop 11.8 for PyTorch/XLA 2.7 release then.

@nWEIdia nWEIdia mentioned this issue Apr 17, 2025
Closed
@atalman atalman mentioned this issue May 6, 2025
Open
@atalman atalman mentioned this issue May 19, 2025
Open
pytorchmergebot pushed a commit that referenced this issue May 20, 2025
@nWEIdia @pytorchmergebot
[CI][CUDA] Move cu118 distributed pull jobs to cu126, move cu124-sm75…
8cabd23 
… to cu126-sm75 (#151594)

This PR moves distributed cuda CI job from cuda 11.8 to cuda 12.6.
In doing so, a few unit test failures were exposed, some if not all of which would take a while to root-cause and fix, so temporarily skip them after creating the issues.

#153479 test_nan_assert tricky behavior (e.g. skip_but_pass_in_sandcastle, ubuntu 20.04 does not work, ubuntu 22.04 works, Amazon Linux 2023 skip - what is Sandcastle OS?)
#153122 CUDA context related
#153517  NCCL regression, future NCCL may fix it

See: #147383

Pull Request resolved: #151594
Approved by: https://github.com/eqy, https://github.com/atalman, https://github.com/cyyever
pytorchmergebot pushed a commit that referenced this issue May 22, 2025
@nWEIdia @pytorchmergebot
[CI][CUDA][Distributed] Move cuda 11.8 distributed pull jobs to cuda …
7128b50 
…12.6 (#151594)

This PR moves distributed cuda CI job from cuda 11.8 to cuda 12.6.
In doing so, a few unit test failures were exposed, some if not all of which would take a while to root-cause and fix, so temporarily skip them after creating the issues.

#153479 test_nan_assert tricky behavior (e.g. skip_but_pass_in_sandcastle, ubuntu 20.04 does not work, ubuntu 22.04 works, Amazon Linux 2023 skip - what is Sandcastle OS?)
#153122 CUDA context related
#153517  NCCL regression, future NCCL may fix it
#154073 skip test_symmetric_memory for cuda 12.6 before it is fixed

See: #147383

Pull Request resolved: #151594
Approved by: https://github.com/eqy, https://github.com/atalman, https://github.com/cyyever, https://github.com/huydhn, https://github.com/kwen2501
@nWEIdia nWEIdia mentioned this issue May 22, 2025
Closed
This was referenced May 23, 2025
Open
Closed
pytorchmergebot pushed a commit that referenced this issue Jun 2, 2025
@nWEIdia @pytorchmergebot
[CI][CUDA] Migrate remaining cu118 jobs to cu128 (#154169)
48807d5 
Contributing to the fix of #147383   and #154119

Additional steps required: https://github.com/pytorch/pytorch/blob/3218b1b68427e21220342bc0317ac8b492bfae2f/.github/workflows/lint.yml cu118 needs to be updated.
Make install_cuda.sh accept both 12.8 and 12.8.* as CUDA_VERSION argument.

Pull Request resolved: #154169
Approved by: https://github.com/eqy, https://github.com/malfet, https://github.com/atalman, https://github.com/tinglvv
@simon-mo simon-mo mentioned this issue Jun 3, 2025
Open
1 task
@scotts scotts mentioned this issue Jun 3, 2025
Closed
qingyi-yan pushed a commit to qingyi-yan/pytorch that referenced this issue Jun 3, 2025
@nWEIdia
[CI][CUDA] Migrate remaining cu118 jobs to cu128 (pytorch#154169)
bedc936 
Contributing to the fix of pytorch#147383   and pytorch#154119

Additional steps required: https://github.com/pytorch/pytorch/blob/3218b1b68427e21220342bc0317ac8b492bfae2f/.github/workflows/lint.yml cu118 needs to be updated.
Make install_cuda.sh accept both 12.8 and 12.8.* as CUDA_VERSION argument.

Pull Request resolved: pytorch#154169
Approved by: https://github.com/eqy, https://github.com/malfet, https://github.com/atalman, https://github.com/tinglvv
iupaikov-amd pushed a commit to ROCm/pytorch that referenced this issue Jun 4, 2025
@nWEIdia @iupaikov-amd
[CI][CUDA] Migrate remaining cu118 jobs to cu128 (pytorch#154169)
3b70687 
Contributing to the fix of pytorch#147383   and pytorch#154119

Additional steps required: https://github.com/pytorch/pytorch/blob/3218b1b68427e21220342bc0317ac8b492bfae2f/.github/workflows/lint.yml cu118 needs to be updated.
Make install_cuda.sh accept both 12.8 and 12.8.* as CUDA_VERSION argument.

Pull Request resolved: pytorch#154169
Approved by: https://github.com/eqy, https://github.com/malfet, https://github.com/atalman, https://github.com/tinglvv
@atalman atalman mentioned this issue Jun 9, 2025
Merged
atalman added a commit to pytorch/test-infra that referenced this issue Jun 9, 2025
@atalman
[CD] Remove CUDA 11.8 builds (#6727)
88bce55 
Please see: pytorch/pytorch#147383
@atalman atalman mentioned this issue Jun 10, 2025
Closed
pytorchmergebot pushed a commit that referenced this issue Jun 10, 2025
@atalman @pytorchmergebot
[CI/CD] Remove CUDA 11.8 builds (#155509)
8153340 
This removes CUDA 11.8 from CI/CD
Please see: #147383

TODO: Will followup of cleaning CUDA 11.8 config from scripts

Pull Request resolved: #155509
Approved by: https://github.com/cyyever, https://github.com/huydhn, https://github.com/malfet
@atalman atalman mentioned this issue Jun 10, 2025
Closed
pytorchmergebot pushed a commit that referenced this issue Jun 10, 2025
@atalman @pytorchmergebot
[BE] Remove CUDA 11 artifacts. Fix Check Binary workflow (#155555)
7a03b0d 
Please see: #147383

1. Remove CUDA 11 build and test artifacts. One place CUDA 12.4
2. Fix Check Binary Workflow to use Stable Cuda version variable rather then hardcoded one

Pull Request resolved: #155555
Approved by: https://github.com/malfet, https://github.com/Skylion007
thatgeeman pushed a commit to thatgeeman/pytorch-docathon that referenced this issue Jun 15, 2025
@atalman @thatgeeman
[CI/CD] Remove CUDA 11.8 builds (pytorch#155509)
978923d 
This removes CUDA 11.8 from CI/CD
Please see: pytorch#147383

TODO: Will followup of cleaning CUDA 11.8 config from scripts

Pull Request resolved: pytorch#155509
Approved by: https://github.com/cyyever, https://github.com/huydhn, https://github.com/malfet
thatgeeman pushed a commit to thatgeeman/pytorch-docathon that referenced this issue Jun 15, 2025
@atalman @thatgeeman
[BE] Remove CUDA 11 artifacts. Fix Check Binary workflow (pytorch#155555
c232c1f 
)

Please see: pytorch#147383

1. Remove CUDA 11 build and test artifacts. One place CUDA 12.4
2. Fix Check Binary Workflow to use Stable Cuda version variable rather then hardcoded one

Pull Request resolved: pytorch#155555
Approved by: https://github.com/malfet, https://github.com/Skylion007
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
module: build Build system issues module: cuda Related to torch.cuda, and CUDA support in general triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@colesbury @atalman @zpcore
0