8000 [ATen][CUDA][CUB] Implement changes to CCCL (CUB/Thrust/LibCUDACXX) usage in ATen by Aidyn-A · Pull Request #153373 · pytorch/pytorch · GitHub
[go: up one dir, main page]
Skip to content

[ATen][CUDA][CUB] Implement changes to CCCL (CUB/Thrust/LibCUDACXX) usage in ATen #153373

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
wants to merge 9 commits into
base: main
Choose a base branch
from
Open

[ATen][CUDA][CUB] Implement changes to CCCL (CUB/Thrust/LibCUDACXX) usage in ATen #153373

wants to merge 9 commits into from

Conversation

Aidyn-A
Copy link
Collaborator
@Aidyn-A Aidyn-A commented May 12, 2025 •
edited by pytorch-bot bot
Loading

A major release of CCCL 3.0.0 will introduce some bc-breaking changes. Namely iterators like TransformInputIterator and ConstantInputIterator were moved from CUB to Thrust, some operators like Max and Sum were moved to LibCUDACXX.

For the more info on changes please visit: https://nvidia.github.io/cccl/cccl/3.0_migration_guide.html

This is a follow up to PR #147493. A description from the original PR:

Several cub iterators have been deprecated and removed in the latest CCCL (cub) development NVIDIA/cccl#3831. This PR replaced the u 8000 sage of those cub iterators with thrust iterators.

Some cub thread operators were also deprecated and removed in NVIDIA/cccl#3918. This PR replaced those operators with libcudacxx ops.

This might also affect ROCM usability a bit.

This patch is tested to work with CCCL commit at NVIDIA/cccl@82befb0

Tracking of CCCL/CUB deprecations in the most recent development NVIDIA/cccl#101

cc @ptrblck @msaroufim @eqy @jerryzh168 @manuelcandales @SherlockNoMad @angelayi @xwang233 @miscco

@Aidyn-A Aidyn-A requested review from eqy and syed-ahmed as code owners May 12, 2025 13:20
@pytorch-bot PyTorch Bot
Copy link
pytorch-bot bot commented May 12, 2025 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/153373

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure

As of commit 8be49e6 with merge base 032ef48 (image):

NEW FAILURE - The following job has failed:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@pytorch-bot pytorch-bot bot added the release notes: cuda release notes category label May 12, 2025
@Aidyn-A Aidyn-A added module: cuda Related to torch.cuda, and CUDA support in general topic: not user facing topic category module: core aten Related to change to the Core ATen opset labels May 12, 2025
@Aidyn-A Aidyn-A requested a review from ngimel May 12, 2025 13:22
miscco
miscco reviewed May 12, 2025
View reviewed changes
Copy link
@miscco miscco left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is technically correct.

However, with the exception of cuda::maximum and cuda::minimumall the types are available unconditionally.

So we can considerably simplify the changes

@xwang233 xwang233 mentioned this pull request May 12, 2025
Closed
@colesbury colesbury added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label May 13, 2025
@Aidyn-A Aidyn-A marked this pull request as draft May 14, 2025 09:12
@Aidyn-A Aidyn-A marked this pull request as ready for review May 15, 2025 12:42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@miscco miscco miscco left review comments

@eqy eqy Awaiting requested review from eqy eqy is a code owner

@syed-ahmed syed-ahmed Awaiting requested review from syed-ahmed syed-ahmed is a code owner

@ngimel ngimel Awaiting requested review from ngimel

Assignees
No one assigned
Labels
module: core aten Related to change to the Core ATen opset module: cuda Related to torch.cuda, and CUDA support in general open source release notes: cuda release notes category topic: not user facing topic category 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

Successfully merging this pull request may close these issues.
5 participants
@Aidyn-A @miscco @ngimel @colesbury @pytorchbot
0