From de05abd8fae510e5b96d634523c2f11efecbeb4a Mon Sep 17 00:00:00 2001 From: James Lamb Date: Mon, 9 Dec 2024 11:04:50 -0600 Subject: [PATCH 1/2] PR CI: make cugraph builds depend on pylibcugraph builds (#4801) Proposes the following change to the wheel jobs for PR CI: ```mermaid --- title: Current --- flowchart LR A[build-pylibcugraph] --> B[test-pylibcugraph] B --> C[build-cugraph] C --> D[test-cugraph] ``` ```mermaid --- title: Proposed --- flowchart LR A[build-pylibcugraph] --> B[test-pylibcugraph] A --> C[build-cugraph] C --> D[test-cugraph] ``` ## Notes for Reviewers I think reducing the end-to-end time for changes here is even more important now that we have new downstream repos (https://github.com/rapidsai/nx-cugraph and https://github.com/rapidsai/cugraph-gnn) that depend on changes made here. ### Benefits of this change * shorter end-to-end time for CI runs - *by parallelizing more work* - *by doing `cugraph-cu{11,12}` builds more frequently, which should mean more frequent population of the `sccache` cache* * faster feedback about `cugraph-cu{11,12}` build issues * consistency with the rest of RAPIDS (every other RAPIDS project has builds depend on builds, not tests, as far as I know) ### Costs of this change * more at-one-moment load on GPU CI runners as a result of `cugraph` PRs (wheel tests for `cugraph` could now occupy 4 GPU runners at once instead of just 2) - *`pylibcugraph` test jobs tend to take around 6-10 minutes once scheduled onto a runner, so this shouldn't be too bad* Authors: - James Lamb (https://github.com/jameslamb) Approvers: - Chuck Hastings (https://github.com/ChuckHastings) - Bradley Dice (https://github.com/bdice) - Alex Barghi (https://github.com/alexbarghi-nv) URL: https://github.com/rapidsai/cugraph/pull/4801 --- .github/workflows/pr.yaml | 2 +- python/cugraph/pytest.ini | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/.github/workflows/pr.yaml b/.github/workflows/pr.yaml index 633d964a3b..c9e41475a1 100644 --- a/.github/workflows/pr.yaml +++ b/.github/workflows/pr.yaml @@ -152,7 +152,7 @@ jobs: build_type: pull-request script: ci/test_wheel_pylibcugraph.sh wheel-build-cugraph: - needs: wheel-tests-pylibcugraph + needs: wheel-build-pylibcugraph secrets: inherit uses: rapidsai/shared-workflows/.github/workflows/wheels-build.yaml@branch-25.02 with: diff --git a/python/cugraph/pytest.ini b/python/cugraph/pytest.ini index bf6e6bdd80..7bdef2d977 100644 --- a/python/cugraph/pytest.ini +++ b/python/cugraph/pytest.ini @@ -62,6 +62,7 @@ filterwarnings = error::FutureWarning error::DeprecationWarning # TODO + ignore:.*cuda..* module is deprecated.*:DeprecationWarning ignore:Multi is deprecated and the removal of multi edges will no longer be supported:FutureWarning ignore:The legacy column names:FutureWarning ignore:The include_hop_column flag is deprecated and will be removed:FutureWarning From b0dc702b15e362ce68d845db15a1247b338e7fae Mon Sep 17 00:00:00 2001 From: Bradley Dice Date: Mon, 9 Dec 2024 16:05:48 -0600 Subject: [PATCH 2/2] Update cuda-python lower bounds to 12.6.2 / 11.8.5 (#4813) We require a newer cuda-python lower bound for new features and to use the new layout. This will fix a number of errors observed when the runtime version of cuda-python is older than the version used to build packages using Cython features from cuda-python. See https://github.com/rapidsai/build-planning/issues/117#issuecomment-2524250915 for details. Authors: - Bradley Dice (https://github.com/bdice) - Ralph Liu (https://github.com/nv-rliu) Approvers: - James Lamb (https://github.com/jameslamb) URL: https://github.com/rapidsai/cugraph/pull/4813 --- conda/recipes/cugraph/meta.yaml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/conda/recipes/cugraph/meta.yaml b/conda/recipes/cugraph/meta.yaml index 9f5a137fab..094b30c5fc 100644 --- a/conda/recipes/cugraph/meta.yaml +++ b/conda/recipes/cugraph/meta.yaml @@ -75,10 +75,10 @@ requirements: - {{ pin_compatible('cuda-version', max_pin='x', min_pin='x') }} {% if cuda_major == "11" %} - cudatoolkit - - cuda-python >=11.7.1,<12.0a0 + - cuda-python >=11.8.5,<12.0a0 {% else %} - cuda-cudart - - cuda-python >=12.0,<13.0a0 + - cuda-python >=12.6.2,<13.0a0 {% endif %} - cudf ={{ minor_version }} - cupy >=12.0.0