Update to RAPIDS 24.10 (#1874)

Closes #1717 Requires nv-morpheus/utilities#75 Requires nv-morpheus/MRC#494 Issues opened during the course of this upgrade: - #1925 ## By Submitting this PR I confirm: - I am familiar with the [Contributing Guidelines](https://github.com/nv-morpheus/Morpheus/blob/main/docs/source/developer_guide/contributing.md). - When the PR is ready for review, new or existing tests cover these changes. - When the PR is ready for review, the documentation is up to date with these changes. Authors: - Christopher Harris (https://github.com/cwharris) - Michael Demoret (https://github.com/mdemoret-nv) Approvers: - Anuradha Karuppiah (https://github.com/AnuradhaKaruppiah) - Tad ZeMicheal (https://github.com/tzemicheal) URL: #1874
nv-morpheus · Oct 9, 2024 · 61cac96 · 61cac96
1 parent f43be04
commit 61cac96
Show file tree

Hide file tree

Showing 49 changed files with 237 additions and 200 deletions.
diff --git a/.devcontainer/cuda12.1-conda/devcontainer.json b/.devcontainer/cuda12.1-conda/devcontainer.json
@@ -5,7 +5,7 @@
     "args": {
       "CUDA": "12.1",
       "PYTHON_PACKAGE_MANAGER": "conda",
-      "BASE": "rapidsai/devcontainers:24.06-cpp-mambaforge-ubuntu22.04"
+      "BASE": "rapidsai/devcontainers:24.10-cpp-mambaforge-ubuntu22.04"
     }
   },
   "privileged": true,

diff --git a/.github/workflows/pr.yaml b/.github/workflows/pr.yaml
@@ -94,8 +94,8 @@ jobs:
       conda_run_build: ${{ !fromJSON(needs.prepare.outputs.is_pr) || fromJSON(needs.prepare.outputs.has_conda_build_label) }}
       # Upload morpheus conda packages only for non PR branches. Use 'main' for main branch and 'dev' for all other branches
       conda_upload_label: ${{ !fromJSON(needs.prepare.outputs.is_pr) && (fromJSON(needs.prepare.outputs.is_main_branch) && 'main' || 'dev') || '' }}
-      container: nvcr.io/ea-nvidia-morpheus/morpheus:morpheus-ci-build-240614
-      test_container: nvcr.io/ea-nvidia-morpheus/morpheus:morpheus-ci-test-240614
+      container: nvcr.io/ea-nvidia-morpheus/morpheus:morpheus-ci-build-241004
+      test_container: nvcr.io/ea-nvidia-morpheus/morpheus:morpheus-ci-test-241004
     secrets:
       CONDA_TOKEN: ${{ secrets.CONDA_TOKEN }}
       NGC_API_KEY: ${{ secrets.NGC_API_KEY }}
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -39,7 +39,7 @@ option(MORPHEUS_USE_IWYU "Enable running include-what-you-use as part of the bui
 
 set(MORPHEUS_PY_INSTALL_DIR "${CMAKE_CURRENT_BINARY_DIR}/wheel" CACHE STRING "Location to install the python directory")
 
-set(MORPHEUS_RAPIDS_VERSION "24.02" CACHE STRING "Sets default versions for RAPIDS libraries.")
+set(MORPHEUS_RAPIDS_VERSION "24.10" CACHE STRING "Sets default versions for RAPIDS libraries.")
 set(MORPHEUS_CACHE_DIR "${CMAKE_SOURCE_DIR}/.cache" CACHE PATH "Directory to contain all CPM and CCache data")
 mark_as_advanced(MORPHEUS_CACHE_DIR)
 

diff --git a/ci/conda/recipes/morpheus-libs/conda_build_config.yaml b/ci/conda/recipes/morpheus-libs/conda_build_config.yaml
@@ -14,19 +14,19 @@
 # limitations under the License.
 
 c_compiler_version:
-  - 11.2
+  - 12.1
 
 cxx_compiler_version:
-  - 11.2
+  - 12.1
 
 cuda_compiler:
   - cuda-nvcc
 
 cuda_compiler_version:
-  - 12.1
+  - 12.5
 
 python:
   - 3.10
 
 rapids_version:
-  - 24.02
+  - 24.10
diff --git a/ci/conda/recipes/morpheus-libs/meta.yaml b/ci/conda/recipes/morpheus-libs/meta.yaml
@@ -53,7 +53,7 @@ outputs:
         - cuda-version {{ cuda_compiler_version }}.*
         - ninja =1.11
         - pkg-config =0.29
-        - sysroot_linux-64 =2.17
+        - sysroot_linux-64 =2.28
       host:
         # CUDA dependencies
         - cuda-cudart-dev {{ cuda_compiler_version }}.*
@@ -65,7 +65,7 @@ outputs:
         # Non-CUDA dependencies
         - cudf {{ rapids_version }}
         - cython 3.0.*
-        - glog 0.6.*
+        - glog >=0.7.1,<0.8
         - libcudf {{ rapids_version }}
         - librdkafka >=1.9.2,<1.10.0a0
         - mrc {{ minor_version }}
@@ -76,7 +76,7 @@ outputs:
         - rapidjson 1.1.0
         - scikit-build 0.17.6
         - versioneer-518
-        - zlib 1.2.13 # required to build triton client
+        - zlib 1.3.1 # required to build triton client
       run:
         # Runtime only requirements. This + setup.py is the definitive runtime requirement list
         # This should be synced with `runtime` in dependencies.yaml
@@ -91,7 +91,7 @@ outputs:
         - docker-py =5.0.*
         - elasticsearch ==8.9.0
         - feedparser =6.0.*
-        - grpcio =1.59.*
+        - grpcio =1.62.*
         - mlflow>=2.10.0,<3
         - mrc
         - networkx=2.8.8
@@ -101,8 +101,6 @@ outputs:
         - python
         - python-confluent-kafka >=1.9.2,<1.10.0a0
         - python-graphviz
-        - pytorch-cuda
-        - pytorch * *cuda*
         - rapids-dask-dependency {{ rapids_version }} # provides dask and distributed
         - requests
         - requests-cache =1.1.*
@@ -188,14 +186,14 @@ outputs:
         # package versions getting installed creating unexpected version conflicts.
         - {{ pin_subpackage('morpheus-core', exact=True) }}
         - cython 3.0.*
-        - glog 0.6.*
+        - glog >=0.7.1,<0.8
         - pip
         - pybind11-stubgen 0.10.5
         - python {{ python }}
         - rapidjson 1.1.0
         - scikit-build 0.17.6
         - versioneer-518
-        - zlib 1.2.13 # required to build triton client
+        - zlib 1.3.1 # required to build triton client
       run:
         - {{ pin_subpackage('morpheus-core', exact=True) }}
 

diff --git a/ci/conda/recipes/morpheus/conda_build_config.yaml b/ci/conda/recipes/morpheus/conda_build_config.yaml
@@ -14,19 +14,19 @@
 # limitations under the License.
 
 c_compiler_version:
-  - 11.2
+  - 12.1
 
 cxx_compiler_version:
-  - 11.2
+  - 12.1
 
 cuda_compiler:
   - cuda-nvcc
 
 cuda_compiler_version:
-  - 12.1
+  - 12.5
 
 python:
   - 3.10
 
 rapids_version:
-  - 24.02
+  - 24.10
diff --git a/ci/conda/recipes/morpheus/meta.yaml b/ci/conda/recipes/morpheus/meta.yaml
@@ -56,7 +56,7 @@ outputs:
         - libtool # Needed for DOCA build
         - ninja =1.11
         - pkg-config =0.29 # for mrc cmake
-        - sysroot_linux-64 =2.17
+        - sysroot_linux-64 =2.28
       host:
         # CUDA dependencies
         - cuda-cudart-dev {{ cuda_compiler_version }}.*
@@ -68,7 +68,7 @@ outputs:
         # Non-CUDA dependencies
         - cudf {{ rapids_version }}
         - cython 3.0.*
-        - glog 0.6.*
+        - glog >=0.7.1,<0.8
         - libcudf {{ rapids_version }}
         - librdkafka >=1.9.2,<1.10.0a0
         - mrc {{ minor_version }}
@@ -80,7 +80,7 @@ outputs:
         - rdma-core >=48 # Needed for DOCA.
         - scikit-build 0.17.6
         - versioneer-518
-        - zlib 1.2.13 # required to build triton client
+        - zlib 1.3.1 # required to build triton client
       run:
         # Runtime only requirements. This + setup.py is the definitive runtime requirement list
         # This should be synced with `runtime` in dependencies.yaml
@@ -95,7 +95,7 @@ outputs:
         - docker-py =5.0.*
         - elasticsearch ==8.9.0
         - feedparser =6.0.*
-        - grpcio =1.59.*
+        - grpcio =1.62.*
         - libwebp>=1.3.2 # Required for CVE mitigation: https://nvd.nist.gov/vuln/detail/CVE-2023-4863
         - mlflow>=2.10.0,<3
         - mrc
@@ -106,8 +106,6 @@ outputs:
         - python
         - python-confluent-kafka >=1.9.2,<1.10.0a0
         - python-graphviz
-        - pytorch-cuda
-        - pytorch * *cuda*
         - rapids-dask-dependency {{ rapids_version }} # provides dask and distributed
         - requests
         - requests-cache =1.1.*

diff --git a/ci/runner/Dockerfile b/ci/runner/Dockerfile
@@ -16,8 +16,8 @@
 # Args used in FROM commands must come first
 ARG FROM_IMAGE="rapidsai/ci-conda"
 ARG CUDA_PKG_VER=12-0
-ARG CUDA_SHORT_VER=12.1
-ARG CUDA_VER=12.1.1
+ARG CUDA_SHORT_VER=12.5
+ARG CUDA_VER=12.5.1
 ARG LINUX_DISTRO=ubuntu
 ARG LINUX_VER=22.04
 ARG PROJ_NAME=morpheus

diff --git a/ci/scripts/run_ci_local.sh b/ci/scripts/run_ci_local.sh
@@ -58,7 +58,7 @@ GIT_BRANCH=$(git branch --show-current)
 GIT_COMMIT=$(git log -n 1 --pretty=format:%H)
 
 LOCAL_CI_TMP=${LOCAL_CI_TMP:-${MORPHEUS_ROOT}/.tmp/local_ci_tmp}
-CONTAINER_VER=${CONTAINER_VER:-240614}
+CONTAINER_VER=${CONTAINER_VER:-241004}
 CUDA_VER=${CUDA_VER:-12.1}
 DOCKER_EXTRA_ARGS=${DOCKER_EXTRA_ARGS:-""}
 

diff --git a/cmake/dependencies.cmake b/cmake/dependencies.cmake
@@ -41,13 +41,11 @@ rapids_find_package(ZLIB
 
 if(MORPHEUS_BUILD_BENCHMARKS)
   # google benchmark
-  # - Expects package to pre-exist in the build environment
   # ================
-  rapids_find_package(benchmark REQUIRED
-    GLOBAL_TARGETS benchmark::benchmark
+  include(${rapids-cmake-dir}/cpm/gbench.cmake)
+  rapids_cpm_gbench(
     BUILD_EXPORT_SET ${PROJECT_NAME}-core-exports
     INSTALL_EXPORT_SET ${PROJECT_NAME}-core-exports
-    FIND_ARGS CONFIG
   )
 endif()
 
@@ -65,13 +63,11 @@ morpheus_utils_configure_glog()
 
 if(MORPHEUS_BUILD_TESTS)
   # google test
-  # - Expects package to pre-exist in the build environment
   # ===========
-  rapids_find_package(GTest REQUIRED
-    GLOBAL_TARGETS GTest::gtest GTest::gmock GTest::gtest_main GTest::gmock_main
+  include(${rapids-cmake-dir}/cpm/gtest.cmake)
+  rapids_cpm_gtest(
     BUILD_EXPORT_SET ${PROJECT_NAME}-core-exports
     INSTALL_EXPORT_SET ${PROJECT_NAME}-core-exports
-    FIND_ARGS CONFIG
   )
 endif()
 

diff --git a/...nvironments/all_cuda-121_arch-x86_64.yaml → ...nvironments/all_cuda-125_arch-x86_64.yaml b/...nvironments/all_cuda-121_arch-x86_64.yaml → ...nvironments/all_cuda-125_arch-x86_64.yaml
@@ -22,18 +22,18 @@ dependencies:
 - clangdev=16
 - click>=8
 - cmake=3.27
-- cuda-cudart-dev=12.1
-- cuda-cudart=12.1
-- cuda-nvcc=12.1
-- cuda-nvml-dev=12.1
-- cuda-nvrtc-dev=12.1
-- cuda-nvrtc=12.1
-- cuda-nvtx-dev=12.1
-- cuda-nvtx=12.1
-- cuda-tools=12.1
-- cuda-version=12.1
-- cudf=24.02
-- cuml=24.02.*
+- cuda-cudart-dev=12.5
+- cuda-cudart=12.5
+- cuda-nvcc=12.5
+- cuda-nvml-dev=12.5
+- cuda-nvrtc-dev=12.5
+- cuda-nvrtc=12.5
+- cuda-nvtx-dev=12.5
+- cuda-nvtx=12.5
+- cuda-sanitizer-api
+- cuda-version=12.5
+- cudf=24.10
+- cuml=24.10.*
 - cupy
 - cxx-compiler
 - cython=3.0
@@ -45,23 +45,28 @@ dependencies:
 - exhale=0.3.6
 - feedparser=6.0
 - flake8
-- gcc_linux-64=11.2
 - git-lfs
-- glog=0.6
-- grpcio-status=1.59
-- grpcio=1.59
-- gxx_linux-64=11.2
+- glog>=0.7.1,<0.8
+- grpcio
+- grpcio-status
+- gtest=1.14
+- gxx=12.1
 - huggingface_hub=0.20.2
 - include-what-you-use=0.20
 - ipython
 - isort
 - jsonpatch>=1.33
 - kfp
-- libcudf=24.02
+- libcublas-dev
+- libcudf=24.10
+- libcufft-dev
+- libcurand-dev
+- libcusolver-dev
 - librdkafka>=1.9.2,<1.10.0a0
 - libtool
 - libwebp=1.3.2
-- mlflow>=2.10.0,<3
+- libzlib >=1.3.1,<2
+- mlflow
 - mrc=24.10
 - myst-parser=0.18.1
 - nbsphinx
@@ -81,6 +86,7 @@ dependencies:
 - pre-commit
 - pybind11-stubgen=0.10.5
 - pydantic
+- pylibcudf=24.10
 - pylint=3.0.3
 - pypdf=3.17.4
 - pypdfium2=4.30
@@ -92,22 +98,20 @@ dependencies:
 - python-docx==1.1.0
 - python-graphviz
 - python=3.10
-- pytorch-cuda
-- pytorch=*=*cuda*
 - rapidjson=1.1.0
-- rapids-dask-dependency=24.02
+- rapids-dask-dependency=24.10
 - rdma-core>=48
 - requests
 - requests-cache=1.1
 - requests-toolbelt=1.0
-- s3fs=2023.12.2
+- s3fs
 - scikit-build=0.17.6
 - scikit-learn=1.3.2
 - sentence-transformers=2.7
 - sphinx
 - sphinx_rtd_theme
 - sqlalchemy<2.0
-- sysroot_linux-64=2.17
+- sysroot_linux-64>=2.28
 - tqdm=4
 - transformers=4.36.2
 - tritonclient=2.34
@@ -120,20 +124,21 @@ dependencies:
 - watchdog=3.0
 - websockets
 - yapf=0.40.1
-- zlib=1.2.13
 - pip:
+  - --extra-index-url https://download.pytorch.org/whl/cu124
   - --find-links https://data.dgl.ai/wheels-test/repo.html
   - --find-links https://data.dgl.ai/wheels/cu121/repo.html
   - databricks-cli < 0.100
   - databricks-connect
   - dgl==2.0.0
   - dglgo
-  - faiss-gpu==1.7.*
+  - faiss-cpu
   - google-search-results==2.4
   - langchain-nvidia-ai-endpoints==0.0.11
   - langchain==0.1.16
   - milvus==2.3.5
   - nemollm==0.3.5
   - pymilvus==2.3.6
   - pytest-kafka==0.6.0
-name: all_cuda-121_arch-x86_64
+  - torch==2.4.0+cu124
+name: all_cuda-125_arch-x86_64