Enable cuml-cpu nightly (#5585)

PR Adds changes necessary to produce cum-cpu nightly packages. PR accomplishes this by: - [x] Using CMake conditional compiling - [x] Using Cython IF compilation temporarily. This will be changed for separating the C++ calls from the Python code, similar to pylibraft. - [x] Adding new conda recipe. Two follow ups from this are: - [ ] Doc and notebook updates - [ ] Additional models added Authors: - Dante Gama Dessavre (https://github.com/dantegd) Approvers: - William Hicks (https://github.com/wphicks) - Ray Douglass (https://github.com/raydouglass) - Corey J. Nolet (https://github.com/cjnolet) URL: #5585
rapidsai · Sep 27, 2023 · 3e5c8e9 · 3e5c8e9
1 parent 4f0f0f3
commit 3e5c8e9
Show file tree

Hide file tree

Showing 53 changed files with 3,058 additions and 2,893 deletions.
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)
 
-VALIDTARGETS="clean libcuml cuml cpp-mgtests prims bench prims-bench cppdocs pydocs"
+VALIDTARGETS="clean libcuml cuml cuml-cpu cpp-mgtests prims bench prims-bench cppdocs pydocs"
 VALIDFLAGS="-v -g -n --allgpuarch --singlegpu --nolibcumltest --nvtx --show_depr_warn --codecov --ccache --configure-only -h --help "
 VALIDARGS="${VALIDTARGETS} ${VALIDFLAGS}"
 HELP="$0 [<target> ...] [<flag> ...]
@@ -27,6 +27,7 @@ HELP="$0 [<target> ...] [<flag> ...]
    libcuml           - build the cuml C++ code only. Also builds the C-wrapper library
                        around the C++ code.
    cuml              - build the cuml Python package
+   cuml-cpu          - build the cuml CPU Python package
    cpp-mgtests       - build libcuml mnmg tests. Builds MPI communicator, adding MPI as dependency.
    prims             - build the ml-prims tests
    bench             - build the libcuml C++ benchmark
@@ -294,3 +295,9 @@ if (! hasArg --configure-only) && (completeBuild || hasArg cuml || hasArg pydocs
         make html
     fi
 fi
+
+if hasArg cuml-cpu; then
+    SKBUILD_CONFIGURE_OPTIONS="-DCUML_CPU=ON -DCMAKE_MESSAGE_LOG_LEVEL=VERBOSE" \
+        SKBUILD_BUILD_OPTIONS="-j${PARALLEL_LEVEL}" \
+        python -m pip install --no-build-isolation --no-deps -v ${REPODIR}/python
+fi
diff --git a/ci/build_python.sh b/ci/build_python.sh
@@ -20,4 +20,13 @@ rapids-conda-retry mambabuild \
   --channel "${CPP_CHANNEL}" \
   conda/recipes/cuml
 
+# Build cuml-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
+  rapids-conda-retry mambabuild \
+  --no-test \
+  conda/recipes/cuml-cpu
+fi
+
 rapids-upload-conda-to-s3 python
diff --git a/conda/recipes/cuml-cpu/build.sh b/conda/recipes/cuml-cpu/build.sh
@@ -0,0 +1,5 @@
+#!/usr/bin/env bash
+# Copyright (c) 2023, NVIDIA CORPORATION.
+
+# This assumes the script is executed from the root of the repo directory
+./build.sh cuml-cpu -v
diff --git a/conda/recipes/cuml-cpu/conda_build_config.yaml b/conda/recipes/cuml-cpu/conda_build_config.yaml
@@ -0,0 +1,11 @@
+c_compiler_version:
+  - 11
+
+cxx_compiler_version:
+  - 11
+
+cmake_version:
+  - ">=3.26.4"
+
+sysroot_version:
+  - "=2.17"
diff --git a/conda/recipes/cuml-cpu/meta.yaml b/conda/recipes/cuml-cpu/meta.yaml
@@ -0,0 +1,50 @@
+# Copyright (c) 2023, NVIDIA CORPORATION.
+
+# Usage:
+#   conda build . -c conda-forge -c numba -c rapidsai -c pytorch
+{% set version = environ.get('GIT_DESCRIBE_TAG', '0.0.0.dev').lstrip('v') + environ.get('VERSION_SUFFIX', '') %}
+{% set py_version = environ['CONDA_PY'] %}
+{% set date_string = environ['RAPIDS_DATE_STRING'] %}
+
+package:
+  name: cuml-cpu
+  version: {{ version }}
+
+source:
+  git_url: ../../..
+
+build:
+  number: {{ GIT_DESCRIBE_NUMBER }}
+  string: py{{ py_version }}_{{ date_string }}_{{ GIT_DESCRIBE_HASH }}_{{ GIT_DESCRIBE_NUMBER }}
+  script_env:
+    - VERSION_SUFFIX
+
+requirements:
+  build:
+    - cmake {{ cmake_version }}
+    - {{ compiler('c') }}
+    - {{ compiler('cxx') }}
+    - sysroot_{{ target_platform }} {{ sysroot_version }}
+    - ninja
+  host:
+    - python x.x
+    - setuptools
+    - scikit-build>=0.13.1
+    - cython>=3.0.0
+  run:
+    - python x.x
+    - numpy
+    - scikit-learn=1.2
+    - hdbscan<=0.8.30
+    - umap-learn=0.5.3
+    - nvtx
+
+tests:                                 # [linux64]
+  imports:                             # [linux64]
+    - cuml-cpu                         # [linux64]
+
+about:
+  home: http://rapids.ai/
+  license: Apache-2.0
+  # license_file: LICENSE
+  summary: cuML-CPU library
diff --git a/docs/source/api.rst b/docs/source/api.rst
@@ -414,10 +414,7 @@ Quasi-Newton
     :members:
 
 Support Vector Machines
-------------------------
-
-.. autoclass:: cuml.svm.SVC
-    :members:
+-----------------------
 
 .. autoclass:: cuml.svm.SVR
     :members:

diff --git a/python/CMakeLists.txt b/python/CMakeLists.txt
@@ -18,10 +18,16 @@ include(../fetch_rapids.cmake)
 
 set(CUML_VERSION 23.10.00)
 
-# We always need CUDA for cuML because the raft dependency brings in a
-# header-only cuco dependency that enables CUDA unconditionally.
-include(rapids-cuda)
-rapids_cuda_init_architectures(cuml-python)
+option(CUML_CPU "Build only cuML CPU Python components." OFF)
+set(language_list "C;CXX")
+
+if(NOT CUML_CPU)
+  # We always need CUDA for cuML GPU because the raft dependency brings in a
+  # header-only cuco dependency that enables CUDA unconditionally.
+  include(rapids-cuda)
+  rapids_cuda_init_architectures(cuml-python)
+  list(APPEND language_list "CUDA")
+endif()
 
 project(
   cuml-python
@@ -30,14 +36,12 @@ project(
             # language to be enabled here. The test project that is built in scikit-build to verify
             # various linking options for the python library is hardcoded to build with C, so until
             # that is fixed we need to keep C.
-            C CXX CUDA
+            ${language_list}
 )
 
 ################################################################################
 # - User Options  --------------------------------------------------------------
 option(CUML_UNIVERSAL "Build all cuML Python components." ON)
-option(CUML_CPU "Build only cuML CPU Python components." OFF)
-option(CUML_GPU "Build only cuML GPU Python components. Not supported yet, defaulting to CUML_UNIVERSAL" OFF)
 option(FIND_CUML_CPP "Search for existing CUML C++ installations before defaulting to local files" OFF)
 option(CUML_BUILD_WHEELS "Whether this build is generating a Python wheel." OFF)
 option(SINGLEGPU "Disable all mnmg components and comms libraries" OFF)
@@ -47,11 +51,11 @@ set(CUML_RAFT_CLONE_ON_PIN OFF)
 
 # todo: use CMAKE_MESSAGE_CONTEXT for prefix for logging.
 # https://github.com/rapidsai/cuml/issues/4843
+message(VERBOSE "CUML_PY: Build only cuML CPU Python components.: ${CUML_CPU}")
 message(VERBOSE "CUML_PY: Searching for existing CUML C++ installations before defaulting to local files: ${FIND_CUML_CPP}")
 message(VERBOSE "CUML_PY: Disabling all mnmg components and comms libraries: ${SINGLEGPU}")
 
-set(CUML_ALGORITHMS "ALL" CACHE STRING "Experimental: Choose which algorithms are built into libcuml++.so. Can specify individual algorithms or groups in a semicolon-separated list.")
-message(VERBOSE "CUML_PY: Building cuML with algorithms: '${CUML_ALGORITHMS}'.")
+set(CUML_ALGORITHMS "ALL" CACHE STRING "Choose which algorithms are built cuML. Can specify individual algorithms or groups in a semicolon-separated list.")
 
 set(CUML_CPP_TARGET "cuml++")
 set(CUML_CPP_SRC "../cpp")
@@ -80,48 +84,68 @@ else()
   set(CUML_PYTHON_TREELITE_TARGET treelite::treelite)
 endif()
 
-if(NOT cuml_FOUND)
-  set(BUILD_CUML_TESTS OFF)
-  set(BUILD_PRIMS_TESTS OFF)
-  set(BUILD_CUML_C_LIBRARY OFF)
-  set(BUILD_CUML_EXAMPLES OFF)
-  set(BUILD_CUML_BENCH OFF)
-  set(BUILD_CUML_PRIMS_BENCH OFF)
-  set(CUML_EXPORT_TREELITE_LINKAGE ON)
-
-  set(_exclude_from_all "")
-  if(CUML_BUILD_WHEELS)
-    # Statically link dependencies if building wheels
-    set(CUDA_STATIC_RUNTIME ON)
-    set(CUML_USE_RAFT_STATIC ON)
-    set(CUML_USE_FAISS_STATIC ON)
-    set(CUML_USE_TREELITE_STATIC ON)
-    set(CUML_USE_CUMLPRIMS_MG_STATIC ON)
-    # Don't install the static libs into wheels
-    set(CUML_EXCLUDE_RAFT_FROM_ALL ON)
-    set(RAFT_EXCLUDE_FAISS_FROM_ALL ON)
-    set(CUML_EXCLUDE_TREELITE_FROM_ALL ON)
-    set(CUML_EXCLUDE_CUMLPRIMS_MG_FROM_ALL ON)
-
-    # Don't install the cuML C++ targets into wheels
-    set(_exclude_from_all EXCLUDE_FROM_ALL)
+if(NOT ${CUML_CPU})
+  if(NOT cuml_FOUND)
+    set(BUILD_CUML_TESTS OFF)
+    set(BUILD_PRIMS_TESTS OFF)
+    set(BUILD_CUML_C_LIBRARY OFF)
+    set(BUILD_CUML_EXAMPLES OFF)
+    set(BUILD_CUML_BENCH OFF)
+    set(BUILD_CUML_PRIMS_BENCH OFF)
+    set(CUML_EXPORT_TREELITE_LINKAGE ON)
+
+    set(_exclude_from_all "")
+    if(CUML_BUILD_WHEELS)
+      # Statically link dependencies if building wheels
+      set(CUDA_STATIC_RUNTIME ON)
+      set(CUML_USE_RAFT_STATIC ON)
+      set(CUML_USE_FAISS_STATIC ON)
+      set(CUML_USE_TREELITE_STATIC ON)
+      set(CUML_USE_CUMLPRIMS_MG_STATIC ON)
+      # Don't install the static libs into wheels
+      set(CUML_EXCLUDE_RAFT_FROM_ALL ON)
+      set(RAFT_EXCLUDE_FAISS_FROM_ALL ON)
+      set(CUML_EXCLUDE_TREELITE_FROM_ALL ON)
+      set(CUML_EXCLUDE_CUMLPRIMS_MG_FROM_ALL ON)
+
+      # Don't install the cuML C++ targets into wheels
+      set(_exclude_from_all EXCLUDE_FROM_ALL)
+    endif()
+
+    add_subdirectory(../cpp cuml-cpp ${_exclude_from_all})
+
+    set(cython_lib_dir cuml)
+    install(TARGETS ${CUML_CPP_TARGET} DESTINATION ${cython_lib_dir})
   endif()
-
-  add_subdirectory(../cpp cuml-cpp ${_exclude_from_all})
-
-  set(cython_lib_dir cuml)
-  install(TARGETS ${CUML_CPP_TARGET} DESTINATION ${cython_lib_dir})
 endif()
 
 if(CUML_CPU)
   set(CUML_UNIVERSAL OFF)
   set(SINGLEGPU ON)
 
   set(CUML_ALGORITHMS "linearregression")
-endif()
+  list(APPEND CUML_ALGORITHMS "pca")
+  list(APPEND CUML_ALGORITHMS "tsvd")
+  list(APPEND CUML_ALGORITHMS "elasticnet")
+  list(APPEND CUML_ALGORITHMS "logisticregression")
+  list(APPEND CUML_ALGORITHMS "ridge")
+  list(APPEND CUML_ALGORITHMS "lasso")
+  list(APPEND CUML_ALGORITHMS "umap")
+  list(APPEND CUML_ALGORITHMS "knn")
+  list(APPEND CUML_ALGORITHMS "hdbscan")
+
+  # this won't be needed when we add CPU libcuml++ (FIL)
+  set(cuml_sg_libraries "")
+
+  list(APPEND CYTHON_FLAGS
+  "--compile-time-env GPUBUILD=0")
+else()
+  set(cuml_sg_libraries cuml::${CUML_CPP_TARGET})
+  set(cuml_mg_libraries cuml::${CUML_CPP_TARGET})
 
-set(cuml_sg_libraries cuml::${CUML_CPP_TARGET})
-set(cuml_mg_libraries cuml::${CUML_CPP_TARGET})
+  list(APPEND CYTHON_FLAGS
+  "--compile-time-env GPUBUILD=1")
+endif()
 
 if(NOT SINGLEGPU)
   include("${CUML_CPP_SRC}/cmake/thirdparty/get_cumlprims_mg.cmake")
@@ -131,9 +155,21 @@ if(NOT SINGLEGPU)
   )
 endif()
 
-rapids_cython_init()
+ ################################################################################
+ # - Build Cython artifacts -----------------------------------------------------
+
+include("${CUML_CPP_SRC}/cmake/modules/ConfigureAlgorithms.cmake")
+include(cmake/ConfigureCythonAlgorithms.cmake)
+
+if(${CUML_CPU})
+  # libcuml requires metrics built if HDSCAN is built, which is not the case
+  # for cuml-cpu
+  unset(metrics_algo)
+endif()
+
+message(VERBOSE "CUML_PY: Building cuML with algorithms: '${CUML_ALGORITHMS}'.")
 
-include(cmake/ConfigureAlgorithmsHelpers.cmake)
+rapids_cython_init()
 
 add_subdirectory(cuml/common)
 add_subdirectory(cuml/internals)

diff --git a/...on/cmake/ConfigureAlgorithmsHelpers.cmake → python/cmake/ConfigureCythonAlgorithms.cmake b/...on/cmake/ConfigureAlgorithmsHelpers.cmake → python/cmake/ConfigureCythonAlgorithms.cmake
@@ -14,19 +14,6 @@
 
 
 function(add_module_gpu_default FILENAME)
-    set (extra_args ${ARGN})
-    list(LENGTH extra_args extra_count)
-    if (${extra_count} GREATER 0 OR
-        ${CUML_UNIVERSAL} OR
-        ${CUML_GPU})
-      list(APPEND cython_sources
-           ${FILENAME})
-      set (cython_sources ${cython_sources} PARENT_SCOPE)
-    endif()
-endfunction()
-
-
-function(add_module_universal_default FILENAME)
     set (extra_args ${ARGN})
     list(LENGTH extra_args extra_count)
     if (${extra_count} GREATER 0 OR