Forward-merge branch-24.04 into branch-24.06 [skip ci]

.pre-commit-config.yaml

@@ -81,12 +81,6 @@ repos:
                 verbose: true
                 require_serial: true
                 exclude: .*/thirdparty/.*
-              - id: copyright-check
-                name: copyright-check
-                entry: python ./ci/checks/copyright.py --git-modified-only --update-current-year
-                language: python
-                pass_filenames: false
-                additional_dependencies: [gitpython]
               - id: include-check
                 name: include-check
                 entry: python ./cpp/scripts/include_checker.py cpp/bench cpp/include cpp/test
@@ -109,6 +103,23 @@ repos:
         rev: v4.5.0
         hooks:
               - id: check-json
+      - repo: https://github.com/rapidsai/pre-commit-hooks
+        rev: v0.0.1
+        hooks:
+          - id: verify-copyright
+            files: |
+              (?x)
+                  [.](cmake|cpp|cu|cuh|h|hpp|sh|pxd|py|pyx)$|
+                  CMakeLists[.]txt$|
+                  CMakeLists_standalone[.]txt$|
+                  meta[.]yaml$|
+                  setup[.]cfg$
+            exclude: |
+              (?x)
+                  cpp/include/raft/neighbors/detail/faiss_select/|
+                  cpp/include/raft/thirdparty/|
+                  docs/source/sphinxext/github_link[.]py|
+                  cpp/cmake/modules/FindAVX[.]cmake
 
 default_language_version:
       python: python3

build.sh

@@ -305,7 +305,7 @@ if hasArg --allgpuarch; then
     BUILD_ALL_GPU_ARCH=1
 fi
 
-if hasArg --compile-lib || (( ${NUMARGS} == 0 )); then
+if hasArg --compile-lib || hasArg pylibraft || (( ${NUMARGS} == 0 )); then
     COMPILE_LIBRARY=ON
     CMAKE_TARGET="${CMAKE_TARGET};raft_lib"
 fi
@@ -405,7 +405,7 @@ fi
 
 ################################################################################
 # Configure for building all C++ targets
-if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || hasArg bench-prims || hasArg bench-ann; then
+if (( ${NUMARGS} == 0 )) || hasArg libraft || hasArg docs || hasArg tests || hasArg bench-prims || hasArg bench-ann || ((${COMPILE_LIBRARY} == ON )); then
     if (( ${BUILD_ALL_GPU_ARCH} == 0 )); then
         RAFT_CMAKE_CUDA_ARCHITECTURES="NATIVE"
         echo "Building for the architecture of the GPU in the system..."

conda/recipes/libraft/build_libraft_template.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env bash
-# Copyright (c) 2022-2023, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 # Just building template so we verify it uses libraft.so and fail if it doesn't build
-./build.sh template
+./build.sh template --no-nvtx

cpp/bench/prims/CMakeLists.txt

@@ -156,6 +156,14 @@ if(BUILD_PRIMS_BENCH)
     bench/prims/neighbors/knn/ivf_pq_filter_float_int64_t.cu
     bench/prims/neighbors/knn/ivf_pq_int8_t_int64_t.cu
     bench/prims/neighbors/knn/ivf_pq_uint8_t_int64_t.cu
+    src/neighbors/detail/ivf_pq_search_filtering_float_int64_t.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_float_float_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_false_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_true_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_float_half_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_false_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_true_bitset64.cu
+    src/neighbors/detail/ivf_pq_compute_similarity_half_half_bitset64.cu
     bench/prims/neighbors/refine_float_int64_t.cu
     bench/prims/neighbors/refine_uint8_t_int64_t.cu
     bench/prims/main.cpp

cpp/bench/prims/neighbors/knn/ivf_pq_filter_float_int64_t.cu

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2023, NVIDIA CORPORATION.
+ * Copyright (c) 2023-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -14,9 +14,10 @@
  * limitations under the License.
  */
 
-#undef RAFT_EXPLICIT_INSTANTIATE_ONLY  // Enable instantiation of search with filter
 #include "../knn.cuh"
 
+#include <raft_internal/neighbors/ivf_pq_compute_similarity_filters_test-ext.cuh>
+#include <raft_internal/neighbors/ivf_pq_search_test-ext.cuh>
 namespace raft::bench::spatial {
 
 KNN_REGISTER(float, int64_t, ivf_pq_filter_knn, kInputsFilter, kNoCopyOnly, kScopeFull);

cpp/include/raft/core/detail/mdspan_numpy_serializer.hpp

@@ -126,7 +126,8 @@ inline dtype_t get_numpy_dtype()
 }
 
 #if defined(_RAFT_HAS_CUDA)
-template <typename T, typename std::enable_if_t<std::is_same_v<T, half>, bool> = true>
+template <typename T,
+          typename std::enable_if_t<std::is_same_v<std::remove_cv_t<T>, half>, bool> = true>
 inline dtype_t get_numpy_dtype()
 {
   return {RAFT_NUMPY_HOST_ENDIAN_CHAR, 'e', sizeof(T)};