Skip to content

update gputreeshap#7121

Merged
rapids-bot[bot] merged 5 commits intorapidsai:branch-25.10from
jameslamb:update-gputreeshap
Aug 22, 2025
Merged

update gputreeshap#7121
rapids-bot[bot] merged 5 commits intorapidsai:branch-25.10from
jameslamb:update-gputreeshap

Conversation

@jameslamb
Copy link
Member

@jameslamb jameslamb commented Aug 20, 2025

Contributes to rapidsai/build-planning#208

As part of adding CUDA 13 support across RAPIDS, in rapidsai/gputreeshap#56 I'm proposing testing gputreeshap with CUDA 13.

This PR updates the gputreeshap commit here to include those changes.

@jameslamb jameslamb added DO NOT MERGE Hold off on merging; see PR for details improvement Improvement / enhancement to an existing function non-breaking Non-breaking change labels Aug 20, 2025
@copy-pr-bot

This comment was marked as resolved.

@jameslamb
Copy link
Member Author

/ok to test

@jameslamb
Copy link
Member Author

/ok to test

@jameslamb
Copy link
Member Author

CI here is failing with an issue affecting all of cuml CI:

error    libmamba Could not solve for environment specs
    The following packages are incompatible
    ├─ cuda-version =12.2 * is requested and can be installed;
    └─ libcuml =25.10,>=0.0.0a0 * is not installable because it requires
       └─ libcuvs =25.10 *, which requires
          └─ libraft-headers =25.10 *, which requires
             ├─ libcusparse-dev =* * with the potential options
             │  ├─ libcusparse-dev 12.1.2.141 would require
             │  │  └─ libnvjitlink >=12.2.140,<12.3.0a0 *, which can be installed;
             │  ├─ libcusparse-dev 12.0.0.76 would require
             │  │  └─ cuda-version [>=12.0,<12.1.0a0 *|>=12.0.0,<12.1.0a0 *], which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev 12.1.0.106 would require
             │  │  └─ cuda-version >=12.1,<12.2.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev 12.2.0.103 would require
             │  │  └─ cuda-version >=12.3,<12.4.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev [12.3.0.142|12.3.1.170] would require
             │  │  └─ cuda-version >=12.4,<12.5.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev [12.4.1.24|12.5.1.3] would require
             │  │  └─ cuda-version >=12.5,<12.6.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev [12.5.10.65|12.5.9.5] would require
             │  │  └─ cuda-version >=12.9,<12.10.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev [12.5.2.23|12.5.3.3|12.5.4.2] would require
             │  │  └─ cuda-version >=12.6,<12.7.0a0 *, which conflicts with any installable versions previously reported;
             │  ├─ libcusparse-dev [12.5.7.53|12.5.8.93] would require
             │  │  └─ cuda-version >=12.8,<12.9.0a0 *, which conflicts with any installable versions previously reported;
             │  └─ libcusparse-dev 12.6.2.49 would require
             │     └─ cuda-version >=13.0,<13.1.0a0 *, which conflicts with any installable versions previously reported;
             └─ ucxx =0.46 * but there are no viable options
                ├─ ucxx [0.46.00a21|0.46.00a22|0.46.00a23] would require
                │  └─ numba-cuda >=0.18.0,<0.19.0a0 * but there are no viable options
                │     ├─ numba-cuda [0.18.0|0.18.1] would require
                │     │  └─ cuda-bindings =* * but there are no viable options
                │     │     ├─ cuda-bindings [12.9.1|12.9.2] would require
                │     │     │  └─ libnvjitlink >=12.3,<13 * but there are no viable options
                │     │     │     ├─ libnvjitlink [12.3.101|12.4.127|...|12.9.86] conflicts with any installable versions previously reported;
                │     │     │     ├─ libnvjitlink 12.4.99 would require
                │     │     │     │  └─ cuda-version >=12.4,<12.5.0a0 *, which conflicts with any installable versions previously reported;
                │     │     │     └─ libnvjitlink [12.6.20|12.6.68] would require
                │     │     │        └─ cuda-version >=12.6,<12.7.0a0 *, which conflicts with any installable versions previously reported;
                │     │     ├─ cuda-bindings [13.0.0|13.0.1] would require
                │     │     │  └─ cuda-version >=13,<14.0a0 *, which conflicts with any installable versions previously reported;
                │     │     ├─ cuda-bindings [11.8.6|11.8.7] would require
                │     │     │  └─ cuda-version >=11,<12.0a0 *, which conflicts with any installable versions previously reported;
                │     │     └─ cuda-bindings [12.8.0|12.9.0] would require
                │     │        └─ libnvjitlink >=12.3 * but there are no viable options
                │     │           ├─ libnvjitlink [12.3.101|12.4.127|...|12.9.86] conflicts with any installable versions previously reported;
                │     │           ├─ libnvjitlink 12.4.99, which cannot be installed (as previously explained);
                │     │           ├─ libnvjitlink [12.6.20|12.6.68], which cannot be installed (as previously explained);
                │     │           └─ libnvjitlink 13.0.39 would require
                │     │              └─ cuda-version >=13,<13.1.0a0 *, which conflicts with any installable versions previously reported;
                │     └─ numba-cuda [0.18.0|0.18.1] would require
                │        └─ cuda-version >=11.2,<12 *, which conflicts with any installable versions previously reported;
                └─ ucxx 0.46.00a24 would require
                   └─ numba-cuda >=0.19.0,<0.20.0a0 *, which requires
                      └─ cuda-bindings >=12.9.1 *, which cannot be installed (as previously explained).
critical libmamba Could not solve for environment specs

(conda-cpp-tests link)

Which should be fixed by conda-forge/conda-forge-repodata-patches-feedstock#1065

@jameslamb
Copy link
Member Author

/ok to test

@jameslamb
Copy link
Member Author

🎉 this is working

build link: https://github.com/rapidsai/cuml/actions/runs/17135809649?pr=7121

Proposed steps:

  1. merge test CUDA 13 support, other small updates gputreeshap#56
  2. update this branch to point at the rapidsai/gputreeshap repo, not my fork
  3. merge this PR

@jameslamb jameslamb changed the title WIP: update gputreeshap update gputreeshap Aug 21, 2025
@jameslamb jameslamb marked this pull request as ready for review August 21, 2025 21:06
@jameslamb jameslamb requested a review from a team as a code owner August 21, 2025 21:06
@jameslamb jameslamb requested review from cjnolet and csadorf August 21, 2025 21:06
@jameslamb jameslamb removed the DO NOT MERGE Hold off on merging; see PR for details label Aug 21, 2025
@jakirkham
Copy link
Member

cc @dantegd

@dantegd
Copy link
Member

dantegd commented Aug 22, 2025

/merge

@rapids-bot rapids-bot bot merged commit 00c3be7 into rapidsai:branch-25.10 Aug 22, 2025
89 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CMake CUDA/C++ improvement Improvement / enhancement to an existing function non-breaking Non-breaking change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants