Consolidate pre-compiled specializations into single libraft binary (…

…#1333) Now that we no longer need to rely on the FAISS dependency, this PR: 1. Consolidates the `libraft-distance` and `libraft-nn` targets into a single `libraft` artifact 2. Introduces a new `raft::compiled` target to link against the new `libraft.so` binary. Removes `raft::distance`and `raft::nn` targets. 3. Consolidates `RAFT_DISTANCE_COMPILED` and `RAFT_NN_COMPILED` pre-processor vars into a single `RAFT_COMPILED` (to match similar pattern implementated by `spdlog`) 4. Consolidates `specializations.cuh` headers 5. Updates all docs, scripts, and build infra This change has been a long time coming and is intended to be a 23.04 feature. This is further going to require updates to several projects downstream. Here's a checklist to track that progress: - [x] offline announcement - [x] cuml rapidsai/cuml#5272 - [x] cugraph rapidsai/cugraph#3348 - [x] cuopt rapidsai/cuopt#1023 - [x] cugraph-ops rapidsai/cugraph-ops#429 This PR depended on #1340 (removing FAISS from the build) and on #1202 (replacing the FAISS bfknn w/ our own), both of which have been merged. Closes #824 Authors: - Corey J. Nolet (https://github.com/cjnolet) - Ben Frederickson (https://github.com/benfred) Approvers: - Sevag H (https://github.com/sevagh) - Divye Gala (https://github.com/divyegala) - Ben Frederickson (https://github.com/benfred) - Vyas Ramasubramani (https://github.com/vyasr) URL: #1333
rapidsai · Mar 20, 2023 · 97f8ad7 · 97f8ad7
1 parent 81fd1e0
commit 97f8ad7
Show file tree

Hide file tree

Showing 208 changed files with 687 additions and 686 deletions.
diff --git a/README.md b/README.md
@@ -195,8 +195,7 @@ RAFT itself can be installed through conda, [CMake Package Manager (CPM)](https:
 
 The easiest way to install RAFT is through conda and several packages are provided.
 - `libraft-headers` RAFT headers
-- `libraft-nn` (optional) contains shared libraries for the nearest neighbors primitives.
-- `libraft-distance` (optional) contains shared libraries for distance primitives.
+- `libraft` (optional) shared library of pre-compiled template specializations and runtime APIs.
 - `pylibraft` (optional) Python wrappers around RAFT algorithms and primitives.
 - `raft-dask` (optional) enables deployment of multi-node multi-GPU algorithms that use RAFT `raft::comms` in Dask clusters.
 
@@ -205,9 +204,9 @@ Use the following command to install all of the RAFT packages with conda (replac
 mamba install -c rapidsai -c conda-forge -c nvidia raft-dask pylibraft
 ```
 
-You can also install the `libraft-*` conda packages individually using the `mamba` command above.
+You can also install the conda packages individually using the `mamba` command above.
 
-After installing RAFT, `find_package(raft COMPONENTS nn distance)` can be used in your CUDA/C++ cmake build to compile and/or link against needed dependencies in your raft target. `COMPONENTS` are optional and will depend on the packages installed.
+After installing RAFT, `find_package(raft COMPONENTS compiled distributed)` can be used in your CUDA/C++ cmake build to compile and/or link against needed dependencies in your raft target. `COMPONENTS` are optional and will depend on the packages installed.
 
 ### Pip
 
@@ -266,12 +265,11 @@ find_and_configure_raft(VERSION    ${RAFT_VERSION}.00
 
 Several CMake targets can be made available by adding components in the table below to the `RAFT_COMPONENTS` list above, separated by spaces. The `raft::raft` target will always be available. RAFT headers require, at a minimum, the CUDA toolkit libraries and RMM dependencies.
 
-| Component | Target | Description | Base Dependencies                                                |
-| --- | --- | --- |------------------------------------------------------------------|
-| n/a | `raft::raft` | Full RAFT header library | CUDA toolkit library, RMM, Thrust (optional), NVTools (optional) |
-| distance | `raft::distance` | Pre-compiled template specializations for raft::distance | raft::raft, cuCollections (optional)                             |
-| nn | `raft::nn` | Pre-compiled template specializations for raft::neighbors | raft::raft                                                       |
-| distributed | `raft::distributed` | No specializations | raft::raft, UCX, NCCL                                            |
+| Component   | Target              | Description                                               | Base Dependencies                     |
+|-------------|---------------------|-----------------------------------------------------------|---------------------------------------|
+| n/a         | `raft::raft`        | Full RAFT header library                                  | CUDA toolkit, RMM, NVTX, CCCL, CUTLASS |
+| compiled    | `raft::compiled`    | Pre-compiled template specializations and runtime library | raft::raft                            |
+| distributed | `raft::distributed` | Dependencies for `raft::comms` APIs                       | raft::raft, UCX, NCCL                 |
 
 ### Source
 
@@ -282,7 +280,7 @@ mamba env create --name raft_dev_env -f conda/environments/all_cuda-118_arch-x86
 mamba activate raft_dev_env
 ```
 ```
-./build.sh raft-dask pylibraft libraft tests bench --compile-libs
+./build.sh raft-dask pylibraft libraft tests bench --compile-lib
 ```
 
 The [build](docs/source/build.md) instructions 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++ from source](docs/source/build.md#building-raft-c-from-source-in-cmake) section of the build instructions.

diff --git a/build.sh b/build.sh
@@ -18,7 +18,7 @@ ARGS=$*
 # script, and that this script resides in the repo dir!
 REPODIR=$(cd $(dirname $0); pwd)
 
-VALIDARGS="clean libraft pylibraft raft-dask docs tests bench clean --uninstall  -v -g -n --compile-libs --compile-nn --compile-dist --allgpuarch --no-nvtx --show_depr_warn -h --minimal-deps"
+VALIDARGS="clean libraft pylibraft raft-dask docs tests bench clean --uninstall  -v -g -n --compile-lib --allgpuarch --no-nvtx --show_depr_warn -h"
 HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<tool>] [--limit-tests=<targets>] [--limit-bench=<targets>]
  where <target> is:
    clean            - remove all existing build artifacts and configuration (start over)
@@ -35,12 +35,7 @@ HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<to
    -g                          - build for debug
    -n                          - no install step
    --uninstall                 - uninstall files for specified targets which were built and installed prior
-   --compile-libs              - compile shared libraries for all components
-   --compile-nn                - compile shared library for nn component
-   --compile-dist              - compile shared library for distance and current random components
-                                 (eventually, this will be renamed to something more generic and
-                                  the only option to be supported)
-   --minimal-deps              - disables dependencies like thrust so they can be overridden.
+   --compile-lib               - compile shared libraries for all components
                                  can be useful for a pure header-only install
    --limit-tests               - semicolon-separated list of test executables to compile (e.g. NEIGHBORS_TEST;CLUSTER_TEST)
    --limit-bench               - semicolon-separated list of benchmark executables to compute (e.g. NEIGHBORS_BENCH;CLUSTER_BENCH)
@@ -68,14 +63,11 @@ BUILD_ALL_GPU_ARCH=0
 BUILD_TESTS=OFF
 BUILD_TYPE=Release
 BUILD_BENCH=OFF
-COMPILE_LIBRARIES=OFF
-COMPILE_NN_LIBRARY=OFF
-COMPILE_DIST_LIBRARY=OFF
+COMPILE_LIBRARY=OFF
 INSTALL_TARGET=install
 
 TEST_TARGETS="CLUSTER_TEST;CORE_TEST;DISTANCE_TEST;LABEL_TEST;LINALG_TEST;MATRIX_TEST;RANDOM_TEST;SOLVERS_TEST;SPARSE_TEST;SPARSE_DIST_TEST;SPARSE_NEIGHBORS_TEST;NEIGHBORS_TEST;STATS_TEST;UTILS_TEST"
 BENCH_TARGETS="CLUSTER_BENCH;NEIGHBORS_BENCH;DISTANCE_BENCH;LINALG_BENCH;MATRIX_BENCH;SPARSE_BENCH;RANDOM_BENCH"
-ENABLE_thrust_DEPENDENCY=ON
 
 CACHE_ARGS=""
 NVTX=ON
@@ -254,10 +246,6 @@ if hasArg -n; then
     INSTALL_TARGET=""
 fi
 
-if hasArg --minimal-deps; then
-    ENABLE_thrust_DEPENDENCY=OFF
-fi
-
 if hasArg -v; then
     VERBOSE_FLAG="-v"
     CMAKE_LOG_LEVEL="VERBOSE"
@@ -270,64 +258,38 @@ if hasArg --allgpuarch; then
     BUILD_ALL_GPU_ARCH=1
 fi
 
-if hasArg --compile-libs || (( ${NUMARGS} == 0 )); then
-    COMPILE_LIBRARIES=ON
-fi
-
-if hasArg --compile-nn || hasArg --compile-libs || (( ${NUMARGS} == 0 )); then
-    COMPILE_NN_LIBRARY=ON
-    CMAKE_TARGET="${CMAKE_TARGET};raft_nn_lib"
-fi
-
-if hasArg --compile-dist || hasArg --compile-libs || (( ${NUMARGS} == 0 )); then
-    COMPILE_DIST_LIBRARY=ON
-    CMAKE_TARGET="${CMAKE_TARGET};raft_distance_lib"
+if hasArg --compile-lib || (( ${NUMARGS} == 0 )); then
+    COMPILE_LIBRARY=ON
+    CMAKE_TARGET="${CMAKE_TARGET};raft_lib"
 fi
 
 if hasArg tests || (( ${NUMARGS} == 0 )); then
     BUILD_TESTS=ON
     CMAKE_TARGET="${CMAKE_TARGET};${TEST_TARGETS}"
 
-    # Force compile nn library when needed test targets are specified
-    if [[ $CMAKE_TARGET == *"CLUSTER_TEST"* || \
-          $CMAKE_TARGET == *"SPARSE_DIST_TEST"* || \
-          $CMAKE_TARGET == *"SPARSE_NEIGHBORS_TEST"* || \
-          $CMAKE_TARGET == *"NEIGHBORS_TEST"* || \
-          $CMAKE_TARGET == *"STATS_TEST"* ]]; then
-      echo "-- Enabling nearest neighbors lib for gtests"
-      COMPILE_NN_LIBRARY=ON
-    fi
-
-    # Force compile distance library when needed test targets are specified
+    # Force compile library when needed test targets are specified
     if [[ $CMAKE_TARGET == *"CLUSTER_TEST"* || \
           $CMAKE_TARGET == *"DISTANCE_TEST"* || \
           $CMAKE_TARGET == *"SPARSE_DIST_TEST" || \
           $CMAKE_TARGET == *"SPARSE_NEIGHBORS_TEST"* || \
           $CMAKE_TARGET == *"MATRIX_TEST"* || \
           $CMAKE_TARGET == *"NEIGHBORS_TEST" || \
           $CMAKE_TARGET == *"STATS_TEST"* ]]; then
-      echo "-- Enabling distance lib for gtests"
-      COMPILE_DIST_LIBRARY=ON
+      echo "-- Enabling compiled lib for gtests"
+      COMPILE_LIBRARY=ON
     fi
 fi
 
 if hasArg bench || (( ${NUMARGS} == 0 )); then
     BUILD_BENCH=ON
     CMAKE_TARGET="${CMAKE_TARGET};${BENCH_TARGETS}"
 
-    # Force compile nn library when needed benchmark targets are specified
-    if [[ $CMAKE_TARGET == *"CLUSTER_BENCH"* || \
-          $CMAKE_TARGET == *"NEIGHBORS_BENCH"*  ]]; then
-      echo "-- Enabling nearest neighbors lib for benchmarks"
-      COMPILE_NN_LIBRARY=ON
-    fi
-
-    # Force compile distance library when needed benchmark targets are specified
+    # Force compile library when needed benchmark targets are specified
     if [[ $CMAKE_TARGET == *"CLUSTER_BENCH"* || \
           $CMAKE_TARGET == *"MATRIX_BENCH"* || \
           $CMAKE_TARGET == *"NEIGHBORS_BENCH"* ]]; then
-      echo "-- Enabling distance lib for benchmarks"
-      COMPILE_DIST_LIBRARY=ON
+      echo "-- Enabling compiled lib for benchmarks"
+      COMPILE_LIBRARY=ON
     fi
 
 fi
@@ -398,9 +360,7 @@ if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || has
           -DBUILD_TESTS=${BUILD_TESTS} \
           -DBUILD_BENCH=${BUILD_BENCH} \
           -DCMAKE_MESSAGE_LOG_LEVEL=${CMAKE_LOG_LEVEL} \
-          -DRAFT_COMPILE_NN_LIBRARY=${COMPILE_NN_LIBRARY} \
-          -DRAFT_COMPILE_DIST_LIBRARY=${COMPILE_DIST_LIBRARY} \
-          -DRAFT_ENABLE_thrust_DEPENDENCY=${ENABLE_thrust_DEPENDENCY} \
+          -DRAFT_COMPILE_LIBRARY=${COMPILE_LIBRARY} \
           ${CACHE_ARGS} \
           ${EXTRA_CMAKE_ARGS}
 
@@ -414,34 +374,35 @@ if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || has
   fi
 fi
 
-# Build and (optionally) install the raft-dask Python package
-if (( ${NUMARGS} == 0 )) || hasArg raft-dask; then
+# Build and (optionally) install the pylibraft Python package
+if (( ${NUMARGS} == 0 )) || hasArg pylibraft; then
     # Append `-DFIND_RAFT_CPP=ON` to EXTRA_CMAKE_ARGS unless a user specified the option.
     if [[ "${EXTRA_CMAKE_ARGS}" != *"DFIND_RAFT_CPP"* ]]; then
         EXTRA_CMAKE_ARGS="${EXTRA_CMAKE_ARGS} -DFIND_RAFT_CPP=ON"
     fi
 
-    cd ${REPODIR}/python/raft-dask
+    cd ${REPODIR}/python/pylibraft
     python setup.py build_ext --inplace -- -DCMAKE_PREFIX_PATH="${RAFT_DASK_BUILD_DIR};${INSTALL_PREFIX}" -DCMAKE_LIBRARY_PATH=${LIBRAFT_BUILD_DIR} ${EXTRA_CMAKE_ARGS} -- -j${PARALLEL_LEVEL:-1}
     if [[ ${INSTALL_TARGET} != "" ]]; then
         python setup.py install --single-version-externally-managed --record=record.txt -- -DCMAKE_PREFIX_PATH=${INSTALL_PREFIX} ${EXTRA_CMAKE_ARGS}
     fi
 fi
 
-# Build and (optionally) install the pylibraft Python package
-if (( ${NUMARGS} == 0 )) || hasArg pylibraft; then
+# Build and (optionally) install the raft-dask Python package
+if (( ${NUMARGS} == 0 )) || hasArg raft-dask; then
     # Append `-DFIND_RAFT_CPP=ON` to EXTRA_CMAKE_ARGS unless a user specified the option.
     if [[ "${EXTRA_CMAKE_ARGS}" != *"DFIND_RAFT_CPP"* ]]; then
         EXTRA_CMAKE_ARGS="${EXTRA_CMAKE_ARGS} -DFIND_RAFT_CPP=ON"
     fi
 
-    cd ${REPODIR}/python/pylibraft
+    cd ${REPODIR}/python/raft-dask
     python setup.py build_ext --inplace -- -DCMAKE_PREFIX_PATH="${RAFT_DASK_BUILD_DIR};${INSTALL_PREFIX}" -DCMAKE_LIBRARY_PATH=${LIBRAFT_BUILD_DIR} ${EXTRA_CMAKE_ARGS} -- -j${PARALLEL_LEVEL:-1}
     if [[ ${INSTALL_TARGET} != "" ]]; then
         python setup.py install --single-version-externally-managed --record=record.txt -- -DCMAKE_PREFIX_PATH=${INSTALL_PREFIX} ${EXTRA_CMAKE_ARGS}
     fi
 fi
 
+
 if hasArg docs; then
     set -x
     cd ${DOXYGEN_BUILD_DIR}

diff --git a/ci/build_docs.sh b/ci/build_docs.sh
@@ -24,9 +24,8 @@ VERSION_NUMBER=$(rapids-get-rapids-version-from-git)
 rapids-mamba-retry install \
   --channel "${CPP_CHANNEL}" \
   --channel "${PYTHON_CHANNEL}" \
-  libraft-distance \
+  libraft \
   libraft-headers \
-  libraft-nn \
   pylibraft \
   raft-dask
 

diff --git a/ci/test_cpp.sh b/ci/test_cpp.sh
@@ -26,7 +26,7 @@ rapids-print-env
 
 rapids-mamba-retry install \
   --channel "${CPP_CHANNEL}" \
-  libraft-headers libraft-distance libraft-nn libraft-tests
+  libraft-headers libraft libraft-tests
 
 rapids-logger "Check GPU usage"
 nvidia-smi

diff --git a/ci/test_python.sh b/ci/test_python.sh
@@ -31,7 +31,7 @@ rapids-print-env
 rapids-mamba-retry install \
   --channel "${CPP_CHANNEL}" \
   --channel "${PYTHON_CHANNEL}" \
-  libraft-distance libraft-headers pylibraft raft-dask
+  libraft libraft-headers pylibraft raft-dask
 
 rapids-logger "Check GPU usage"
 nvidia-smi

diff --git a/conda/recipes/libraft/build_libraft_nn.sh → conda/recipes/libraft/build_libraft.sh b/conda/recipes/libraft/build_libraft_nn.sh → conda/recipes/libraft/build_libraft.sh
@@ -1,4 +1,4 @@
 #!/usr/bin/env bash
 # Copyright (c) 2022-2023, NVIDIA CORPORATION.
 
-./build.sh libraft --allgpuarch --compile-nn --no-nvtx
+./build.sh libraft --allgpuarch --compile-lib --no-nvtx
diff --git a/conda/recipes/libraft/build_libraft_distance.sh b/conda/recipes/libraft/build_libraft_distance.sh
diff --git a/conda/recipes/libraft/meta.yaml b/conda/recipes/libraft/meta.yaml
@@ -76,9 +76,9 @@ outputs:
       home: https://rapids.ai/
       license: Apache-2.0
       summary: libraft-headers library
-  - name: libraft-distance
+  - name: libraft
     version: {{ version }}
-    script: build_libraft_distance.sh
+    script: build_libraft.sh
     build:
       script_env: *script_env
       number: {{ GIT_DESCRIBE_NUMBER }}
@@ -109,42 +109,7 @@ outputs:
     about:
       home: https://rapids.ai/
       license: Apache-2.0
-      summary: libraft-distance library
-  - name: libraft-nn
-    version: {{ version }}
-    script: build_libraft_nn.sh
-    build:
-      script_env: *script_env
-      number: {{ GIT_DESCRIBE_NUMBER }}
-      string: cuda{{ cuda_major }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }}
-      ignore_run_exports_from:
-        - {{ compiler('cuda') }}
-    requirements:
-      build:
-        - {{ compiler('c') }}
-        - {{ compiler('cuda') }} {{ cuda_version }}
-        - {{ compiler('cxx') }}
-        - cmake {{ cmake_version }}
-        - ninja
-        - sysroot_{{ target_platform }} {{ sysroot_version }}
-      host:
-        - {{ pin_subpackage('libraft-headers', exact=True) }}
-        - cuda-profiler-api {{ cuda_profiler_api_host_version }}
-        - lapack
-        - libcublas {{ libcublas_host_version }}
-        - libcublas-dev {{ libcublas_host_version }}
-        - libcurand {{ libcurand_host_version }}
-        - libcurand-dev {{ libcurand_host_version }}
-        - libcusolver {{ libcusolver_host_version }}
-        - libcusolver-dev {{ libcusolver_host_version }}
-        - libcusparse {{ libcusparse_host_version }}
-        - libcusparse-dev {{ libcusparse_host_version }}
-      run:
-        - {{ pin_subpackage('libraft-headers', exact=True) }}
-    about:
-      home: https://rapids.ai/
-      license: Apache-2.0
-      summary: libraft-nn library
+      summary: libraft library
   - name: libraft-tests
     version: {{ version }}
     script: build_libraft_tests.sh
@@ -163,9 +128,8 @@ outputs:
         - ninja
         - sysroot_{{ target_platform }} {{ sysroot_version }}
       host:
-        - {{ pin_subpackage('libraft-distance', exact=True) }}
+        - {{ pin_subpackage('libraft', exact=True) }}
         - {{ pin_subpackage('libraft-headers', exact=True) }}
-        - {{ pin_subpackage('libraft-nn', exact=True) }}
         - cuda-profiler-api {{ cuda_profiler_api_host_version }}
         - gmock {{ gtest_version }}
         - gtest {{ gtest_version }}
@@ -178,9 +142,8 @@ outputs:
         - libcusparse {{ libcusparse_host_version }}
         - libcusparse-dev {{ libcusparse_host_version }}
       run:
-        - {{ pin_subpackage('libraft-distance', exact=True) }}
+        - {{ pin_subpackage('libraft', exact=True) }}
         - {{ pin_subpackage('libraft-headers', exact=True) }}
-        - {{ pin_subpackage('libraft-nn', exact=True) }}
         - gmock {{ gtest_version }}
         - gtest {{ gtest_version }}
     about:

diff --git a/conda/recipes/pylibraft/meta.yaml b/conda/recipes/pylibraft/meta.yaml
@@ -34,7 +34,7 @@ requirements:
     - cuda-python >=11.7.1,<12.0
     - cudatoolkit ={{ cuda_version }}
     - cython >=0.29,<0.30
-    - libraft-distance {{ version }}
+    - libraft {{ version }}
     - libraft-headers {{ version }}
     - python x.x
     - rmm ={{ minor_version }}
@@ -43,7 +43,7 @@ requirements:
   run:
     - {{ pin_compatible('cudatoolkit', max_pin='x', min_pin='x') }}
     - cuda-python >=11.7.1,<12.0
-    - libraft-distance {{ version }}
+    - libraft {{ version }}
     - libraft-headers {{ version }}
     - python x.x