Skip to content

Upgrade to CUDA 12.9#11972

Merged
hcho3 merged 6 commits into
dmlc:masterfrom
hcho3:test_cuda129
Jan 30, 2026
Merged

Upgrade to CUDA 12.9#11972
hcho3 merged 6 commits into
dmlc:masterfrom
hcho3:test_cuda129

Conversation

@hcho3

@hcho3 hcho3 commented Jan 29, 2026

Copy link
Copy Markdown
Collaborator

Test CTK 12.9.

Note. The Windows CI pipeline is already using CTK 12.9.

Closes #11899

@hcho3
hcho3 marked this pull request as draft January 29, 2026 05:46
@hcho3 hcho3 changed the title [Don't merge] Test dmlc/xgboost-devops#77 [WIP] Upgrade to CUDA 12.9 Jan 29, 2026
Comment thread src/common/algorithm.cuh
void InclusiveSum(Context const *ctx, InputIteratorT d_in, OutputIteratorT d_out,
OffsetT num_items) {
#if CUB_VERSION >= 300000
#if CUB_VERSION >= 200800

@hcho3 hcho3 Jan 29, 2026

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

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

cub::Sum is marked as deprecated starting from CCCL 2.8.0 (which is used by CTK 12.9).

@hcho3 hcho3 changed the title [WIP] Upgrade to CUDA 12.9 Upgrade to CUDA 12.9 Jan 30, 2026
@hcho3
hcho3 marked this pull request as ready for review January 30, 2026 20:35
@hcho3
hcho3 requested a review from trivialfis January 30, 2026 20:36

@trivialfis trivialfis left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Overall looks good.

  • Please update the CMakeLists.txt to reflect the required CTK version bump. #11968
  • Let's ignore the clang-format error for this PR. I would like to get #11980 first, if that's preferred.

@hcho3
hcho3 merged commit 73d7637 into dmlc:master Jan 30, 2026
75 of 76 checks passed
@hcho3
hcho3 deleted the test_cuda129 branch January 30, 2026 22:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[CI] Bump CUDA requirement to 12.9

2 participants