Merge branch-24.12 into branch-25.02

.github/workflows/build.yaml

@@ -52,7 +52,6 @@ jobs:
       branch: ${{ inputs.branch }}
       date: ${{ inputs.date }}
       sha: ${{ inputs.sha }}
-      skip_upload_pkgs: libraft-template
   docs-build:
     if: github.ref_type == 'branch'
     needs: python-build

.github/workflows/pr.yaml

@@ -43,16 +43,8 @@ jobs:
           - '!README.md'
           - '!docs/**'
           - '!img/**'
-          - '!notebooks/**'
           - '!python/**'
           - '!thirdparty/LICENSES/**'
-        test_notebooks:
-          - '**'
-          - '!.devcontainer/**'
-          - '!.pre-commit-config.yaml'
-          - '!CONTRIBUTING.md'
-          - '!README.md'
-          - '!thirdparty/LICENSES/**'
         test_python:
           - '**'
           - '!.devcontainer/**'
@@ -61,7 +53,6 @@ jobs:
           - '!README.md'
           - '!docs/**'
           - '!img/**'
-          - '!notebooks/**'
           - '!thirdparty/LICENSES/**'
   checks:
     secrets: inherit
@@ -89,7 +80,6 @@ jobs:
     with:
       build_type: pull-request
       enable_check_symbols: true
-      symbol_exclusions: raft_cutlass
   conda-python-build:
     needs: conda-cpp-build
     secrets: inherit
@@ -151,5 +141,5 @@ jobs:
       cuda: '["12.5"]'
       build_command: |
         sccache -z;
-        build-all -DBUILD_PRIMS_BENCH=ON -DBUILD_ANN_BENCH=ON --verbose;
+        build-all -DBUILD_PRIMS_BENCH=ON --verbose;
         sccache -s;

.github/workflows/test.yaml

@@ -23,7 +23,6 @@ jobs:
       date: ${{ inputs.date }}
       sha: ${{ inputs.sha }}
       enable_check_symbols: true
-      symbol_exclusions: raft_cutlass
   conda-cpp-tests:
     secrets: inherit
     uses: rapidsai/shared-workflows/.github/workflows/[email protected]

.gitignore

@@ -25,7 +25,6 @@ log
 dask-worker-space/
 *.egg-info/
 *.bin
-bench/ann/data
 temporary_*.json
 
 ## scikit-build

.pre-commit-config.yaml

@@ -62,7 +62,7 @@ repos:
                 entry: ./cpp/scripts/run-cmake-format.sh cmake-format
                 language: python
                 types: [cmake]
-                exclude: .*/thirdparty/.*|.*FindAVX.cmake.*
+                exclude: .*/thirdparty/.*
                 # Note that pre-commit autoupdate does not update the versions
                 # of dependencies, so we'll have to update this manually.
                 additional_dependencies:
@@ -93,7 +93,10 @@ repos:
               - id: codespell
                 additional_dependencies: [tomli]
                 args: ["--toml", "pyproject.toml"]
-                exclude: (?x)^(^CHANGELOG.md$)
+                exclude: |
+                  (?x)
+                    ^CHANGELOG[.]md$|
+                    ^cpp/cmake/patches/cutlass/build-export[.]patch$
       - repo: https://github.com/pre-commit/pre-commit-hooks
         rev: v4.5.0
         hooks:
@@ -114,7 +117,6 @@ repos:
                   cpp/include/raft/neighbors/detail/faiss_select/|
                   cpp/include/raft/thirdparty/|
                   docs/source/sphinxext/github_link[.]py|
-                  cpp/cmake/modules/FindAVX[.]cmake
           - id: verify-alpha-spec
       - repo: https://github.com/rapidsai/dependency-file-generator
         rev: v1.16.0

README.md

@@ -1,7 +1,7 @@
 # <div align="left"><img src="https://rapids.ai/assets/images/rapids_logo.png" width="90px"/>&nbsp;RAFT: Reusable Accelerated Functions and Tools for Vector Search and More</div>
 
 > [!IMPORTANT]
-> The vector search and clustering algorithms in RAFT are being migrated to a new library dedicated to vector search called [cuVS](https://github.com/rapidsai/cuvs). We will continue to support the vector search algorithms in RAFT during this move, but will no longer update them after the RAPIDS 24.06 (June) release. We plan to complete the migration by RAPIDS 24.10 (October) release and will be removing them altogether in the 24.12 (December) release.
+> The vector search and clustering algorithms in RAFT have been formally migrated to a new library dedicated to vector search called [cuVS](https://github.com/rapidsai/cuvs). The headers for the vector search and clustering algorithms in RAFT will remain for a bried period, but will no longer be tested, benchmarked, included in the pre-compiled libraft binary, or otherwise updated after the 24.12 (December 2024) release. We will be removing these headers altogether in a future release. It is strongly suggested to use cuVS for these routines, which include any headers in the `distance`, `neighbors`, `cluster` and `spatial` directories, and use the RAFT versions at your own risk.
 
 ![RAFT tech stack](img/raft-tech-stack-vss.png)
 
@@ -27,7 +27,6 @@
 - [RAFT Reference Documentation](https://docs.rapids.ai/api/raft/stable/): API Documentation.
 - [RAFT Getting Started](./docs/source/quick_start.md): Getting started with RAFT.
 - [Build and Install RAFT](./docs/source/build.md): Instructions for installing and building RAFT.
-- [Example Notebooks](./notebooks): Example jupyter notebooks
 - [RAPIDS Community](https://rapids.ai/community.html): Get help, contribute, and collaborate.
 - [GitHub repository](https://github.com/rapidsai/raft): Download the RAFT source code.
 - [Issue tracker](https://github.com/rapidsai/raft/issues): Report issues or request features.
@@ -120,13 +119,13 @@ auto metric = raft::distance::DistanceType::L2SqrtExpanded;
 raft::distance::pairwise_distance(handle, input.view(), input.view(), output.view(), metric);
 ```
 
-It's also possible to create `raft::device_mdspan` views to invoke the same API with raw pointers and shape information:
+It's also possible to create `raft::device_mdspan` views to invoke the same API with raw pointers and shape information. Take this example from the [NVIDIA cuVS](https://github.com/rapidsai/cuvs) library:
 
 ```c++
 #include <raft/core/device_resources.hpp>
 #include <raft/core/device_mdspan.hpp>
 #include <raft/random/make_blobs.cuh>
-#include <raft/distance/distance.cuh>
+#include <cuvs/distance/distance.hpp>
 
 raft::device_resources handle;
 
@@ -147,21 +146,21 @@ auto output_view = raft::make_device_matrix_view(output, n_samples, n_samples);
 
 raft::random::make_blobs(handle, input_view, labels_view);
 
-auto metric = raft::distance::DistanceType::L2SqrtExpanded;
-raft::distance::pairwise_distance(handle, input_view, input_view, output_view, metric);
+auto metric = cuvs::distance::DistanceType::L2SqrtExpanded;
+cuvs::distance::pairwise_distance(handle, input_view, input_view, output_view, metric);
 ```
 
 
 ### Python Example
 
 The `pylibraft` package contains a Python API for RAFT algorithms and primitives. `pylibraft` integrates nicely into other libraries by being very lightweight with minimal dependencies and accepting any object that supports the `__cuda_array_interface__`, such as [CuPy's ndarray](https://docs.cupy.dev/en/stable/user_guide/interoperability.html#rmm). The number of RAFT algorithms exposed in this package is continuing to grow from release to release.
 
-The example below demonstrates computing the pairwise Euclidean distances between CuPy arrays. Note that CuPy is not a required dependency for `pylibraft`.
+The example below demonstrates computing the pairwise Euclidean distances between CuPy arrays using the [NVIDIA cuVS](https://github.com/rapidsai/cuvs) library. Note that CuPy is not a required dependency for `pylibraft`.
 
 ```python
 import cupy as cp
 
-from pylibraft.distance import pairwise_distance
+from cuvs.distance import pairwise_distance
 
 n_samples = 5000
 n_features = 50
@@ -208,7 +207,7 @@ pylibraft.config.set_output_as(lambda device_ndarray: return device_ndarray.copy
 ```python
 import cupy as cp
 
-from pylibraft.distance import pairwise_distance
+from cuvs.distance import pairwise_distance
 
 n_samples = 5000
 n_features = 50
@@ -223,18 +222,15 @@ pairwise_distance(in1, in2, out=output, metric="euclidean")
 
 ## Installing
 
-RAFT's C++ and Python libraries can both be installed through Conda and the Python libraries through Pip. 
+RAFT's C++ and Python libraries can both be installed through Conda and the Python libraries through Pip.
 
 
 ### Installing C++ and Python through Conda
 
 The easiest way to install RAFT is through conda and several packages are provided.
 - `libraft-headers` C++ headers
-- `libraft` (optional) C++ shared library containing pre-compiled template instantiations and runtime API.
 - `pylibraft` (optional) Python library
 - `raft-dask` (optional) Python library for deployment of multi-node multi-GPU algorithms that use the RAFT `raft::comms` abstraction layer in Dask clusters.
-- `raft-ann-bench` (optional) Benchmarking tool for easily producing benchmarks that compare RAFT's vector search algorithms against other state-of-the-art implementations.
-- `raft-ann-bench-cpu` (optional) Reproducible benchmarking tool similar to above, but doesn't require CUDA to be installed on the machine. Can be used to test in environments with competitive CPUs.
 
 Use the following command, depending on your CUDA version, to install all of the RAFT packages with conda (replace `rapidsai` with `rapidsai-nightly` to install more up-to-date but less stable nightly packages). `mamba` is preferred over the `conda` command.
 ```bash
@@ -255,8 +251,6 @@ You can also install the conda packages individually using the `mamba` command a
 mamba install -c rapidsai -c conda-forge -c nvidia libraft libraft-headers cuda-version=12.5
 ```
 
-If installing the C++ APIs please see [using libraft](https://docs.rapids.ai/api/raft/nightly/using_libraft/) for more information on using the pre-compiled shared library. You can also refer to the [example C++ template project](https://github.com/rapidsai/raft/tree/branch-25.02/cpp/template) for a ready-to-go CMake configuration that you can drop into your project and build against installed RAFT development artifacts above.
-
 ### Installing Python through Pip
 
 `pylibraft` and `raft-dask` both have experimental packages that can be [installed through pip](https://rapids.ai/pip.html#install):
@@ -265,12 +259,10 @@ pip install pylibraft-cu11 --extra-index-url=https://pypi.nvidia.com
 pip install raft-dask-cu11 --extra-index-url=https://pypi.nvidia.com
 ```
 
-These packages statically build RAFT's pre-compiled instantiations and so the C++ headers and pre-compiled shared library won't be readily available to use in your code.
+These packages statically build RAFT's pre-compiled instantiations and so the C++ headers won't be readily available to use in your code.
 
 The [build instructions](https://docs.rapids.ai/api/raft/nightly/build/) contain more details on building RAFT from source and including it in downstream projects. You can also find a more comprehensive version of the above CPM code snippet the [Building RAFT C++ and Python from source](https://docs.rapids.ai/api/raft/nightly/build/#building-c-and-python-from-source) section of the build instructions.
 
-You can find an example [RAFT project template](cpp/template/README.md) in the `cpp/template` directory, which demonstrates how to build a new application with RAFT or incorporate RAFT into an existing CMake project.
-
 
 ## Contributing
 
@@ -284,7 +276,7 @@ When citing RAFT generally, please consider referencing this Github project.
   title={Rapidsai/raft: RAFT contains fundamental widely-used algorithms and primitives for data science, Graph and machine learning.},
   url={https://github.com/rapidsai/raft},
   journal={GitHub},
-  publisher={Nvidia RAPIDS},
+  publisher={NVIDIA RAPIDS},
   author={Rapidsai},
   year={2022}
 }

build.sh

@@ -18,8 +18,8 @@ ARGS=$*
 # scripts, and that this script resides in the repo dir!
 REPODIR=$(cd $(dirname $0); pwd)
 
-VALIDARGS="clean libraft pylibraft raft-dask docs tests template bench-prims bench-ann clean --uninstall  -v -g -n --compile-lib --compile-static-lib --allgpuarch --no-nvtx --cpu-only --show_depr_warn --incl-cache-stats --time -h"
-HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<tool>] [--limit-tests=<targets>] [--limit-bench-prims=<targets>] [--limit-bench-ann=<targets>] [--build-metrics=<filename>]
+VALIDARGS="clean libraft pylibraft raft-dask docs tests bench-prims clean --uninstall  -v -g -n --compile-lib --compile-static-lib --allgpuarch --no-nvtx --show_depr_warn --incl-cache-stats --time -h"
+HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<tool>] [--limit-tests=<targets>] [--limit-bench-prims=<targets>] [--build-metrics=<filename>]
  where <target> is:
    clean            - remove all existing build artifacts and configuration (start over)
    libraft          - build the raft C++ code only. Also builds the C-wrapper library
@@ -29,8 +29,6 @@ HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<to
    docs             - build the documentation
    tests            - build the tests
    bench-prims      - build micro-benchmarks for primitives
-   bench-ann        - build end-to-end ann benchmarks
-   template         - build the example RAFT application template
 
  and <flag> is:
    -v                          - verbose build mode
@@ -39,10 +37,8 @@ HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<to
    --uninstall                 - uninstall files for specified targets which were built and installed prior
    --compile-lib               - compile shared library for all components
    --compile-static-lib        - compile static library for all components
-   --cpu-only                  - build CPU only components without CUDA. Applies to bench-ann only currently.
    --limit-tests               - semicolon-separated list of test executables to compile (e.g. NEIGHBORS_TEST;CLUSTER_TEST)
    --limit-bench-prims         - semicolon-separated list of prims benchmark executables to compute (e.g. NEIGHBORS_PRIMS_BENCH;CLUSTER_PRIMS_BENCH)
-   --limit-bench-ann           - semicolon-separated list of ann benchmark executables to compute (e.g. HNSWLIB_ANN_BENCH;RAFT_IVF_PQ_ANN_BENCH)
    --allgpuarch                - build for all supported GPU architectures
    --no-nvtx                   - disable nvtx (profiling markers), but allow enabling it in downstream projects
    --show_depr_warn            - show cmake deprecation warnings
@@ -71,15 +67,13 @@ BUILD_ALL_GPU_ARCH=0
 BUILD_TESTS=OFF
 BUILD_TYPE=Release
 BUILD_PRIMS_BENCH=OFF
-BUILD_ANN_BENCH=OFF
-BUILD_CPU_ONLY=OFF
 COMPILE_LIBRARY=OFF
 INSTALL_TARGET=install
 BUILD_REPORT_METRICS=""
 BUILD_REPORT_INCL_CACHE_STATS=OFF
 
-TEST_TARGETS="CLUSTER_TEST;CORE_TEST;DISTANCE_TEST;LABEL_TEST;LINALG_TEST;MATRIX_TEST;NEIGHBORS_TEST;NEIGHBORS_ANN_BRUTE_FORCE_TEST;NEIGHBORS_ANN_CAGRA_TEST;NEIGHBORS_ANN_NN_DESCENT_TEST;NEIGHBORS_ANN_IVF_TEST;RANDOM_TEST;SOLVERS_TEST;SPARSE_TEST;SPARSE_DIST_TEST;SPARSE_NEIGHBORS_TEST;STATS_TEST;UTILS_TEST"
-BENCH_TARGETS="CLUSTER_BENCH;CORE_BENCH;NEIGHBORS_BENCH;DISTANCE_BENCH;LINALG_BENCH;MATRIX_BENCH;SPARSE_BENCH;RANDOM_BENCH"
+TEST_TARGETS="CORE_TEST;LABEL_TEST;LINALG_TEST;MATRIX_TEST;RANDOM_TEST;SOLVERS_TEST;SPARSE_TEST;STATS_TEST;UTILS_TEST"
+BENCH_TARGETS="CORE_BENCH;LINALG_BENCH;MATRIX_BENCH;SPARSE_BENCH;RANDOM_BENCH"
 
 CACHE_ARGS=""
 NVTX=ON
@@ -174,21 +168,6 @@ function limitBench {
     fi
 }
 
-function limitAnnBench {
-    # Check for option to limit the set of test binaries to build
-    if [[ -n $(echo $ARGS | { grep -E "\-\-limit\-bench-ann" || true; } ) ]]; then
-        # There are possible weird edge cases that may cause this regex filter to output nothing and fail silently
-        # the true pipe will catch any weird edge cases that may happen and will cause the program to fall back
-        # on the invalid option error
-        LIMIT_ANN_BENCH_TARGETS=$(echo $ARGS | sed -e 's/.*--limit-bench-ann=//' -e 's/ .*//')
-        if [[ -n ${LIMIT_ANN_BENCH_TARGETS} ]]; then
-            # Remove the full LIMIT_TEST_TARGETS argument from list of args so that it passes validArgs function
-            ARGS=${ARGS//--limit-bench-ann=$LIMIT_ANN_BENCH_TARGETS/}
-            ANN_BENCH_TARGETS=${LIMIT_ANN_BENCH_TARGETS}
-        fi
-    fi
-}
-
 function buildMetrics {
     # Check for multiple build-metrics options
     if [[ $(echo $ARGS | { grep -Eo "\-\-build\-metrics" || true; } | wc -l ) -gt 1 ]]; then
@@ -219,7 +198,6 @@ if (( ${NUMARGS} != 0 )); then
     cacheTool
     limitTests
     limitBench
-    limitAnnBench
     buildMetrics
     for a in ${ARGS}; do
         if ! (echo " ${VALIDARGS} " | grep -q " ${a} "); then
@@ -349,18 +327,6 @@ if hasArg bench-prims || (( ${NUMARGS} == 0 )); then
     fi
 fi
 
-if hasArg bench-ann || (( ${NUMARGS} == 0 )); then
-    BUILD_ANN_BENCH=ON
-    CMAKE_TARGET="${CMAKE_TARGET};${ANN_BENCH_TARGETS}"
-    if hasArg --cpu-only; then
-        COMPILE_LIBRARY=OFF
-        BUILD_CPU_ONLY=ON
-        NVTX=OFF
-    else
-        COMPILE_LIBRARY=ON
-    fi
-fi
-
 if hasArg --no-nvtx; then
     NVTX=OFF
 fi
@@ -405,7 +371,7 @@ fi
 
 ################################################################################
 # Configure for building all C++ targets
-if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || hasArg bench-prims || hasArg bench-ann; then
+if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || hasArg bench-prims; then
     if (( ${BUILD_ALL_GPU_ARCH} == 0 )); then
         RAFT_CMAKE_CUDA_ARCHITECTURES="NATIVE"
         echo "Building for the architecture of the GPU in the system..."
@@ -432,8 +398,6 @@ if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || has
           -DDISABLE_DEPRECATION_WARNINGS=${DISABLE_DEPRECATION_WARNINGS} \
           -DBUILD_TESTS=${BUILD_TESTS} \
           -DBUILD_PRIMS_BENCH=${BUILD_PRIMS_BENCH} \
-          -DBUILD_ANN_BENCH=${BUILD_ANN_BENCH} \
-          -DBUILD_CPU_ONLY=${BUILD_CPU_ONLY} \
           -DCMAKE_MESSAGE_LOG_LEVEL=${CMAKE_LOG_LEVEL} \
           ${CACHE_ARGS} \
           ${EXTRA_CMAKE_ARGS}
@@ -505,11 +469,6 @@ if (( ${NUMARGS} == 0 )) || hasArg raft-dask; then
         python -m pip install --no-build-isolation --no-deps --config-settings rapidsai.disable-cuda=true ${REPODIR}/python/raft-dask
 fi
 
-# Build and (optionally) install the raft-ann-bench Python package
-if (( ${NUMARGS} == 0 )) || hasArg bench-ann; then
-    python -m pip install --no-build-isolation --no-deps --config-settings rapidsai.disable-cuda=true ${REPODIR}/python/raft-ann-bench -vvv
-fi
-
 if hasArg docs; then
     set -x
     export RAPIDS_VERSION="$(sed -E -e 's/^([0-9]{2})\.([0-9]{2})\.([0-9]{2}).*$/\1.\2.\3/' "${REPODIR}/VERSION")"
@@ -520,11 +479,3 @@ if hasArg docs; then
     sphinx-build -b html source _html
 fi
 
-################################################################################
-# Initiate build for example RAFT application template (if needed)
-
-if hasArg template; then
-    pushd ${REPODIR}/cpp/template
-    ./build.sh
-    popd
-fi

ci/build_python.sh

@@ -41,30 +41,5 @@ rapids-conda-retry mambabuild \
   conda/recipes/raft-dask
 
 sccache --show-adv-stats
-sccache --zero-stats
-
-# Build ann-bench for each cuda and python version
-rapids-conda-retry mambabuild \
-  --no-test \
-  --channel "${CPP_CHANNEL}" \
-  --channel "${RAPIDS_CONDA_BLD_OUTPUT_DIR}" \
-  conda/recipes/raft-ann-bench
-
-sccache --show-adv-stats
-
-# Build ann-bench-cpu only in CUDA 11 jobs since it only depends on python
-# version
-RAPIDS_CUDA_MAJOR="${RAPIDS_CUDA_VERSION%%.*}"
-if [[ ${RAPIDS_CUDA_MAJOR} == "11" ]]; then
-  sccache --zero-stats
-
-  rapids-conda-retry mambabuild \
-  --no-test \
-  --channel "${CPP_CHANNEL}" \
-  --channel "${RAPIDS_CONDA_BLD_OUTPUT_DIR}" \
-  conda/recipes/raft-ann-bench-cpu
-
-  sccache --show-adv-stats
-fi
 
 rapids-upload-conda-to-s3 python