Merge branch 'branch-24.04' into fix-cagra-prune-illegal-memory-access

rapidsai · Feb 28, 2024 · 192fdea · 192fdea
2 parents bd0eb24 + 6789367
commit 192fdea
Show file tree

Hide file tree

Showing 9 changed files with 74 additions and 31 deletions.
diff --git a/.devcontainer/cuda11.8-pip/devcontainer.json b/.devcontainer/cuda11.8-pip/devcontainer.json
@@ -5,16 +5,26 @@
     "args": {
       "CUDA": "11.8",
       "PYTHON_PACKAGE_MANAGER": "pip",
-      "BASE": "rapidsai/devcontainers:24.04-cpp-llvm16-cuda11.8-ubuntu22.04"
+      "BASE": "rapidsai/devcontainers:24.04-cpp-cuda11.8-ubuntu22.04"
     }
   },
   "hostRequirements": {"gpu": "optional"},
   "features": {
-    "ghcr.io/rapidsai/devcontainers/features/ucx:24.4": {"version": "1.14.1"},
+    "ghcr.io/rapidsai/devcontainers/features/ucx:24.4": {
+      "version": "1.14.1"
+    },
+    "ghcr.io/rapidsai/devcontainers/features/cuda:24.4": {
+      "version": "11.8",
+      "installcuBLAS": true,
+      "installcuSOLVER": true,
+      "installcuRAND": true,
+      "installcuSPARSE": true
+    },
     "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.4": {}
   },
   "overrideFeatureInstallOrder": [
     "ghcr.io/rapidsai/devcontainers/features/ucx",
+    "ghcr.io/rapidsai/devcontainers/features/cuda",
     "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
   ],
   "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config/pip,local/share/${localWorkspaceFolderBasename}-cuda11.8-venvs}"],

diff --git a/...ontainer/cuda12.0-conda/devcontainer.json → ...ontainer/cuda12.2-conda/devcontainer.json b/...ontainer/cuda12.0-conda/devcontainer.json → ...ontainer/cuda12.2-conda/devcontainer.json
@@ -3,7 +3,7 @@
     "context": "${localWorkspaceFolder}/.devcontainer",
     "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
     "args": {
-      "CUDA": "12.0",
+      "CUDA": "12.2",
       "PYTHON_PACKAGE_MANAGER": "conda",
       "BASE": "rapidsai/devcontainers:24.04-cpp-mambaforge-ubuntu22.04"
     }
@@ -15,7 +15,7 @@
   "overrideFeatureInstallOrder": [
     "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
   ],
-  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config,conda/pkgs,conda/${localWorkspaceFolderBasename}-cuda12.0-envs}"],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config,conda/pkgs,conda/${localWorkspaceFolderBasename}-cuda12.2-envs}"],
   "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
   "workspaceFolder": "/home/coder",
   "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/raft,type=bind,consistency=consistent",
@@ -24,7 +24,7 @@
     "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
     "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
     "source=${localWorkspaceFolder}/../.conda/pkgs,target=/home/coder/.conda/pkgs,type=bind,consistency=consistent",
-    "source=${localWorkspaceFolder}/../.conda/${localWorkspaceFolderBasename}-cuda12.0-envs,target=/home/coder/.conda/envs,type=bind,consistency=consistent"
+    "source=${localWorkspaceFolder}/../.conda/${localWorkspaceFolderBasename}-cuda12.2-envs,target=/home/coder/.conda/envs,type=bind,consistency=consistent"
   ],
   "customizations": {
     "vscode": {

diff --git a/.devcontainer/cuda12.0-pip/devcontainer.json → .devcontainer/cuda12.2-pip/devcontainer.json b/.devcontainer/cuda12.0-pip/devcontainer.json → .devcontainer/cuda12.2-pip/devcontainer.json
@@ -3,29 +3,39 @@
     "context": "${localWorkspaceFolder}/.devcontainer",
     "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
     "args": {
-      "CUDA": "12.0",
+      "CUDA": "12.2",
       "PYTHON_PACKAGE_MANAGER": "pip",
-      "BASE": "rapidsai/devcontainers:24.04-cpp-llvm16-cuda12.0-ubuntu22.04"
+      "BASE": "rapidsai/devcontainers:24.04-cpp-cuda12.2-ubuntu22.04"
     }
   },
   "hostRequirements": {"gpu": "optional"},
   "features": {
-    "ghcr.io/rapidsai/devcontainers/features/ucx:24.4": {"version": "1.14.1"},
+    "ghcr.io/rapidsai/devcontainers/features/ucx:24.4": {
+      "version": "1.14.1"
+    },
+    "ghcr.io/rapidsai/devcontainers/features/cuda:24.4": {
+      "version": "12.2",
+      "installcuBLAS": true,
+      "installcuSOLVER": true,
+      "installcuRAND": true,
+      "installcuSPARSE": true
+    },
     "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.4": {}
   },
   "overrideFeatureInstallOrder": [
     "ghcr.io/rapidsai/devcontainers/features/ucx",
+    "ghcr.io/rapidsai/devcontainers/features/cuda",
     "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
   ],
-  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config/pip,local/share/${localWorkspaceFolderBasename}-cuda12.0-venvs}"],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config/pip,local/share/${localWorkspaceFolderBasename}-cuda12.2-venvs}"],
   "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
   "workspaceFolder": "/home/coder",
   "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/raft,type=bind,consistency=consistent",
   "mounts": [
     "source=${localWorkspaceFolder}/../.aws,target=/home/coder/.aws,type=bind,consistency=consistent",
     "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
     "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
-    "source=${localWorkspaceFolder}/../.local/share/${localWorkspaceFolderBasename}-cuda12.0-venvs,target=/home/coder/.local/share/venvs,type=bind,consistency=consistent"
+    "source=${localWorkspaceFolder}/../.local/share/${localWorkspaceFolderBasename}-cuda12.2-venvs,target=/home/coder/.local/share/venvs,type=bind,consistency=consistent"
   ],
   "customizations": {
     "vscode": {

diff --git a/.github/workflows/pr.yaml b/.github/workflows/pr.yaml
@@ -104,8 +104,10 @@ jobs:
       script: ci/test_wheel_raft_dask.sh
   devcontainer:
     secrets: inherit
-    uses: rapidsai/shared-workflows/.github/workflows/build-in-devcontainer.yaml@branch-24.04
+    uses: rapidsai/shared-workflows/.github/workflows/build-in-devcontainer.yaml@fix/devcontainer-json-location
     with:
+      arch: '["amd64"]'
+      cuda: '["12.2"]'
       build_command: |
         sccache -z;
         build-all -DBUILD_PRIMS_BENCH=ON -DBUILD_ANN_BENCH=ON --verbose;

diff --git a/ci/run_ctests.sh b/ci/run_ctests.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+set -euo pipefail
+
+# Support customizing the ctests' install location
+cd "${INSTALL_PREFIX:-${CONDA_PREFIX:-/usr}}/bin/gtests/libraft/"
+
+ctest --output-on-failure "$@"
diff --git a/ci/run_pylibraft_pytests.sh b/ci/run_pylibraft_pytests.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+set -euo pipefail
+
+# Support invoking run_pylibraft_pytests.sh outside the script directory
+cd "$(dirname "$(realpath "${BASH_SOURCE[0]}")")"/../python/pylibraft/pylibraft
+
+pytest --cache-clear "$@" test
diff --git a/ci/run_raft_dask_pytests.sh b/ci/run_raft_dask_pytests.sh
@@ -0,0 +1,9 @@
+#!/bin/bash
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+set -euo pipefail
+
+# Support invoking run_raft_dask_pytests.sh outside the script directory
+cd "$(dirname "$(realpath "${BASH_SOURCE[0]}")")"/../python/raft-dask/raft_dask
+
+pytest --cache-clear "$@" test
diff --git a/ci/test_cpp.sh b/ci/test_cpp.sh
@@ -1,8 +1,11 @@
 #!/bin/bash
-# Copyright (c) 2022-2023, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 set -euo pipefail
 
+# Support invoking test_cpp.sh outside the script directory
+cd "$(dirname "$(realpath "${BASH_SOURCE[0]}")")"/../
+
 . /opt/conda/etc/profile.d/conda.sh
 
 rapids-logger "Generate C++ testing dependencies"
@@ -31,13 +34,9 @@ rapids-mamba-retry install \
 rapids-logger "Check GPU usage"
 nvidia-smi
 
-EXITCODE=0
-trap "EXITCODE=1" ERR
-set +e
-
+export GTEST_OUTPUT=xml:${RAPIDS_TESTS_DIR}/
 # Run libraft gtests from libraft-tests package
-cd "$CONDA_PREFIX"/bin/gtests/libraft
-ctest -j8 --output-on-failure
+./ci/run_ctests.sh -j8 && EXITCODE=$? || EXITCODE=$?;
 
 rapids-logger "Test script exiting with value: $EXITCODE"
 exit ${EXITCODE}
diff --git a/ci/test_python.sh b/ci/test_python.sh
@@ -1,8 +1,11 @@
 #!/bin/bash
-# Copyright (c) 2022-2023, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 set -euo pipefail
 
+# Support invoking test_python.sh outside the script directory
+cd "$(dirname "$(realpath "${BASH_SOURCE[0]}")")"/../
+
 . /opt/conda/etc/profile.d/conda.sh
 
 rapids-logger "Generate Python testing dependencies"
@@ -41,28 +44,20 @@ trap "EXITCODE=1" ERR
 set +e
 
 rapids-logger "pytest pylibraft"
-pushd python/pylibraft/pylibraft
-pytest \
-  --cache-clear \
+./ci/run_pylibraft_pytests.sh \
   --junitxml="${RAPIDS_TESTS_DIR}/junit-pylibraft.xml" \
   --cov-config=../.coveragerc \
   --cov=pylibraft \
   --cov-report=xml:"${RAPIDS_COVERAGE_DIR}/pylibraft-coverage.xml" \
-  --cov-report=term \
-  test
-popd
+  --cov-report=term
 
 rapids-logger "pytest raft-dask"
-pushd python/raft-dask/raft_dask
-pytest \
-  --cache-clear \
+./ci/run_raft_dask_pytests.sh \
   --junitxml="${RAPIDS_TESTS_DIR}/junit-raft-dask.xml" \
   --cov-config=../.coveragerc \
   --cov=raft_dask \
   --cov-report=xml:"${RAPIDS_COVERAGE_DIR}/raft-dask-coverage.xml" \
-  --cov-report=term \
-  test
-popd
+  --cov-report=term
 
 rapids-logger "Test script exiting with value: $EXITCODE"
 exit ${EXITCODE}