diff --git a/build.sh b/build.sh
index ea69f8d0d..eb360ff32 100755
--- a/build.sh
+++ b/build.sh
@@ -40,7 +40,7 @@ HELP="$0 [<target> ...] [<flag> ...] [--cmake-args=\"<args>\"] [--cache-tool=<to
    --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-prims         - semicolon-separated list of prims benchmark executables to compute (e.g. NEIGHBORS_MICRO_BENCH;CLUSTER_MICRO_BENCH)
    --limit-bench-ann           - semicolon-separated list of ann benchmark executables to compute (e.g. HNSWLIB_ANN_BENCH;CUVS_IVF_PQ_ANN_BENCH)
    --allgpuarch                - build for all supported GPU architectures
    --no-nvtx                   - disable nvtx (profiling markers), but allow enabling it in downstream projects
@@ -69,7 +69,7 @@ VERBOSE_FLAG=""
 BUILD_ALL_GPU_ARCH=0
 BUILD_TESTS=OFF
 BUILD_TYPE=Release
-BUILD_PRIMS_BENCH=OFF
+BUILD_MICRO_BENCH=OFF
 BUILD_ANN_BENCH=OFF
 BUILD_CPU_ONLY=OFF
 COMPILE_LIBRARY=OFF
@@ -77,8 +77,8 @@ 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_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="CLUSTER_TEST;DISTANCE_TEST;NEIGHBORS_TEST;NEIGHBORS_ANN_CAGRA_TEST;NEIGHBORS_ANN_NN_DESCENT_TEST;NEIGHBORS_ANN_IVF_TEST"
+BENCH_TARGETS="CLUSTER_BENCH;NEIGHBORS_BENCH;DISTANCE_BENCH"
 
 CACHE_ARGS=""
 NVTX=ON
@@ -164,11 +164,11 @@ function limitBench {
         # 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_PRIMS_BENCH_TARGETS=$(echo $ARGS | sed -e 's/.*--limit-bench-prims=//' -e 's/ .*//')
-        if [[ -n ${LIMIT_PRIMS_BENCH_TARGETS} ]]; then
-            # Remove the full LIMIT_PRIMS_BENCH_TARGETS argument from list of args so that it passes validArgs function
-            ARGS=${ARGS//--limit-bench-prims=$LIMIT_PRIMS_BENCH_TARGETS/}
-            PRIMS_BENCH_TARGETS=${LIMIT_PRIMS_BENCH_TARGETS}
+        LIMIT_MICRO_BENCH_TARGETS=$(echo $ARGS | sed -e 's/.*--limit-bench-prims=//' -e 's/ .*//')
+        if [[ -n ${LIMIT_MICRO_BENCH_TARGETS} ]]; then
+            # Remove the full LIMIT_MICRO_BENCH_TARGETS argument from list of args so that it passes validArgs function
+            ARGS=${ARGS//--limit-bench-prims=$LIMIT_MICRO_BENCH_TARGETS/}
+            MICRO_BENCH_TARGETS=${LIMIT_MICRO_BENCH_TARGETS}
         fi
     fi
 }
@@ -306,13 +306,13 @@ fi
 
 if hasArg --compile-lib || (( ${NUMARGS} == 0 )); then
     COMPILE_LIBRARY=ON
-    CMAKE_TARGET="${CMAKE_TARGET};cuvs_lib"
+    CMAKE_TARGET="${CMAKE_TARGET};cuvs"
 fi
 
-if hasArg --compile-static-lib || (( ${NUMARGS} == 0 )); then
-    COMPILE_LIBRARY=ON
-    CMAKE_TARGET="${CMAKE_TARGET};cuvs_lib_static"
-fi
+#if hasArg --compile-static-lib || (( ${NUMARGS} == 0 )); then
+#    COMPILE_LIBRARY=ON
+#    CMAKE_TARGET="${CMAKE_TARGET};cuvs_lib_static"
+#fi
 
 if hasArg tests || (( ${NUMARGS} == 0 )); then
     BUILD_TESTS=ON
@@ -321,13 +321,10 @@ if hasArg tests || (( ${NUMARGS} == 0 )); then
     # Force compile library when needed test targets are specified
     if [[ $CMAKE_TARGET == *"CLUSTER_TEST"* || \
           $CMAKE_TARGET == *"DISTANCE_TEST"* || \
-          $CMAKE_TARGET == *"MATRIX_TEST"* || \
           $CMAKE_TARGET == *"NEIGHBORS_ANN_CAGRA_TEST"* || \
           $CMAKE_TARGET == *"NEIGHBORS_ANN_IVF_TEST"* || \
           $CMAKE_TARGET == *"NEIGHBORS_ANN_NN_DESCENT_TEST"* || \
           $CMAKE_TARGET == *"NEIGHBORS_TEST"* || \
-          $CMAKE_TARGET == *"SPARSE_DIST_TEST" || \
-          $CMAKE_TARGET == *"SPARSE_NEIGHBORS_TEST"* || \
           $CMAKE_TARGET == *"STATS_TEST"* ]]; then
       echo "-- Enabling compiled lib for gtests"
       COMPILE_LIBRARY=ON
@@ -335,13 +332,12 @@ if hasArg tests || (( ${NUMARGS} == 0 )); then
 fi
 
 if hasArg bench-prims || (( ${NUMARGS} == 0 )); then
-    BUILD_PRIMS_BENCH=ON
-    CMAKE_TARGET="${CMAKE_TARGET};${PRIMS_BENCH_TARGETS}"
+    BUILD_MICRO_BENCH=ON
+    CMAKE_TARGET="${CMAKE_TARGET};${MICRO_BENCH_TARGETS}"
 
     # Force compile library when needed benchmark targets are specified
-    if [[ $CMAKE_TARGET == *"CLUSTER_PRIMS_BENCH"* || \
-          $CMAKE_TARGET == *"MATRIX_PRIMS_BENCH"* || \
-          $CMAKE_TARGET == *"NEIGHBORS_PRIMS_BENCH"* ]]; then
+    if [[ $CMAKE_TARGET == *"CLUSTER_MICRO_BENCH"* || \
+          $CMAKE_TARGET == *"NEIGHBORS_MICRO_BENCH"* ]]; then
       echo "-- Enabling compiled lib for benchmarks"
       COMPILE_LIBRARY=ON
     fi
@@ -427,7 +423,7 @@ if (( ${NUMARGS} == 0 )) || hasArg libcuvs || hasArg docs || hasArg tests || has
           -DCUDA_LOG_COMPILE_TIME=${LOG_COMPILE_TIME} \
           -DDISABLE_DEPRECATION_WARNINGS=${DISABLE_DEPRECATION_WARNINGS} \
           -DBUILD_TESTS=${BUILD_TESTS} \
-          -DBUILD_PRIMS_BENCH=${BUILD_PRIMS_BENCH} \
+          -DBUILD_MICRO_BENCH=${BUILD_MICRO_BENCH} \
           -DBUILD_ANN_BENCH=${BUILD_ANN_BENCH} \
           -DBUILD_CPU_ONLY=${BUILD_CPU_ONLY} \
           -DCMAKE_MESSAGE_LOG_LEVEL=${CMAKE_LOG_LEVEL} \
diff --git a/cpp/CMakeLists.txt b/cpp/CMakeLists.txt
index acb77ec8c..9b9475c93 100644
--- a/cpp/CMakeLists.txt
+++ b/cpp/CMakeLists.txt
@@ -11,7 +11,7 @@
 # or implied. See the License for the specific language governing permissions and limitations under
 # the License.
 set(RAPIDS_VERSION "24.02")
-set(RAFT_VERSION "24.02.00")
+set(CUVS_VERSION "24.02.00")
 
 cmake_minimum_required(VERSION 3.26.4 FATAL_ERROR)
 include(../fetch_rapids.cmake)
@@ -20,7 +20,7 @@ include(rapids-cpm)
 include(rapids-export)
 include(rapids-find)
 
-option(BUILD_CPU_ONLY "Build CPU only components. Applies to RAFT ANN benchmarks currently" OFF)
+option(BUILD_CPU_ONLY "Build CPU only components. Applies to CUVS benchmarks currently" OFF)
 
 # workaround for rapids_cuda_init_architectures not working for arch detection with
 # enable_language(CUDA)
@@ -28,18 +28,18 @@ set(lang_list "CXX")
 
 if(NOT BUILD_CPU_ONLY)
   include(rapids-cuda)
-  rapids_cuda_init_architectures(RAFT)
+  rapids_cuda_init_architectures(cuVS)
   list(APPEND lang_list "CUDA")
 endif()
 
 project(
-  RAFT
-  VERSION ${RAFT_VERSION}
+  cuVS
+  VERSION ${CUVS_VERSION}
   LANGUAGES ${lang_list}
 )
 
 # Write the version header
-rapids_cmake_write_version_file(include/raft/version_config.hpp)
+rapids_cmake_write_version_file(include/cuvs/version_config.hpp)
 
 # ##################################################################################################
 # * build type ---------------------------------------------------------------
@@ -53,10 +53,10 @@ set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
 # ##################################################################################################
 # * User Options  ------------------------------------------------------------
 
-option(BUILD_SHARED_LIBS "Build raft shared libraries" ON)
-option(BUILD_TESTS "Build raft unit-tests" ON)
-option(BUILD_PRIMS_BENCH "Build raft C++ benchmark tests" OFF)
-option(BUILD_ANN_BENCH "Build raft ann benchmarks" OFF)
+option(BUILD_SHARED_LIBS "Build cuvs shared libraries" ON)
+option(BUILD_TESTS "Build cuvs unit-tests" ON)
+option(BUILD_MICRO_BENCH "Build cuvs C++ micro benchmarks" OFF)
+option(BUILD_ANN_BENCH "Build cuvs ann benchmarks" OFF)
 option(CUDA_ENABLE_KERNELINFO "Enable kernel resource usage info" OFF)
 option(CUDA_ENABLE_LINEINFO
        "Enable the -lineinfo option for nvcc (useful for cuda-memcheck / profiler)" OFF
@@ -66,20 +66,16 @@ option(CUDA_LOG_COMPILE_TIME "Write a log of compilation times to nvcc_compile_l
 option(DETECT_CONDA_ENV "Enable detection of conda environment for dependencies" ON)
 option(DISABLE_DEPRECATION_WARNINGS "Disable deprecaction warnings " ON)
 option(DISABLE_OPENMP "Disable OpenMP" OFF)
-option(RAFT_NVTX "Enable nvtx markers" OFF)
+option(CUVS_NVTX "Enable nvtx markers" OFF)
 
-set(RAFT_COMPILE_LIBRARY_DEFAULT OFF)
 if((BUILD_TESTS
-    OR BUILD_PRIMS_BENCH
+    OR BUILD_MICRO_BENCH
     OR BUILD_ANN_BENCH
    )
    AND NOT BUILD_CPU_ONLY
 )
-  set(RAFT_COMPILE_LIBRARY_DEFAULT ON)
+
 endif()
-option(RAFT_COMPILE_LIBRARY "Enable building raft shared library instantiations"
-       ${RAFT_COMPILE_LIBRARY_DEFAULT}
-)
 
 if(BUILD_CPU_ONLY)
   set(BUILD_SHARED_LIBS OFF)
@@ -92,22 +88,21 @@ endif()
 set(THREADS_PREFER_PTHREAD_FLAG ON)
 
 include(CMakeDependentOption)
-# cmake_dependent_option( RAFT_USE_FAISS_STATIC "Build and statically link the FAISS library for
-# nearest neighbors search on GPU" ON RAFT_COMPILE_LIBRARY OFF )
-
-message(VERBOSE "RAFT: Building optional components: ${raft_FIND_COMPONENTS}")
-message(VERBOSE "RAFT: Build RAFT unit-tests: ${BUILD_TESTS}")
-message(VERBOSE "RAFT: Building raft C++ benchmarks: ${BUILD_PRIMS_BENCH}")
-message(VERBOSE "RAFT: Building ANN benchmarks: ${BUILD_ANN_BENCH}")
-message(VERBOSE "RAFT: Build CPU only components: ${BUILD_CPU_ONLY}")
-message(VERBOSE "RAFT: Enable detection of conda environment for dependencies: ${DETECT_CONDA_ENV}")
-message(VERBOSE "RAFT: Disable depreaction warnings " ${DISABLE_DEPRECATION_WARNINGS})
-message(VERBOSE "RAFT: Disable OpenMP: ${DISABLE_OPENMP}")
-message(VERBOSE "RAFT: Enable kernel resource usage info: ${CUDA_ENABLE_KERNELINFO}")
-message(VERBOSE "RAFT: Enable lineinfo in nvcc: ${CUDA_ENABLE_LINEINFO}")
-message(VERBOSE "RAFT: Enable nvtx markers: ${RAFT_NVTX}")
+# cmake_dependent_option( CUVS_USE_FAISS_STATIC "Build and statically link the FAISS library for
+# nearest neighbors search on GPU" ON CUVS_COMPILE_LIBRARY OFF )
+
+message(VERBOSE "cuVS: Build cuVS unit-tests: ${BUILD_TESTS}")
+message(VERBOSE "cuVS: Building cuvs C++ benchmarks: ${BUILD_MICRO_BENCH}")
+message(VERBOSE "cuVS: Building ANN benchmarks: ${BUILD_ANN_BENCH}")
+message(VERBOSE "cuVS: Build CPU only components: ${BUILD_CPU_ONLY}")
+message(VERBOSE "cuVS: Enable detection of conda environment for dependencies: ${DETECT_CONDA_ENV}")
+message(VERBOSE "cuVS: Disable depreaction warnings " ${DISABLE_DEPRECATION_WARNINGS})
+message(VERBOSE "cuVS: Disable OpenMP: ${DISABLE_OPENMP}")
+message(VERBOSE "cuVS: Enable kernel resource usage info: ${CUDA_ENABLE_KERNELINFO}")
+message(VERBOSE "cuVS: Enable lineinfo in nvcc: ${CUDA_ENABLE_LINEINFO}")
+message(VERBOSE "cuVS: Enable nvtx markers: ${CUVS_NVTX}")
 message(VERBOSE
-        "RAFT: Statically link the CUDA toolkit runtime and libraries: ${CUDA_STATIC_RUNTIME}"
+        "cuVS: Statically link the CUDA toolkit runtime and libraries: ${CUDA_STATIC_RUNTIME}"
 )
 
 # Set RMM logging level
@@ -118,7 +113,7 @@ set(RMM_LOGGING_LEVEL
 set_property(
   CACHE RMM_LOGGING_LEVEL PROPERTY STRINGS "TRACE" "DEBUG" "INFO" "WARN" "ERROR" "CRITICAL" "OFF"
 )
-message(VERBOSE "RAFT: RMM_LOGGING_LEVEL = '${RMM_LOGGING_LEVEL}'.")
+message(VERBOSE "cuVS: RMM_LOGGING_LEVEL = '${RMM_LOGGING_LEVEL}'.")
 
 # ##################################################################################################
 # * Conda environment detection ----------------------------------------------
@@ -127,7 +122,7 @@ if(DETECT_CONDA_ENV)
   rapids_cmake_support_conda_env(conda_env MODIFY_PREFIX_PATH)
   if(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT AND DEFINED ENV{CONDA_PREFIX})
     message(
-      STATUS "RAFT: No CMAKE_INSTALL_PREFIX argument detected, setting to: $ENV{CONDA_PREFIX}"
+      STATUS "cuVS: No CMAKE_INSTALL_PREFIX argument detected, setting to: $ENV{CONDA_PREFIX}"
     )
     set(CMAKE_INSTALL_PREFIX "$ENV{CONDA_PREFIX}")
   endif()
@@ -150,8 +145,8 @@ if(NOT BUILD_CPU_ONLY)
   # * set other CUDA compilation flags
   rapids_find_package(
     CUDAToolkit REQUIRED
-    BUILD_EXPORT_SET raft-exports
-    INSTALL_EXPORT_SET raft-exports
+    BUILD_EXPORT_SET cuvs-exports
+    INSTALL_EXPORT_SET cuvs-exports
   )
 else()
   add_compile_definitions(BUILD_CPU_ONLY)
@@ -160,11 +155,11 @@ endif()
 if(NOT DISABLE_OPENMP)
   rapids_find_package(
     OpenMP REQUIRED
-    BUILD_EXPORT_SET raft-exports
-    INSTALL_EXPORT_SET raft-exports
+    BUILD_EXPORT_SET cuvs-exports
+    INSTALL_EXPORT_SET cuvs-exports
   )
   if(OPENMP_FOUND)
-    message(VERBOSE "RAFT: OpenMP found in ${OpenMP_CXX_INCLUDE_DIRS}")
+    message(VERBOSE "cuVS: OpenMP found in ${OpenMP_CXX_INCLUDE_DIRS}")
   endif()
 endif()
 
@@ -177,375 +172,265 @@ include(cmake/modules/ConfigureCUDA.cmake)
 rapids_cpm_init()
 
 if(NOT BUILD_CPU_ONLY)
-  # thrust before rmm/cuco so we get the right version of thrust/cub
-  include(cmake/thirdparty/get_thrust.cmake)
-  include(cmake/thirdparty/get_rmm.cmake)
-  include(cmake/thirdparty/get_cutlass.cmake)
-
-  include(${rapids-cmake-dir}/cpm/cuco.cmake)
-  rapids_cpm_cuco(BUILD_EXPORT_SET raft-exports INSTALL_EXPORT_SET raft-exports)
+  include(cmake/thirdparty/get_raft.cmake)
 endif()
 
 if(BUILD_TESTS)
   include(cmake/thirdparty/get_gtest.cmake)
 endif()
 
-if(BUILD_PRIMS_BENCH OR BUILD_ANN_BENCH)
+if(BUILD_MICRO_BENCH OR BUILD_ANN_BENCH)
   include(${rapids-cmake-dir}/cpm/gbench.cmake)
   rapids_cpm_gbench()
 endif()
 
 # ##################################################################################################
-# * raft ---------------------------------------------------------------------
-add_library(raft INTERFACE)
-add_library(raft::raft ALIAS raft)
+# * cuvs ---------------------------------------------------------------------
+
+add_library(
+  cuvs SHARED
+  src/distance/detail/pairwise_matrix/dispatch_canberra_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_canberra_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_correlation_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_correlation_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_cosine_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_cosine_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_kl_divergence_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_kl_divergence_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l1_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l1_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l2_expanded_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l2_expanded_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l_inf_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_l_inf_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_float_float_float_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_rbf.cu
+  src/distance/detail/pairwise_matrix/dispatch_russel_rao_double_double_double_int.cu
+  src/distance/detail/pairwise_matrix/dispatch_russel_rao_float_float_float_int.cu
+  src/distance/distance.cu
+  src/distance/fused_l2_nn.cu
+  src/matrix/detail/select_k_double_int64_t.cu
+  src/matrix/detail/select_k_double_uint32_t.cu
+  src/matrix/detail/select_k_float_int64_t.cu
+  src/matrix/detail/select_k_float_uint32_t.cu
+  src/matrix/detail/select_k_float_int32.cu
+  src/matrix/detail/select_k_half_int64_t.cu
+  src/matrix/detail/select_k_half_uint32_t.cu
+  src/neighbors/ball_cover.cu
+  src/neighbors/brute_force_fused_l2_knn_float_int64_t.cu
+  src/neighbors/brute_force_knn_int64_t_float_int64_t.cu
+  src/neighbors/brute_force_knn_int64_t_float_uint32_t.cu
+  src/neighbors/brute_force_knn_int_float_int.cu
+  src/neighbors/brute_force_knn_uint32_t_float_uint32_t.cu
+  src/neighbors/brute_force_knn_index_float.cu
+  src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim1024_t32.cu
+  src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim1024_t32.cu
+  src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim1024_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_float_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_single_cta_float_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_single_cta_float_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_float_uint32_dim1024_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim1024_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim128_t8.cu
+  src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim256_t16.cu
+  src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim512_t32.cu
+  src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim1024_t32.cu
+  src/neighbors/detail/ivf_flat_interleaved_scan_float_float_int64_t.cu
+  src/neighbors/detail/ivf_flat_interleaved_scan_int8_t_int32_t_int64_t.cu
+  src/neighbors/detail/ivf_flat_interleaved_scan_uint8_t_uint32_t_int64_t.cu
+  src/neighbors/detail/ivf_flat_search.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_float_float.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_false.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_true.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_float_half.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_false.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_true.cu
+  src/neighbors/detail/ivf_pq_compute_similarity_half_half.cu
+  src/neighbors/detail/refine_host_float_float.cpp
+  src/neighbors/detail/refine_host_int8_t_float.cpp
+  src/neighbors/detail/refine_host_uint8_t_float.cpp
+  src/neighbors/detail/selection_faiss_int32_t_float.cu
+  src/neighbors/detail/selection_faiss_int_double.cu
+  src/neighbors/detail/selection_faiss_long_float.cu
+  src/neighbors/detail/selection_faiss_size_t_double.cu
+  src/neighbors/detail/selection_faiss_size_t_float.cu
+  src/neighbors/detail/selection_faiss_uint32_t_float.cu
+  src/neighbors/detail/selection_faiss_int64_t_double.cu
+  src/neighbors/detail/selection_faiss_int64_t_half.cu
+  src/neighbors/detail/selection_faiss_uint32_t_double.cu
+  src/neighbors/detail/selection_faiss_uint32_t_half.cu
+  src/neighbors/ivf_flat_build_float_int64_t.cu
+  src/neighbors/ivf_flat_build_int8_t_int64_t.cu
+  src/neighbors/ivf_flat_build_uint8_t_int64_t.cu
+  src/neighbors/ivf_flat_extend_float_int64_t.cu
+  src/neighbors/ivf_flat_extend_int8_t_int64_t.cu
+  src/neighbors/ivf_flat_extend_uint8_t_int64_t.cu
+  src/neighbors/ivf_flat_search_float_int64_t.cu
+  src/neighbors/ivf_flat_search_int8_t_int64_t.cu
+  src/neighbors/ivf_flat_search_uint8_t_int64_t.cu
+  src/neighbors/ivfpq_build_float_int64_t.cu
+  src/neighbors/ivfpq_build_int8_t_int64_t.cu
+  src/neighbors/ivfpq_build_uint8_t_int64_t.cu
+  src/neighbors/ivfpq_extend_float_int64_t.cu
+  src/neighbors/ivfpq_extend_int8_t_int64_t.cu
+  src/neighbors/ivfpq_extend_uint8_t_int64_t.cu
+  src/neighbors/ivfpq_search_float_int64_t.cu
+  src/neighbors/ivfpq_search_int8_t_int64_t.cu
+  src/neighbors/ivfpq_search_uint8_t_int64_t.cu
+  src/neighbors/refine_float_float.cu
+  src/neighbors/refine_int8_t_float.cu
+  src/neighbors/refine_uint8_t_float.cu
+  src/cuvs_runtime/cluster/cluster_cost.cuh
+  src/cuvs_runtime/cluster/cluster_cost_double.cu
+  src/cuvs_runtime/cluster/cluster_cost_float.cu
+  src/cuvs_runtime/cluster/kmeans_fit_double.cu
+  src/cuvs_runtime/cluster/kmeans_fit_float.cu
+  src/cuvs_runtime/cluster/kmeans_init_plus_plus_double.cu
+  src/cuvs_runtime/cluster/kmeans_init_plus_plus_float.cu
+  src/cuvs_runtime/cluster/update_centroids.cuh
+  src/cuvs_runtime/cluster/update_centroids_double.cu
+  src/cuvs_runtime/cluster/update_centroids_float.cu
+  src/cuvs_runtime/distance/fused_l2_min_arg.cu
+  src/cuvs_runtime/distance/pairwise_distance.cu
+  src/cuvs_runtime/matrix/select_k_float_int64_t.cu
+  src/cuvs_runtime/neighbors/brute_force_knn_int64_t_float.cu
+  src/cuvs_runtime/neighbors/cagra_build.cu
+  src/cuvs_runtime/neighbors/cagra_search.cu
+  src/cuvs_runtime/neighbors/cagra_serialize.cu
+  src/cuvs_runtime/neighbors/ivf_flat_build.cu
+  src/cuvs_runtime/neighbors/ivf_flat_search.cu
+  src/cuvs_runtime/neighbors/ivf_flat_serialize.cu
+  src/cuvs_runtime/neighbors/ivfpq_build.cu
+  src/cuvs_runtime/neighbors/ivfpq_deserialize.cu
+  src/cuvs_runtime/neighbors/ivfpq_search_float_int64_t.cu
+  src/cuvs_runtime/neighbors/ivfpq_search_int8_t_int64_t.cu
+  src/cuvs_runtime/neighbors/ivfpq_search_uint8_t_int64_t.cu
+  src/cuvs_runtime/neighbors/ivfpq_serialize.cu
+  src/cuvs_runtime/neighbors/refine_d_int64_t_float.cu
+  src/cuvs_runtime/neighbors/refine_d_int64_t_int8_t.cu
+  src/cuvs_runtime/neighbors/refine_d_int64_t_uint8_t.cu
+  src/cuvs_runtime/neighbors/refine_h_int64_t_float.cu
+  src/cuvs_runtime/neighbors/refine_h_int64_t_int8_t.cu
+  src/cuvs_runtime/neighbors/refine_h_int64_t_uint8_t.cu
+  src/cuvs_runtime/random/rmat_rectangular_generator_int64_double.cu
+  src/cuvs_runtime/random/rmat_rectangular_generator_int64_float.cu
+  src/cuvs_runtime/random/rmat_rectangular_generator_int_double.cu
+  src/cuvs_runtime/random/rmat_rectangular_generator_int_float.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_2d_dist.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_2d_euclidean.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_2d_haversine.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_3d_dist.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_3d_euclidean.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_one_3d_haversine.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_2d_dist.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_2d_euclidean.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_2d_haversine.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_3d_dist.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_3d_euclidean.cu
+  src/spatial/knn/detail/ball_cover/registers_pass_two_3d_haversine.cu
+  src/spatial/knn/detail/fused_l2_knn_int32_t_float.cu
+  src/spatial/knn/detail/fused_l2_knn_int64_t_float.cu
+  src/spatial/knn/detail/fused_l2_knn_uint32_t_float.cu
+)
+
+add_library(cuvs::cuvs ALIAS cuvs)
 
 target_include_directories(
-  raft INTERFACE "$<BUILD_INTERFACE:${RAFT_SOURCE_DIR}/include>" "$<INSTALL_INTERFACE:include>"
+  cuvs PUBLIC "$<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>"
+              "$<INSTALL_INTERFACE:include>"
 )
 
 if(NOT BUILD_CPU_ONLY)
-  # Keep RAFT as lightweight as possible. Only CUDA libs and rmm should be used in global target.
-  target_link_libraries(raft INTERFACE rmm::rmm cuco::cuco nvidia::cutlass::cutlass raft::Thrust)
+  # Keep cuVS as lightweight as possible. Only CUDA libs and rmm should be used in global target.
+  target_link_libraries(cuvs PUBLIC raft::raft)
 endif()
 
-target_compile_features(raft INTERFACE cxx_std_17 $<BUILD_INTERFACE:cuda_std_17>)
-target_compile_options(
-  raft INTERFACE $<$<COMPILE_LANG_AND_ID:CUDA,NVIDIA>:--expt-extended-lambda
-                 --expt-relaxed-constexpr>
-)
-
-set(RAFT_CUSOLVER_DEPENDENCY CUDA::cusolver${_ctk_static_suffix})
-set(RAFT_CUBLAS_DEPENDENCY CUDA::cublas${_ctk_static_suffix})
-set(RAFT_CURAND_DEPENDENCY CUDA::curand${_ctk_static_suffix})
-set(RAFT_CUSPARSE_DEPENDENCY CUDA::cusparse${_ctk_static_suffix})
-
-set(RAFT_CTK_MATH_DEPENDENCIES ${RAFT_CUBLAS_DEPENDENCY} ${RAFT_CUSOLVER_DEPENDENCY}
-                               ${RAFT_CUSPARSE_DEPENDENCY} ${RAFT_CURAND_DEPENDENCY}
-)
-
 # Endian detection
 include(TestBigEndian)
 test_big_endian(BIG_ENDIAN)
 if(BIG_ENDIAN)
-  target_compile_definitions(raft INTERFACE RAFT_SYSTEM_LITTLE_ENDIAN=0)
+  target_compile_definitions(cuvs PRIVATE CUVS_SYSTEM_LITTLE_ENDIAN=0)
 else()
-  target_compile_definitions(raft INTERFACE RAFT_SYSTEM_LITTLE_ENDIAN=1)
+  target_compile_definitions(cuvs PRIVATE CUVS_SYSTEM_LITTLE_ENDIAN=1)
 endif()
 
-if(RAFT_COMPILE_LIBRARY)
-  file(
-    WRITE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld"
-    [=[
+file(
+  WRITE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld"
+  [=[
 SECTIONS
 {
 .nvFatBinSegment : { *(.nvFatBinSegment) }
 .nv_fatbin : { *(.nv_fatbin) }
 }
 ]=]
-  )
-endif()
+)
 
 # ##################################################################################################
-# * NVTX support in raft -----------------------------------------------------
+# * NVTX support in cuvs -----------------------------------------------------
 
-if(RAFT_NVTX)
+if(CUVS_NVTX)
   # This enables NVTX within the project with no option to disable it downstream.
-  target_link_libraries(raft INTERFACE CUDA::nvToolsExt)
-  target_compile_definitions(raft INTERFACE NVTX_ENABLED)
+  target_link_libraries(cuvs PUBLIC CUDA::nvToolsExt)
+  target_compile_definitions(cuvs PUBLIC NVTX_ENABLED)
 else()
   # Allow enable NVTX downstream if not set here. This creates a new option at build/install time,
   # which is set by default to OFF, but can be enabled in the dependent project.
   get_property(
     nvtx_option_help_string
-    CACHE RAFT_NVTX
+    CACHE CUVS_NVTX
     PROPERTY HELPSTRING
   )
   string(
     CONCAT
       nvtx_export_string
-      "option(RAFT_NVTX \""
+      "option(CUVS_NVTX \""
       ${nvtx_option_help_string}
       "\" OFF)"
       [=[
 
-target_link_libraries(raft::raft INTERFACE $<$<BOOL:${RAFT_NVTX}>:CUDA::nvToolsExt>)
-target_compile_definitions(raft::raft INTERFACE $<$<BOOL:${RAFT_NVTX}>:NVTX_ENABLED>)
+target_link_libraries(cuvs::cuvs INTERFACE $<$<BOOL:${CUVS_NVTX}>:CUDA::nvToolsExt>)
+target_compile_definitions(cuvs::cuvs INTERFACE $<$<BOOL:${CUVS_NVTX}>:NVTX_ENABLED>)
 
   ]=]
   )
 endif()
 
-# ##################################################################################################
-# * raft_compiled ------------------------------------------------------------
-add_library(raft_compiled INTERFACE)
-
-if(TARGET raft_compiled AND (NOT TARGET raft::compiled))
-  add_library(raft::compiled ALIAS raft_compiled)
-endif()
-
-set_target_properties(raft_compiled PROPERTIES EXPORT_NAME compiled)
-
-if(RAFT_COMPILE_LIBRARY)
-  add_library(
-    raft_objs OBJECT
-    src/core/logger.cpp
-    src/distance/detail/pairwise_matrix/dispatch_canberra_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_canberra_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_correlation_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_correlation_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_cosine_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_cosine_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_hamming_unexpanded_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_hellinger_expanded_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_jensen_shannon_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_kl_divergence_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_kl_divergence_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l1_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l1_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l2_expanded_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l2_expanded_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l2_unexpanded_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l_inf_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_l_inf_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_lp_unexpanded_float_float_float_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_rbf.cu
-    src/distance/detail/pairwise_matrix/dispatch_russel_rao_double_double_double_int.cu
-    src/distance/detail/pairwise_matrix/dispatch_russel_rao_float_float_float_int.cu
-    src/distance/distance.cu
-    src/distance/fused_l2_nn.cu
-    src/linalg/detail/coalesced_reduction.cu
-    src/matrix/detail/select_k_double_int64_t.cu
-    src/matrix/detail/select_k_double_uint32_t.cu
-    src/matrix/detail/select_k_float_int64_t.cu
-    src/matrix/detail/select_k_float_uint32_t.cu
-    src/matrix/detail/select_k_float_int32.cu
-    src/matrix/detail/select_k_half_int64_t.cu
-    src/matrix/detail/select_k_half_uint32_t.cu
-    src/neighbors/ball_cover.cu
-    src/neighbors/brute_force_fused_l2_knn_float_int64_t.cu
-    src/neighbors/brute_force_knn_int64_t_float_int64_t.cu
-    src/neighbors/brute_force_knn_int64_t_float_uint32_t.cu
-    src/neighbors/brute_force_knn_int_float_int.cu
-    src/neighbors/brute_force_knn_uint32_t_float_uint32_t.cu
-    src/neighbors/brute_force_knn_index_float.cu
-    src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_multi_cta_float_uint32_dim1024_t32.cu
-    src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_multi_cta_int8_uint32_dim1024_t32.cu
-    src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_multi_cta_uint8_uint32_dim1024_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_float_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_single_cta_float_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_single_cta_float_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_float_uint32_dim1024_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_int8_uint32_dim1024_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim128_t8.cu
-    src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim256_t16.cu
-    src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim512_t32.cu
-    src/neighbors/detail/cagra/search_single_cta_uint8_uint32_dim1024_t32.cu
-    src/neighbors/detail/ivf_flat_interleaved_scan_float_float_int64_t.cu
-    src/neighbors/detail/ivf_flat_interleaved_scan_int8_t_int32_t_int64_t.cu
-    src/neighbors/detail/ivf_flat_interleaved_scan_uint8_t_uint32_t_int64_t.cu
-    src/neighbors/detail/ivf_flat_search.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_float_float.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_false.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_float_fp8_true.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_float_half.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_false.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_half_fp8_true.cu
-    src/neighbors/detail/ivf_pq_compute_similarity_half_half.cu
-    src/neighbors/detail/refine_host_float_float.cpp
-    src/neighbors/detail/refine_host_int8_t_float.cpp
-    src/neighbors/detail/refine_host_uint8_t_float.cpp
-    src/neighbors/detail/selection_faiss_int32_t_float.cu
-    src/neighbors/detail/selection_faiss_int_double.cu
-    src/neighbors/detail/selection_faiss_long_float.cu
-    src/neighbors/detail/selection_faiss_size_t_double.cu
-    src/neighbors/detail/selection_faiss_size_t_float.cu
-    src/neighbors/detail/selection_faiss_uint32_t_float.cu
-    src/neighbors/detail/selection_faiss_int64_t_double.cu
-    src/neighbors/detail/selection_faiss_int64_t_half.cu
-    src/neighbors/detail/selection_faiss_uint32_t_double.cu
-    src/neighbors/detail/selection_faiss_uint32_t_half.cu
-    src/neighbors/ivf_flat_build_float_int64_t.cu
-    src/neighbors/ivf_flat_build_int8_t_int64_t.cu
-    src/neighbors/ivf_flat_build_uint8_t_int64_t.cu
-    src/neighbors/ivf_flat_extend_float_int64_t.cu
-    src/neighbors/ivf_flat_extend_int8_t_int64_t.cu
-    src/neighbors/ivf_flat_extend_uint8_t_int64_t.cu
-    src/neighbors/ivf_flat_search_float_int64_t.cu
-    src/neighbors/ivf_flat_search_int8_t_int64_t.cu
-    src/neighbors/ivf_flat_search_uint8_t_int64_t.cu
-    src/neighbors/ivfpq_build_float_int64_t.cu
-    src/neighbors/ivfpq_build_int8_t_int64_t.cu
-    src/neighbors/ivfpq_build_uint8_t_int64_t.cu
-    src/neighbors/ivfpq_extend_float_int64_t.cu
-    src/neighbors/ivfpq_extend_int8_t_int64_t.cu
-    src/neighbors/ivfpq_extend_uint8_t_int64_t.cu
-    src/neighbors/ivfpq_search_float_int64_t.cu
-    src/neighbors/ivfpq_search_int8_t_int64_t.cu
-    src/neighbors/ivfpq_search_uint8_t_int64_t.cu
-    src/neighbors/refine_float_float.cu
-    src/neighbors/refine_int8_t_float.cu
-    src/neighbors/refine_uint8_t_float.cu
-    src/raft_runtime/cluster/cluster_cost.cuh
-    src/raft_runtime/cluster/cluster_cost_double.cu
-    src/raft_runtime/cluster/cluster_cost_float.cu
-    src/raft_runtime/cluster/kmeans_fit_double.cu
-    src/raft_runtime/cluster/kmeans_fit_float.cu
-    src/raft_runtime/cluster/kmeans_init_plus_plus_double.cu
-    src/raft_runtime/cluster/kmeans_init_plus_plus_float.cu
-    src/raft_runtime/cluster/update_centroids.cuh
-    src/raft_runtime/cluster/update_centroids_double.cu
-    src/raft_runtime/cluster/update_centroids_float.cu
-    src/raft_runtime/distance/fused_l2_min_arg.cu
-    src/raft_runtime/distance/pairwise_distance.cu
-    src/raft_runtime/matrix/select_k_float_int64_t.cu
-    src/raft_runtime/neighbors/brute_force_knn_int64_t_float.cu
-    src/raft_runtime/neighbors/cagra_build.cu
-    src/raft_runtime/neighbors/cagra_search.cu
-    src/raft_runtime/neighbors/cagra_serialize.cu
-    src/raft_runtime/neighbors/ivf_flat_build.cu
-    src/raft_runtime/neighbors/ivf_flat_search.cu
-    src/raft_runtime/neighbors/ivf_flat_serialize.cu
-    src/raft_runtime/neighbors/ivfpq_build.cu
-    src/raft_runtime/neighbors/ivfpq_deserialize.cu
-    src/raft_runtime/neighbors/ivfpq_search_float_int64_t.cu
-    src/raft_runtime/neighbors/ivfpq_search_int8_t_int64_t.cu
-    src/raft_runtime/neighbors/ivfpq_search_uint8_t_int64_t.cu
-    src/raft_runtime/neighbors/ivfpq_serialize.cu
-    src/raft_runtime/neighbors/refine_d_int64_t_float.cu
-    src/raft_runtime/neighbors/refine_d_int64_t_int8_t.cu
-    src/raft_runtime/neighbors/refine_d_int64_t_uint8_t.cu
-    src/raft_runtime/neighbors/refine_h_int64_t_float.cu
-    src/raft_runtime/neighbors/refine_h_int64_t_int8_t.cu
-    src/raft_runtime/neighbors/refine_h_int64_t_uint8_t.cu
-    src/raft_runtime/random/rmat_rectangular_generator_int64_double.cu
-    src/raft_runtime/random/rmat_rectangular_generator_int64_float.cu
-    src/raft_runtime/random/rmat_rectangular_generator_int_double.cu
-    src/raft_runtime/random/rmat_rectangular_generator_int_float.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_2d_dist.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_2d_euclidean.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_2d_haversine.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_3d_dist.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_3d_euclidean.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_one_3d_haversine.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_2d_dist.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_2d_euclidean.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_2d_haversine.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_3d_dist.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_3d_euclidean.cu
-    src/spatial/knn/detail/ball_cover/registers_pass_two_3d_haversine.cu
-    src/spatial/knn/detail/fused_l2_knn_int32_t_float.cu
-    src/spatial/knn/detail/fused_l2_knn_int64_t_float.cu
-    src/spatial/knn/detail/fused_l2_knn_uint32_t_float.cu
-    src/util/memory_pool.cpp
-  )
-  set_target_properties(
-    raft_objs
-    PROPERTIES CXX_STANDARD 17
-               CXX_STANDARD_REQUIRED ON
-               CUDA_STANDARD 17
-               CUDA_STANDARD_REQUIRED ON
-               POSITION_INDEPENDENT_CODE ON
-  )
-
-  target_compile_definitions(raft_objs PRIVATE "RAFT_EXPLICIT_INSTANTIATE_ONLY")
-  target_compile_options(
-    raft_objs PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${RAFT_CXX_FLAGS}>"
-                      "$<$<COMPILE_LANGUAGE:CUDA>:${RAFT_CUDA_FLAGS}>"
-  )
-
-  add_library(raft_lib SHARED $<TARGET_OBJECTS:raft_objs>)
-  add_library(raft_lib_static STATIC $<TARGET_OBJECTS:raft_objs>)
-
-  set_target_properties(
-    raft_lib raft_lib_static
-    PROPERTIES OUTPUT_NAME raft
-               BUILD_RPATH "\$ORIGIN"
-               INSTALL_RPATH "\$ORIGIN"
-               INTERFACE_POSITION_INDEPENDENT_CODE ON
-  )
-
-  foreach(target raft_lib raft_lib_static raft_objs)
-    target_link_libraries(
-      ${target}
-      PUBLIC raft::raft
-             ${RAFT_CTK_MATH_DEPENDENCIES} # TODO: Once `raft::resources` is used everywhere, this
-                                           # will just be cublas
-             $<TARGET_NAME_IF_EXISTS:OpenMP::OpenMP_CXX>
-    )
-
-    # So consumers know when using libraft.so/libraft.a
-    target_compile_definitions(${target} PUBLIC "RAFT_COMPILED")
-    # ensure CUDA symbols aren't relocated to the middle of the debug build binaries
-    target_link_options(${target} PRIVATE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld")
-  endforeach()
-endif()
-
-if(TARGET raft_lib AND (NOT TARGET raft::raft_lib))
-  add_library(raft::raft_lib ALIAS raft_lib)
-endif()
-
-target_link_libraries(raft_compiled INTERFACE raft::raft $<TARGET_NAME_IF_EXISTS:raft::raft_lib>)
-
-# ##################################################################################################
-# * raft_compiled_static----------------------------------------------------------------------------
-
-add_library(raft_compiled_static INTERFACE)
-
-if(TARGET raft_compiled_static AND (NOT TARGET raft::compiled_static))
-  add_library(raft::compiled_static ALIAS raft_compiled_static)
-endif()
-set_target_properties(raft_compiled_static PROPERTIES EXPORT_NAME compiled_static)
-
-if(TARGET raft_lib_static AND (NOT TARGET raft::raft_lib_static))
-  add_library(raft::raft_lib_static ALIAS raft_lib_static)
-endif()
-
-target_link_libraries(
-  raft_compiled_static INTERFACE raft::raft $<TARGET_NAME_IF_EXISTS:raft::raft_lib_static>
+set_target_properties(
+  cuvs
+  PROPERTIES CXX_STANDARD 17
+             CXX_STANDARD_REQUIRED ON
+             CUDA_STANDARD 17
+             CUDA_STANDARD_REQUIRED ON
+             POSITION_INDEPENDENT_CODE ON
 )
 
-# ##################################################################################################
-# * raft_distributed -------------------------------------------------------------------------------
-add_library(raft_distributed INTERFACE)
-
-if(TARGET raft_distributed AND (NOT TARGET raft::distributed))
-  add_library(raft::distributed ALIAS raft_distributed)
-endif()
-
-set_target_properties(raft_distributed PROPERTIES EXPORT_NAME distributed)
-
-rapids_find_generate_module(
-  NCCL
-  HEADER_NAMES nccl.h
-  LIBRARY_NAMES nccl
-  BUILD_EXPORT_SET raft-distributed-exports
-  INSTALL_EXPORT_SET raft-distributed-exports
+target_compile_options(
+  cuvs PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${CUVS_CXX_FLAGS}>"
+               "$<$<COMPILE_LANGUAGE:CUDA>:${CUVS_CUDA_FLAGS}>"
 )
-
-rapids_export_package(BUILD ucx raft-distributed-exports)
-rapids_export_package(INSTALL ucx raft-distributed-exports)
-rapids_export_package(BUILD NCCL raft-distributed-exports)
-rapids_export_package(INSTALL NCCL raft-distributed-exports)
-
-target_link_libraries(raft_distributed INTERFACE ucx::ucp NCCL::NCCL)
+# ensure CUDA symbols aren't relocated to the middle of the debug build binaries
+target_link_options(cuvs PRIVATE "${CMAKE_CURRENT_BINARY_DIR}/fatbin.ld")
 
 # ##################################################################################################
 # * install targets-----------------------------------------------------------
@@ -554,168 +439,45 @@ include(GNUInstallDirs)
 include(CPack)
 
 install(
-  TARGETS raft
+  TARGETS cuvs
   DESTINATION ${lib_dir}
-  COMPONENT raft
-  EXPORT raft-exports
+  COMPONENT cuvs
+  EXPORT cuvs-exports
 )
 
 install(
-  TARGETS raft_compiled raft_compiled_static
-  DESTINATION ${lib_dir}
-  COMPONENT raft
-  EXPORT raft-compiled-exports
-)
-
-if(TARGET raft_lib)
-  install(
-    TARGETS raft_lib
-    DESTINATION ${lib_dir}
-    COMPONENT compiled
-    EXPORT raft-compiled-lib-exports
-  )
-  install(
-    TARGETS raft_lib_static
-    DESTINATION ${lib_dir}
-    COMPONENT compiled-static
-    EXPORT raft-compiled-static-lib-exports
-  )
-  install(
-    DIRECTORY include/raft_runtime
-    DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
-    COMPONENT compiled
-  )
-endif()
-
-install(
-  TARGETS raft_distributed
-  DESTINATION ${lib_dir}
-  COMPONENT distributed
-  EXPORT raft-distributed-exports
-)
-
-install(
-  DIRECTORY include/raft
-  COMPONENT raft
+  DIRECTORY include/cuvs
+  COMPONENT cuvs
   DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
 )
 
-# Temporary install of raft.hpp while the file is removed
 install(
-  FILES include/raft.hpp
-  COMPONENT raft
-  DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/raft
-)
-
-install(
-  FILES ${CMAKE_CURRENT_BINARY_DIR}/include/raft/version_config.hpp
-  COMPONENT raft
-  DESTINATION include/raft
-)
-
-# ##################################################################################################
-# * install export -----------------------------------------------------------
-set(doc_string
-    [=[
-Provide targets for the RAFT: Reusable Accelerated Functions and Tools
-
-RAFT contains fundamental widely-used algorithms and primitives
-for data science and machine learning.
-
-Optional Components:
-  - compiled
-  - compiled_static
-  - distributed
-
-Imported Targets:
-  - raft::raft
-  - raft::compiled brought in by the `compiled` optional component
-  - raft::compiled_static brought in by the `compiled_static` optional component
-  - raft::distributed brought in by the `distributed` optional component
-
-]=]
-)
-
-set(code_string ${nvtx_export_string})
-
-string(
-  APPEND
-  code_string
-  [=[
-if(NOT TARGET raft::Thrust)
-  thrust_create_target(raft::Thrust FROM_OPTIONS)
-endif()
-]=]
-)
-
-string(
-  APPEND
-  code_string
-  [=[
-if(compiled IN_LIST raft_FIND_COMPONENTS)
-  enable_language(CUDA)
-endif()
-]=]
-)
-set(raft_components compiled distributed)
-set(raft_export_sets raft-compiled-exports raft-distributed-exports)
-if(TARGET raft_lib)
-  list(APPEND raft_components compiled compiled-static)
-  list(APPEND raft_export_sets raft-compiled-lib-exports raft-compiled-static-lib-exports)
-endif()
-
-string(
-  APPEND
-  code_string
-  [=[
- option(RAFT_ENABLE_CUSOLVER_DEPENDENCY "Enable cusolver dependency" ON)
- option(RAFT_ENABLE_CUBLAS_DEPENDENCY "Enable cublas dependency" ON)
- option(RAFT_ENABLE_CURAND_DEPENDENCY "Enable curand dependency" ON)
- option(RAFT_ENABLE_CUSPARSE_DEPENDENCY "Enable cusparse dependency" ON)
-
-mark_as_advanced(RAFT_ENABLE_CUSOLVER_DEPENDENCY)
-mark_as_advanced(RAFT_ENABLE_CUBLAS_DEPENDENCY)
-mark_as_advanced(RAFT_ENABLE_CURAND_DEPENDENCY)
-mark_as_advanced(RAFT_ENABLE_CUSPARSE_DEPENDENCY)
-
-target_link_libraries(raft::raft INTERFACE
-  $<$<BOOL:${RAFT_ENABLE_CUSOLVER_DEPENDENCY}>:${RAFT_CUSOLVER_DEPENDENCY}>
-  $<$<BOOL:${RAFT_ENABLE_CUBLAS_DEPENDENCY}>:${RAFT_CUBLAS_DEPENDENCY}>
-  $<$<BOOL:${RAFT_ENABLE_CUSPARSE_DEPENDENCY}>:${RAFT_CUSPARSE_DEPENDENCY}>
-  $<$<BOOL:${RAFT_ENABLE_CURAND_DEPENDENCY}>:${RAFT_CURAND_DEPENDENCY}>
-)
-]=]
+  FILES ${CMAKE_CURRENT_BINARY_DIR}/include/cuvs/version_config.hpp
+  COMPONENT cuvs
+  DESTINATION include/cuvs
 )
 
 # Use `rapids_export` for 22.04 as it will have COMPONENT support
 rapids_export(
-  INSTALL raft
-  EXPORT_SET raft-exports
-  COMPONENTS ${raft_components}
-  COMPONENTS_EXPORT_SET ${raft_export_sets}
-  GLOBAL_TARGETS raft compiled distributed
-  NAMESPACE raft::
-  DOCUMENTATION doc_string
-  FINAL_CODE_BLOCK code_string
+  INSTALL cuvs
+  EXPORT_SET cuvs-exports
+  GLOBAL_TARGETS cuvs
+  NAMESPACE cuvs::
 )
 
 # ##################################################################################################
 # * build export -------------------------------------------------------------
 rapids_export(
-  BUILD raft
-  EXPORT_SET raft-exports
-  COMPONENTS ${raft_components}
-  COMPONENTS_EXPORT_SET ${raft_export_sets}
-  GLOBAL_TARGETS raft compiled distributed
-  DOCUMENTATION doc_string
-  NAMESPACE raft::
-  FINAL_CODE_BLOCK code_string
+  BUILD cuvs
+  EXPORT_SET cuvs-exports
+  GLOBAL_TARGETS cuvs
+  NAMESPACE cuvs::
 )
 
 # ##################################################################################################
 # * shared test/bench headers ------------------------------------------------
 
-if(BUILD_TESTS OR BUILD_PRIMS_BENCH)
+if(BUILD_TESTS OR BUILD_MICRO_BENCH)
   include(internal/CMakeLists.txt)
 endif()
 
@@ -729,8 +491,8 @@ endif()
 # ##################################################################################################
 # * build benchmark executable -----------------------------------------------
 
-if(BUILD_PRIMS_BENCH)
-  include(bench/prims/CMakeLists.txt)
+if(BUILD_MICRO_BENCH)
+  include(bench/micro/CMakeLists.txt)
 endif()
 
 # ##################################################################################################
diff --git a/cpp/bench/ann/CMakeLists.txt b/cpp/bench/ann/CMakeLists.txt
index 5919de07e..d1f4d281d 100644
--- a/cpp/bench/ann/CMakeLists.txt
+++ b/cpp/bench/ann/CMakeLists.txt
@@ -15,26 +15,22 @@
 # ##################################################################################################
 # * benchmark options ------------------------------------------------------------------------------
 
-option(RAFT_ANN_BENCH_USE_FAISS_GPU_FLAT "Include faiss' brute-force knn algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_FLAT "Include faiss' ivf flat algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_PQ "Include faiss' ivf pq algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_FAISS_CPU_FLAT
-       "Include faiss' cpu brute-force knn algorithm in benchmark" ON
-)
-option(RAFT_ANN_BENCH_USE_FAISS_CPU_FLAT "Include faiss' cpu brute-force algorithm in benchmark" ON)
-
-option(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_FLAT "Include faiss' cpu ivf flat algorithm in benchmark"
-       ON
-)
-option(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_PQ "Include faiss' cpu ivf pq algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT "Include raft's ivf flat algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ "Include raft's ivf pq algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_RAFT_CAGRA "Include raft's CAGRA in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB "Include raft's CAGRA in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_HNSWLIB "Include hnsw algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_USE_GGNN "Include ggnn algorithm in benchmark" ON)
-option(RAFT_ANN_BENCH_SINGLE_EXE
-       "Make a single executable with benchmark as shared library modules" OFF
+option(CUVS_BENCH_USE_FAISS_GPU_FLAT "Include faiss' brute-force knn algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_FAISS_GPU_IVF_FLAT "Include faiss' ivf flat algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_FAISS_GPU_IVF_PQ "Include faiss' ivf pq algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_FAISS_CPU_FLAT "Include faiss' cpu brute-force knn algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_FAISS_CPU_FLAT "Include faiss' cpu brute-force algorithm in benchmark" ON)
+
+option(CUVS_BENCH_USE_FAISS_CPU_IVF_FLAT "Include faiss' cpu ivf flat algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_FAISS_CPU_IVF_PQ "Include faiss' cpu ivf pq algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_RAFT_IVF_FLAT "Include raft's ivf flat algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_RAFT_IVF_PQ "Include raft's ivf pq algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_RAFT_CAGRA "Include raft's CAGRA in benchmark" ON)
+option(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB "Include raft's CAGRA in benchmark" ON)
+option(CUVS_BENCH_USE_HNSWLIB "Include hnsw algorithm in benchmark" ON)
+option(CUVS_BENCH_USE_GGNN "Include ggnn algorithm in benchmark" ON)
+option(CUVS_BENCH_SINGLE_EXE "Make a single executable with benchmark as shared library modules"
+       OFF
 )
 
 # ##################################################################################################
@@ -48,67 +44,67 @@ if(BUILD_CPU_ONLY)
   include(cmake/thirdparty/get_fmt.cmake)
   include(cmake/thirdparty/get_spdlog.cmake)
 
-  set(RAFT_FAISS_ENABLE_GPU OFF)
-  set(RAFT_ANN_BENCH_USE_FAISS_GPU_FLAT OFF)
-  set(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_FLAT OFF)
-  set(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_PQ OFF)
-  set(RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT OFF)
-  set(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ OFF)
-  set(RAFT_ANN_BENCH_USE_RAFT_CAGRA OFF)
-  set(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB OFF)
-  set(RAFT_ANN_BENCH_USE_GGNN OFF)
+  set(CUVS_FAISS_ENABLE_GPU OFF)
+  set(CUVS_BENCH_USE_FAISS_GPU_FLAT OFF)
+  set(CUVS_BENCH_USE_FAISS_GPU_IVF_FLAT OFF)
+  set(CUVS_BENCH_USE_FAISS_GPU_IVF_PQ OFF)
+  set(CUVS_BENCH_USE_RAFT_IVF_FLAT OFF)
+  set(CUVS_BENCH_USE_RAFT_IVF_PQ OFF)
+  set(CUVS_BENCH_USE_RAFT_CAGRA OFF)
+  set(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB OFF)
+  set(CUVS_BENCH_USE_GGNN OFF)
 else()
   # Disable faiss benchmarks on CUDA 12 since faiss is not yet CUDA 12-enabled.
   # https://github.com/rapidsai/raft/issues/1627
   if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.0.0)
-    set(RAFT_FAISS_ENABLE_GPU OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_GPU_FLAT OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_FLAT OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_PQ OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_CPU_FLAT OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_PQ OFF)
-    set(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_FLAT OFF)
+    set(CUVS_FAISS_ENABLE_GPU OFF)
+    set(CUVS_BENCH_USE_FAISS_GPU_FLAT OFF)
+    set(CUVS_BENCH_USE_FAISS_GPU_IVF_FLAT OFF)
+    set(CUVS_BENCH_USE_FAISS_GPU_IVF_PQ OFF)
+    set(CUVS_BENCH_USE_FAISS_CPU_FLAT OFF)
+    set(CUVS_BENCH_USE_FAISS_CPU_IVF_PQ OFF)
+    set(CUVS_BENCH_USE_FAISS_CPU_IVF_FLAT OFF)
   else()
-    set(RAFT_FAISS_ENABLE_GPU ON)
+    set(CUVS_FAISS_ENABLE_GPU ON)
   endif()
 endif()
 
-set(RAFT_ANN_BENCH_USE_FAISS OFF)
-if(RAFT_ANN_BENCH_USE_FAISS_GPU_FLAT
-   OR RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_PQ
-   OR RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_FLAT
-   OR RAFT_ANN_BENCH_USE_FAISS_CPU_FLAT
-   OR RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_PQ
-   OR RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_FLAT
+set(CUVS_BENCH_USE_FAISS OFF)
+if(CUVS_BENCH_USE_FAISS_GPU_FLAT
+   OR CUVS_BENCH_USE_FAISS_GPU_IVF_PQ
+   OR CUVS_BENCH_USE_FAISS_GPU_IVF_FLAT
+   OR CUVS_BENCH_USE_FAISS_CPU_FLAT
+   OR CUVS_BENCH_USE_FAISS_CPU_IVF_PQ
+   OR CUVS_BENCH_USE_FAISS_CPU_IVF_FLAT
 )
-  set(RAFT_ANN_BENCH_USE_FAISS ON)
+  set(CUVS_BENCH_USE_FAISS ON)
   set(RAFT_USE_FAISS_STATIC ON)
 endif()
 
-set(RAFT_ANN_BENCH_USE_RAFT OFF)
-if(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ
-   OR RAFT_ANN_BENCH_USE_RAFT_BRUTE_FORCE
-   OR RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT
-   OR RAFT_ANN_BENCH_USE_RAFT_CAGRA
-   OR RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB
+set(CUVS_BENCH_USE_RAFT OFF)
+if(CUVS_BENCH_USE_RAFT_IVF_PQ
+   OR CUVS_BENCH_USE_RAFT_BRUTE_FORCE
+   OR CUVS_BENCH_USE_RAFT_IVF_FLAT
+   OR CUVS_BENCH_USE_RAFT_CAGRA
+   OR CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB
 )
-  set(RAFT_ANN_BENCH_USE_RAFT ON)
+  set(CUVS_BENCH_USE_RAFT ON)
 endif()
 
 # ##################################################################################################
 # * Fetch requirements -------------------------------------------------------------
 
-if(RAFT_ANN_BENCH_USE_HNSWLIB OR RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+if(CUVS_BENCH_USE_HNSWLIB OR CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
   include(cmake/thirdparty/get_hnswlib.cmake)
 endif()
 
 include(cmake/thirdparty/get_nlohmann_json.cmake)
 
-if(RAFT_ANN_BENCH_USE_GGNN)
+if(CUVS_BENCH_USE_GGNN)
   include(cmake/thirdparty/get_ggnn.cmake)
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS)
+if(CUVS_BENCH_USE_FAISS)
   # We need to ensure that faiss has all the conda information. So we currently use the very ugly
   # hammer of `link_libraries` to ensure that all targets in this directory and the faiss directory
   # will have the conda includes/link dirs
@@ -132,16 +128,16 @@ function(ConfigureAnnBench)
     ConfigureAnnBench "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN}
   )
 
-  set(BENCH_NAME ${ConfigureAnnBench_NAME}_ANN_BENCH)
+  set(BENCH_NAME ${ConfigureAnnBench_NAME}_CUVS_BENCH)
 
-  if(RAFT_ANN_BENCH_SINGLE_EXE)
+  if(CUVS_BENCH_SINGLE_EXE)
     add_library(${BENCH_NAME} SHARED ${ConfigureAnnBench_PATH})
     string(TOLOWER ${BENCH_NAME} BENCH_LIB_NAME)
     set_target_properties(${BENCH_NAME} PROPERTIES OUTPUT_NAME ${BENCH_LIB_NAME})
-    add_dependencies(${BENCH_NAME} ANN_BENCH)
+    add_dependencies(${BENCH_NAME} CUVS_BENCH)
   else()
     add_executable(${BENCH_NAME} ${ConfigureAnnBench_PATH})
-    target_compile_definitions(${BENCH_NAME} PRIVATE ANN_BENCH_BUILD_MAIN)
+    target_compile_definitions(${BENCH_NAME} PRIVATE CUVS_BENCH_BUILD_MAIN)
     target_link_libraries(${BENCH_NAME} PRIVATE benchmark::benchmark)
   endif()
 
@@ -180,11 +176,10 @@ function(ConfigureAnnBench)
                           "$<$<COMPILE_LANGUAGE:CUDA>:${RAFT_CUDA_FLAGS}>"
   )
 
-  if(RAFT_ANN_BENCH_USE_${ConfigureAnnBench_NAME})
+  if(CUVS_BENCH_USE_${ConfigureAnnBench_NAME})
     target_compile_definitions(
       ${BENCH_NAME}
-      PUBLIC
-        RAFT_ANN_BENCH_USE_${ConfigureAnnBench_NAME}=RAFT_ANN_BENCH_USE_${ConfigureAnnBench_NAME}
+      PUBLIC CUVS_BENCH_USE_${ConfigureAnnBench_NAME}=CUVS_BENCH_USE_${ConfigureAnnBench_NAME}
     )
   endif()
 
@@ -204,56 +199,46 @@ endfunction()
 # ##################################################################################################
 # * Configure tests-------------------------------------------------------------
 
-if(RAFT_ANN_BENCH_USE_HNSWLIB)
+if(CUVS_BENCH_USE_HNSWLIB)
   ConfigureAnnBench(
     NAME HNSWLIB PATH bench/ann/src/hnswlib/hnswlib_benchmark.cpp INCLUDES
     ${CMAKE_CURRENT_BINARY_DIR}/_deps/hnswlib-src/hnswlib CXXFLAGS "${HNSW_CXX_FLAGS}"
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ)
+if(CUVS_BENCH_USE_RAFT_IVF_PQ)
   ConfigureAnnBench(
-    NAME
-    RAFT_IVF_PQ
-    PATH
-    bench/ann/src/raft/raft_benchmark.cu
-    $<$<BOOL:${RAFT_ANN_BENCH_USE_RAFT_IVF_PQ}>:bench/ann/src/raft/raft_ivf_pq.cu>
-    LINKS
-    raft::compiled
+    NAME RAFT_IVF_PQ PATH bench/ann/src/raft/raft_benchmark.cu
+    $<$<BOOL:${CUVS_BENCH_USE_RAFT_IVF_PQ}>:bench/ann/src/raft/raft_ivf_pq.cu> LINKS raft::compiled
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT)
+if(CUVS_BENCH_USE_RAFT_IVF_FLAT)
   ConfigureAnnBench(
     NAME
     RAFT_IVF_FLAT
     PATH
     bench/ann/src/raft/raft_benchmark.cu
-    $<$<BOOL:${RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT}>:bench/ann/src/raft/raft_ivf_flat.cu>
+    $<$<BOOL:${CUVS_BENCH_USE_RAFT_IVF_FLAT}>:bench/ann/src/raft/raft_ivf_flat.cu>
     LINKS
     raft::compiled
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_RAFT_BRUTE_FORCE)
+if(CUVS_BENCH_USE_RAFT_BRUTE_FORCE)
   ConfigureAnnBench(
     NAME RAFT_BRUTE_FORCE PATH bench/ann/src/raft/raft_benchmark.cu LINKS raft::compiled
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_RAFT_CAGRA)
+if(CUVS_BENCH_USE_RAFT_CAGRA)
   ConfigureAnnBench(
-    NAME
-    RAFT_CAGRA
-    PATH
-    bench/ann/src/raft/raft_benchmark.cu
-    $<$<BOOL:${RAFT_ANN_BENCH_USE_RAFT_CAGRA}>:bench/ann/src/raft/raft_cagra.cu>
-    LINKS
-    raft::compiled
+    NAME RAFT_CAGRA PATH bench/ann/src/raft/raft_benchmark.cu
+    $<$<BOOL:${CUVS_BENCH_USE_RAFT_CAGRA}>:bench/ann/src/raft/raft_cagra.cu> LINKS raft::compiled
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+if(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
   ConfigureAnnBench(
     NAME
     RAFT_CAGRA_HNSWLIB
@@ -275,48 +260,48 @@ endif()
 
 message("RAFT_FAISS_TARGETS: ${RAFT_FAISS_TARGETS}")
 message("CUDAToolkit_LIBRARY_DIR: ${CUDAToolkit_LIBRARY_DIR}")
-if(RAFT_ANN_BENCH_USE_FAISS_CPU_FLAT)
+if(CUVS_BENCH_USE_FAISS_CPU_FLAT)
   ConfigureAnnBench(
     NAME FAISS_CPU_FLAT PATH bench/ann/src/faiss/faiss_cpu_benchmark.cpp LINKS
     ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_FLAT)
+if(CUVS_BENCH_USE_FAISS_CPU_IVF_FLAT)
   ConfigureAnnBench(
     NAME FAISS_CPU_IVF_FLAT PATH bench/ann/src/faiss/faiss_cpu_benchmark.cpp LINKS
     ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS_CPU_IVF_PQ)
+if(CUVS_BENCH_USE_FAISS_CPU_IVF_PQ)
   ConfigureAnnBench(
     NAME FAISS_CPU_IVF_PQ PATH bench/ann/src/faiss/faiss_cpu_benchmark.cpp LINKS
     ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_FLAT)
+if(CUVS_BENCH_USE_FAISS_GPU_IVF_FLAT)
   ConfigureAnnBench(
     NAME FAISS_GPU_IVF_FLAT PATH bench/ann/src/faiss/faiss_gpu_benchmark.cu LINKS
     ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS_GPU_IVF_PQ)
+if(CUVS_BENCH_USE_FAISS_GPU_IVF_PQ)
   ConfigureAnnBench(
     NAME FAISS_GPU_IVF_PQ PATH bench/ann/src/faiss/faiss_gpu_benchmark.cu LINKS
     ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_FAISS_GPU_FLAT)
+if(CUVS_BENCH_USE_FAISS_GPU_FLAT)
   ConfigureAnnBench(
     NAME FAISS_GPU_FLAT PATH bench/ann/src/faiss/faiss_gpu_benchmark.cu LINKS ${RAFT_FAISS_TARGETS}
   )
 endif()
 
-if(RAFT_ANN_BENCH_USE_GGNN)
+if(CUVS_BENCH_USE_GGNN)
   include(cmake/thirdparty/get_glog.cmake)
   ConfigureAnnBench(
     NAME GGNN PATH bench/ann/src/ggnn/ggnn_benchmark.cu INCLUDES
@@ -325,11 +310,12 @@ if(RAFT_ANN_BENCH_USE_GGNN)
 endif()
 
 # ##################################################################################################
-# * Dynamically-loading ANN_BENCH executable -------------------------------------------------------
-if(RAFT_ANN_BENCH_SINGLE_EXE)
-  add_executable(ANN_BENCH bench/ann/src/common/benchmark.cpp)
+# * Dynamically-loading CUVS_BENCH executable
+# -------------------------------------------------------
+if(CUVS_BENCH_SINGLE_EXE)
+  add_executable(CUVS_BENCH bench/ann/src/common/benchmark.cpp)
 
-  # Build and link static version of the GBench to keep ANN_BENCH self-contained.
+  # Build and link static version of the GBench to keep CUVS_BENCH self-contained.
   get_target_property(TMP_PROP benchmark::benchmark SOURCES)
   add_library(benchmark_static STATIC ${TMP_PROP})
   get_target_property(TMP_PROP benchmark::benchmark INCLUDE_DIRECTORIES)
@@ -337,14 +323,14 @@ if(RAFT_ANN_BENCH_SINGLE_EXE)
   get_target_property(TMP_PROP benchmark::benchmark LINK_LIBRARIES)
   target_link_libraries(benchmark_static PUBLIC ${TMP_PROP})
 
-  target_include_directories(ANN_BENCH PRIVATE ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})
+  target_include_directories(CUVS_BENCH PRIVATE ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})
 
   target_link_libraries(
-    ANN_BENCH PRIVATE nlohmann_json::nlohmann_json benchmark_static dl -static-libgcc
-                      -static-libstdc++ CUDA::nvtx3
+    CUVS_BENCH PRIVATE nlohmann_json::nlohmann_json benchmark_static dl -static-libgcc
+                       -static-libstdc++ CUDA::nvtx3
   )
   set_target_properties(
-    ANN_BENCH
+    CUVS_BENCH
     PROPERTIES # set target compile options
                CXX_STANDARD 17
                CXX_STANDARD_REQUIRED ON
@@ -358,21 +344,21 @@ if(RAFT_ANN_BENCH_SINGLE_EXE)
 
   # Disable NVTX when the nvtx3 headers are missing
   set(_CMAKE_REQUIRED_INCLUDES_ORIG ${CMAKE_REQUIRED_INCLUDES})
-  get_target_property(CMAKE_REQUIRED_INCLUDES ANN_BENCH INCLUDE_DIRECTORIES)
+  get_target_property(CMAKE_REQUIRED_INCLUDES CUVS_BENCH INCLUDE_DIRECTORIES)
   CHECK_INCLUDE_FILE_CXX(nvtx3/nvToolsExt.h NVTX3_HEADERS_FOUND)
   set(CMAKE_REQUIRED_INCLUDES ${_CMAKE_REQUIRED_INCLUDES_ORIG})
   target_compile_definitions(
-    ANN_BENCH
+    CUVS_BENCH
     PRIVATE
-      $<$<BOOL:${CUDAToolkit_FOUND}>:ANN_BENCH_LINK_CUDART="libcudart.so.${CUDAToolkit_VERSION_MAJOR}.${CUDAToolkit_VERSION_MINOR}.${CUDAToolkit_VERSION_PATCH}
+      $<$<BOOL:${CUDAToolkit_FOUND}>:CUVS_BENCH_LINK_CUDART="libcudart.so.${CUDAToolkit_VERSION_MAJOR}.${CUDAToolkit_VERSION_MINOR}.${CUDAToolkit_VERSION_PATCH}
   ">
-      $<$<BOOL:${NVTX3_HEADERS_FOUND}>:ANN_BENCH_NVTX3_HEADERS_FOUND>
+      $<$<BOOL:${NVTX3_HEADERS_FOUND}>:CUVS_BENCH_NVTX3_HEADERS_FOUND>
   )
 
-  target_link_options(ANN_BENCH PRIVATE -export-dynamic)
+  target_link_options(CUVS_BENCH PRIVATE -export-dynamic)
 
   install(
-    TARGETS ANN_BENCH
+    TARGETS CUVS_BENCH
     COMPONENT ann_bench
     DESTINATION bin/ann
     EXCLUDE_FROM_ALL
diff --git a/cpp/bench/ann/src/common/cuda_stub.hpp b/cpp/bench/ann/src/common/cuda_stub.hpp
index 6e3b63cd3..d4d02d5f3 100644
--- a/cpp/bench/ann/src/common/cuda_stub.hpp
+++ b/cpp/bench/ann/src/common/cuda_stub.hpp
@@ -19,10 +19,10 @@
 The content of this header is governed by two preprocessor definitions:
 
   - BUILD_CPU_ONLY - whether none of the CUDA functions are used.
-  - ANN_BENCH_LINK_CUDART - dynamically link against this string if defined.
+  - CUVS_BENCH_LINK_CUDART - dynamically link against this string if defined.
 
 ______________________________________________________________________________
-|BUILD_CPU_ONLY | ANN_BENCH_LINK_CUDART |         cudart      | cuda_runtime_api.h |
+|BUILD_CPU_ONLY | CUVS_BENCH_LINK_CUDART |         cudart      | cuda_runtime_api.h |
 |         |                       |  found    |  needed |      included      |
 |---------|-----------------------|-----------|---------|--------------------|
 |   ON    |    <not defined>      |  false    |  false  |       NO           |
@@ -34,7 +34,7 @@ ______________________________________________________________________________
 
 #ifndef BUILD_CPU_ONLY
 #include <cuda_runtime_api.h>
-#ifdef ANN_BENCH_LINK_CUDART
+#ifdef CUVS_BENCH_LINK_CUDART
 #include <cstring>
 #include <dlfcn.h>
 #endif
@@ -49,11 +49,11 @@ struct cuda_lib_handle {
   void* handle{nullptr};
   explicit cuda_lib_handle()
   {
-#ifdef ANN_BENCH_LINK_CUDART
+#ifdef CUVS_BENCH_LINK_CUDART
     constexpr int kFlags = RTLD_NOW | RTLD_GLOBAL | RTLD_DEEPBIND | RTLD_NODELETE;
     // The full name of the linked cudart library 'cudart.so.MAJOR.MINOR.PATCH'
-    char libname[] = ANN_BENCH_LINK_CUDART;  // NOLINT
-    handle         = dlopen(ANN_BENCH_LINK_CUDART, kFlags);
+    char libname[] = CUVS_BENCH_LINK_CUDART;  // NOLINT
+    handle         = dlopen(CUVS_BENCH_LINK_CUDART, kFlags);
     if (handle != nullptr) { return; }
     // try strip the PATCH
     auto p = strrchr(libname, '.');
@@ -78,7 +78,7 @@ struct cuda_lib_handle {
   }
   ~cuda_lib_handle() noexcept
   {
-#ifdef ANN_BENCH_LINK_CUDART
+#ifdef CUVS_BENCH_LINK_CUDART
     if (handle != nullptr) { dlclose(handle); }
 #endif
   }
@@ -86,7 +86,7 @@ struct cuda_lib_handle {
   template <typename Symbol>
   auto sym(const char* name) -> Symbol
   {
-#ifdef ANN_BENCH_LINK_CUDART
+#ifdef CUVS_BENCH_LINK_CUDART
     return reinterpret_cast<Symbol>(dlsym(handle, name));
 #else
     return nullptr;
@@ -108,7 +108,7 @@ struct cuda_lib_handle {
   {
 #if defined(BUILD_CPU_ONLY)
     return false;
-#elif defined(ANN_BENCH_LINK_CUDART)
+#elif defined(CUVS_BENCH_LINK_CUDART)
     return handle != nullptr;
 #else
     return true;
@@ -118,7 +118,7 @@ struct cuda_lib_handle {
 
 static inline cuda_lib_handle cudart{};
 
-#ifdef ANN_BENCH_LINK_CUDART
+#ifdef CUVS_BENCH_LINK_CUDART
 namespace stub {
 
 [[gnu::weak, gnu::noinline]] cudaError_t cudaMemcpy(void* dst,
diff --git a/cpp/bench/ann/src/common/util.hpp b/cpp/bench/ann/src/common/util.hpp
index e9e4a9ad2..544aa789a 100644
--- a/cpp/bench/ann/src/common/util.hpp
+++ b/cpp/bench/ann/src/common/util.hpp
@@ -18,7 +18,7 @@
 #include "ann_types.hpp"
 #include "cuda_stub.hpp"  // cuda-related utils
 
-#ifdef ANN_BENCH_NVTX3_HEADERS_FOUND
+#ifdef CUVS_BENCH_NVTX3_HEADERS_FOUND
 #include <nvtx3/nvToolsExt.h>
 #endif
 
@@ -182,7 +182,7 @@ inline auto cuda_info()
 }
 
 struct nvtx_case {
-#ifdef ANN_BENCH_NVTX3_HEADERS_FOUND
+#ifdef CUVS_BENCH_NVTX3_HEADERS_FOUND
  private:
   std::string case_name_;
   std::array<char, 32> iter_name_{0};
@@ -194,7 +194,7 @@ struct nvtx_case {
 
  public:
   struct nvtx_lap {
-#ifdef ANN_BENCH_NVTX3_HEADERS_FOUND
+#ifdef CUVS_BENCH_NVTX3_HEADERS_FOUND
    private:
     nvtxDomainHandle_t domain_;
 
@@ -208,7 +208,7 @@ struct nvtx_case {
 #endif
   };
 
-#ifdef ANN_BENCH_NVTX3_HEADERS_FOUND
+#ifdef CUVS_BENCH_NVTX3_HEADERS_FOUND
   explicit nvtx_case(std::string case_name)
     : case_name_(std::move(case_name)), domain_(nvtxDomainCreateA("ANN benchmark"))
   {
@@ -237,7 +237,7 @@ struct nvtx_case {
 
   [[nodiscard]] auto lap() -> nvtx_case::nvtx_lap
   {
-#ifdef ANN_BENCH_NVTX3_HEADERS_FOUND
+#ifdef CUVS_BENCH_NVTX3_HEADERS_FOUND
     auto i     = iteration_++;
     uint32_t c = (i % 5);
     uint32_t r = 150 + c * 20;
diff --git a/cpp/bench/ann/src/faiss/faiss_cpu_benchmark.cpp b/cpp/bench/ann/src/faiss/faiss_cpu_benchmark.cpp
index 97d1bbf30..82122e8fc 100644
--- a/cpp/bench/ann/src/faiss/faiss_cpu_benchmark.cpp
+++ b/cpp/bench/ann/src/faiss/faiss_cpu_benchmark.cpp
@@ -156,7 +156,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv) { return raft::bench::ann::run_main(argc, argv); }
 #endif
diff --git a/cpp/bench/ann/src/faiss/faiss_gpu_benchmark.cu b/cpp/bench/ann/src/faiss/faiss_gpu_benchmark.cu
index 8b04ba198..77ad2c102 100644
--- a/cpp/bench/ann/src/faiss/faiss_gpu_benchmark.cu
+++ b/cpp/bench/ann/src/faiss/faiss_gpu_benchmark.cu
@@ -156,7 +156,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv) { return raft::bench::ann::run_main(argc, argv); }
 #endif
diff --git a/cpp/bench/ann/src/ggnn/ggnn_benchmark.cu b/cpp/bench/ann/src/ggnn/ggnn_benchmark.cu
index 3b2e97062..01fd1359c 100644
--- a/cpp/bench/ann/src/ggnn/ggnn_benchmark.cu
+++ b/cpp/bench/ann/src/ggnn/ggnn_benchmark.cu
@@ -121,7 +121,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv) { return raft::bench::ann::run_main(argc, argv); }
 #endif
diff --git a/cpp/bench/ann/src/hnswlib/hnswlib_benchmark.cpp b/cpp/bench/ann/src/hnswlib/hnswlib_benchmark.cpp
index 1af19a22c..8e87c64d6 100644
--- a/cpp/bench/ann/src/hnswlib/hnswlib_benchmark.cpp
+++ b/cpp/bench/ann/src/hnswlib/hnswlib_benchmark.cpp
@@ -114,7 +114,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv) { return raft::bench::ann::run_main(argc, argv); }
 #endif
diff --git a/cpp/bench/ann/src/raft/raft_ann_bench_param_parser.h b/cpp/bench/ann/src/raft/raft_ann_bench_param_parser.h
index 1eb0e53cc..a5d1a75e3 100644
--- a/cpp/bench/ann/src/raft/raft_ann_bench_param_parser.h
+++ b/cpp/bench/ann/src/raft/raft_ann_bench_param_parser.h
@@ -20,34 +20,34 @@
 #include <nlohmann/json.hpp>
 
 #undef WARP_SIZE
-#ifdef RAFT_ANN_BENCH_USE_RAFT_BFKNN
+#ifdef CUVS_BENCH_USE_RAFT_BFKNN
 #include "raft_wrapper.h"
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT
+#ifdef CUVS_BENCH_USE_RAFT_IVF_FLAT
 #include "raft_ivf_flat_wrapper.h"
 extern template class raft::bench::ann::RaftIvfFlatGpu<float, int64_t>;
 extern template class raft::bench::ann::RaftIvfFlatGpu<uint8_t, int64_t>;
 extern template class raft::bench::ann::RaftIvfFlatGpu<int8_t, int64_t>;
 #endif
-#if defined(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ) || defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA) || \
-  defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+#if defined(CUVS_BENCH_USE_RAFT_IVF_PQ) || defined(CUVS_BENCH_USE_RAFT_CAGRA) || \
+  defined(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
 #include "raft_ivf_pq_wrapper.h"
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_PQ
+#ifdef CUVS_BENCH_USE_RAFT_IVF_PQ
 extern template class raft::bench::ann::RaftIvfPQ<float, int64_t>;
 extern template class raft::bench::ann::RaftIvfPQ<uint8_t, int64_t>;
 extern template class raft::bench::ann::RaftIvfPQ<int8_t, int64_t>;
 #endif
-#if defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA) || defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+#if defined(CUVS_BENCH_USE_RAFT_CAGRA) || defined(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
 #include "raft_cagra_wrapper.h"
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_CAGRA
+#ifdef CUVS_BENCH_USE_RAFT_CAGRA
 extern template class raft::bench::ann::RaftCagra<float, uint32_t>;
 extern template class raft::bench::ann::RaftCagra<uint8_t, uint32_t>;
 extern template class raft::bench::ann::RaftCagra<int8_t, uint32_t>;
 #endif
 
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT
+#ifdef CUVS_BENCH_USE_RAFT_IVF_FLAT
 template <typename T, typename IdxT>
 void parse_build_param(const nlohmann::json& conf,
                        typename raft::bench::ann::RaftIvfFlatGpu<T, IdxT>::BuildParam& param)
@@ -65,8 +65,8 @@ void parse_search_param(const nlohmann::json& conf,
 }
 #endif
 
-#if defined(RAFT_ANN_BENCH_USE_RAFT_IVF_PQ) || defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA) || \
-  defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+#if defined(CUVS_BENCH_USE_RAFT_IVF_PQ) || defined(CUVS_BENCH_USE_RAFT_CAGRA) || \
+  defined(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
 template <typename T, typename IdxT>
 void parse_build_param(const nlohmann::json& conf,
                        typename raft::bench::ann::RaftIvfPQ<T, IdxT>::BuildParam& param)
@@ -132,7 +132,7 @@ void parse_search_param(const nlohmann::json& conf,
 }
 #endif
 
-#if defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA) || defined(RAFT_ANN_BENCH_USE_RAFT_CAGRA_HNSWLIB)
+#if defined(CUVS_BENCH_USE_RAFT_CAGRA) || defined(CUVS_BENCH_USE_RAFT_CAGRA_HNSWLIB)
 template <typename T, typename IdxT>
 void parse_build_param(const nlohmann::json& conf,
                        raft::neighbors::experimental::nn_descent::index_params& param)
diff --git a/cpp/bench/ann/src/raft/raft_benchmark.cu b/cpp/bench/ann/src/raft/raft_benchmark.cu
index f8c65a2d6..5489e7a86 100644
--- a/cpp/bench/ann/src/raft/raft_benchmark.cu
+++ b/cpp/bench/ann/src/raft/raft_benchmark.cu
@@ -47,28 +47,28 @@ std::unique_ptr<raft::bench::ann::ANN<T>> create_algo(const std::string& algo,
   std::unique_ptr<raft::bench::ann::ANN<T>> ann;
 
   if constexpr (std::is_same_v<T, float>) {
-#ifdef RAFT_ANN_BENCH_USE_RAFT_BFKNN
+#ifdef CUVS_BENCH_USE_RAFT_BFKNN
     if (algo == "raft_bfknn") { ann = std::make_unique<raft::bench::ann::RaftGpu<T>>(metric, dim); }
 #endif
   }
 
   if constexpr (std::is_same_v<T, uint8_t>) {}
 
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT
+#ifdef CUVS_BENCH_USE_RAFT_IVF_FLAT
   if (algo == "raft_ivf_flat") {
     typename raft::bench::ann::RaftIvfFlatGpu<T, int64_t>::BuildParam param;
     parse_build_param<T, int64_t>(conf, param);
     ann = std::make_unique<raft::bench::ann::RaftIvfFlatGpu<T, int64_t>>(metric, dim, param);
   }
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_PQ
+#ifdef CUVS_BENCH_USE_RAFT_IVF_PQ
   if (algo == "raft_ivf_pq") {
     typename raft::bench::ann::RaftIvfPQ<T, int64_t>::BuildParam param;
     parse_build_param<T, int64_t>(conf, param);
     ann = std::make_unique<raft::bench::ann::RaftIvfPQ<T, int64_t>>(metric, dim, param);
   }
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_CAGRA
+#ifdef CUVS_BENCH_USE_RAFT_CAGRA
   if (algo == "raft_cagra") {
     typename raft::bench::ann::RaftCagra<T, uint32_t>::BuildParam param;
     parse_build_param<T, uint32_t>(conf, param);
@@ -85,13 +85,13 @@ template <typename T>
 std::unique_ptr<typename raft::bench::ann::ANN<T>::AnnSearchParam> create_search_param(
   const std::string& algo, const nlohmann::json& conf)
 {
-#ifdef RAFT_ANN_BENCH_USE_RAFT_BFKNN
+#ifdef CUVS_BENCH_USE_RAFT_BFKNN
   if (algo == "raft_brute_force") {
     auto param = std::make_unique<typename raft::bench::ann::ANN<T>::AnnSearchParam>();
     return param;
   }
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_FLAT
+#ifdef CUVS_BENCH_USE_RAFT_IVF_FLAT
   if (algo == "raft_ivf_flat") {
     auto param =
       std::make_unique<typename raft::bench::ann::RaftIvfFlatGpu<T, int64_t>::SearchParam>();
@@ -99,14 +99,14 @@ std::unique_ptr<typename raft::bench::ann::ANN<T>::AnnSearchParam> create_search
     return param;
   }
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_IVF_PQ
+#ifdef CUVS_BENCH_USE_RAFT_IVF_PQ
   if (algo == "raft_ivf_pq") {
     auto param = std::make_unique<typename raft::bench::ann::RaftIvfPQ<T, int64_t>::SearchParam>();
     parse_search_param<T, int64_t>(conf, *param);
     return param;
   }
 #endif
-#ifdef RAFT_ANN_BENCH_USE_RAFT_CAGRA
+#ifdef CUVS_BENCH_USE_RAFT_CAGRA
   if (algo == "raft_cagra") {
     auto param = std::make_unique<typename raft::bench::ann::RaftCagra<T, uint32_t>::SearchParam>();
     parse_search_param<T, uint32_t>(conf, *param);
@@ -124,7 +124,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv)
 {
diff --git a/cpp/bench/ann/src/raft/raft_cagra_hnswlib.cu b/cpp/bench/ann/src/raft/raft_cagra_hnswlib.cu
index ce6fa255b..7b5ff0eba 100644
--- a/cpp/bench/ann/src/raft/raft_cagra_hnswlib.cu
+++ b/cpp/bench/ann/src/raft/raft_cagra_hnswlib.cu
@@ -79,7 +79,7 @@ REGISTER_ALGO_INSTANCE(float);
 REGISTER_ALGO_INSTANCE(std::int8_t);
 REGISTER_ALGO_INSTANCE(std::uint8_t);
 
-#ifdef ANN_BENCH_BUILD_MAIN
+#ifdef CUVS_BENCH_BUILD_MAIN
 #include "../common/benchmark.hpp"
 int main(int argc, char** argv)
 {
diff --git a/cpp/bench/micro/CMakeLists.txt b/cpp/bench/micro/CMakeLists.txt
new file mode 100644
index 000000000..ca22baaf1
--- /dev/null
+++ b/cpp/bench/micro/CMakeLists.txt
@@ -0,0 +1,121 @@
+# =============================================================================
+# Copyright (c) 2022-2023, NVIDIA CORPORATION.
+#
+# Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+# is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
+# or implied. See the License for the specific language governing permissions and limitations under
+# the License.
+# =============================================================================
+
+# ##################################################################################################
+# * compiler function -----------------------------------------------------------------------------
+
+function(ConfigureBench)
+
+  set(options OPTIONAL LIB EXPLICIT_INSTANTIATE_ONLY)
+  set(oneValueArgs NAME)
+  set(multiValueArgs PATH TARGETS CONFIGURATIONS)
+
+  cmake_parse_arguments(ConfigureBench "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
+
+  set(BENCH_NAME ${ConfigureBench_NAME})
+
+  add_executable(${BENCH_NAME} ${ConfigureBench_PATH})
+
+  target_link_libraries(
+    ${BENCH_NAME}
+    PRIVATE raft::raft ${CUVS_CTK_MATH_DEPENDENCIES} benchmark::benchmark Threads::Threads
+            $<TARGET_NAME_IF_EXISTS:OpenMP::OpenMP_CXX> $<TARGET_NAME_IF_EXISTS:conda_env>
+  )
+
+  set_target_properties(
+    ${BENCH_NAME}
+    PROPERTIES # set target compile options
+               INSTALL_RPATH "\$ORIGIN/../../../lib"
+               CXX_STANDARD 17
+               CXX_STANDARD_REQUIRED ON
+               CUDA_STANDARD 17
+               CUDA_STANDARD_REQUIRED ON
+               POSITION_INDEPENDENT_CODE ON
+               INTERFACE_POSITION_INDEPENDENT_CODE ON
+  )
+
+  target_compile_options(
+    ${BENCH_NAME} PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${CUVS_CXX_FLAGS}>"
+                          "$<$<COMPILE_LANGUAGE:CUDA>:${CUVS_CUDA_FLAGS}>"
+  )
+
+  if(ConfigureTest_EXPLICIT_INSTANTIATE_ONLY)
+    target_compile_definitions(${BENCH_NAME} PRIVATE "CUVS_EXPLICIT_INSTANTIATE_ONLY")
+  endif()
+
+  target_include_directories(
+    ${BENCH_NAME} PUBLIC "$<BUILD_INTERFACE:${CUVS_SOURCE_DIR}/bench/micro>"
+  )
+
+  install(
+    TARGETS ${BENCH_NAME}
+    COMPONENT testing
+    DESTINATION bin/gbench/micro/libcuvs
+    EXCLUDE_FROM_ALL
+  )
+
+endfunction()
+
+if(BUILD_PRIMS_BENCH)
+  ConfigureBench(
+    NAME CLUSTER_BENCH PATH bench/micro/cluster/kmeans_balanced.cu bench/micro/cluster/kmeans.cu
+    bench/micro/main.cpp OPTIONAL LIB EXPLICIT_INSTANTIATE_ONLY
+  )
+
+  ConfigureBench(
+    NAME TUNE_DISTANCE PATH bench/micro/distance/tune_pairwise/kernel.cu
+    bench/micro/distance/tune_pairwise/bench.cu bench/micro/main.cpp
+  )
+
+  ConfigureBench(
+    NAME
+    DISTANCE_BENCH
+    PATH
+    bench/micro/distance/distance_cosine.cu
+    bench/micro/distance/distance_exp_l2.cu
+    bench/micro/distance/distance_l1.cu
+    bench/micro/distance/distance_unexp_l2.cu
+    bench/micro/distance/fused_l2_nn.cu
+    bench/micro/distance/masked_nn.cu
+    bench/micro/distance/kernels.cu
+    bench/micro/main.cpp
+    OPTIONAL
+    LIB
+    EXPLICIT_INSTANTIATE_ONLY
+  )
+
+  ConfigureBench(
+    NAME
+    NEIGHBORS_BENCH
+    PATH
+    bench/micro/neighbors/knn/brute_force_float_int64_t.cu
+    bench/micro/neighbors/knn/brute_force_float_uint32_t.cu
+    bench/micro/neighbors/knn/cagra_float_uint32_t.cu
+    bench/micro/neighbors/knn/ivf_flat_filter_float_int64_t.cu
+    bench/micro/neighbors/knn/ivf_flat_float_int64_t.cu
+    bench/micro/neighbors/knn/ivf_flat_int8_t_int64_t.cu
+    bench/micro/neighbors/knn/ivf_flat_uint8_t_int64_t.cu
+    bench/micro/neighbors/knn/ivf_pq_float_int64_t.cu
+    bench/micro/neighbors/knn/ivf_pq_filter_float_int64_t.cu
+    bench/micro/neighbors/knn/ivf_pq_int8_t_int64_t.cu
+    bench/micro/neighbors/knn/ivf_pq_uint8_t_int64_t.cu
+    bench/micro/neighbors/refine_float_int64_t.cu
+    bench/micro/neighbors/refine_uint8_t_int64_t.cu
+    bench/micro/main.cpp
+    OPTIONAL
+    LIB
+    EXPLICIT_INSTANTIATE_ONLY
+  )
+
+endif()
diff --git a/cpp/bench/prims/cluster/kmeans.cu b/cpp/bench/micro/cluster/kmeans.cu
similarity index 100%
rename from cpp/bench/prims/cluster/kmeans.cu
rename to cpp/bench/micro/cluster/kmeans.cu
diff --git a/cpp/bench/prims/cluster/kmeans_balanced.cu b/cpp/bench/micro/cluster/kmeans_balanced.cu
similarity index 100%
rename from cpp/bench/prims/cluster/kmeans_balanced.cu
rename to cpp/bench/micro/cluster/kmeans_balanced.cu
diff --git a/cpp/bench/prims/common/benchmark.hpp b/cpp/bench/micro/common/benchmark.hpp
similarity index 100%
rename from cpp/bench/prims/common/benchmark.hpp
rename to cpp/bench/micro/common/benchmark.hpp
diff --git a/cpp/bench/prims/distance/distance_common.cuh b/cpp/bench/micro/distance/distance_common.cuh
similarity index 100%
rename from cpp/bench/prims/distance/distance_common.cuh
rename to cpp/bench/micro/distance/distance_common.cuh
diff --git a/cpp/bench/prims/distance/distance_cosine.cu b/cpp/bench/micro/distance/distance_cosine.cu
similarity index 100%
rename from cpp/bench/prims/distance/distance_cosine.cu
rename to cpp/bench/micro/distance/distance_cosine.cu
diff --git a/cpp/bench/prims/distance/distance_exp_l2.cu b/cpp/bench/micro/distance/distance_exp_l2.cu
similarity index 100%
rename from cpp/bench/prims/distance/distance_exp_l2.cu
rename to cpp/bench/micro/distance/distance_exp_l2.cu
diff --git a/cpp/bench/prims/distance/distance_l1.cu b/cpp/bench/micro/distance/distance_l1.cu
similarity index 100%
rename from cpp/bench/prims/distance/distance_l1.cu
rename to cpp/bench/micro/distance/distance_l1.cu
diff --git a/cpp/bench/prims/distance/distance_unexp_l2.cu b/cpp/bench/micro/distance/distance_unexp_l2.cu
similarity index 100%
rename from cpp/bench/prims/distance/distance_unexp_l2.cu
rename to cpp/bench/micro/distance/distance_unexp_l2.cu
diff --git a/cpp/bench/prims/distance/fused_l2_nn.cu b/cpp/bench/micro/distance/fused_l2_nn.cu
similarity index 100%
rename from cpp/bench/prims/distance/fused_l2_nn.cu
rename to cpp/bench/micro/distance/fused_l2_nn.cu
diff --git a/cpp/bench/prims/distance/kernels.cu b/cpp/bench/micro/distance/kernels.cu
similarity index 100%
rename from cpp/bench/prims/distance/kernels.cu
rename to cpp/bench/micro/distance/kernels.cu
diff --git a/cpp/bench/prims/distance/masked_nn.cu b/cpp/bench/micro/distance/masked_nn.cu
similarity index 100%
rename from cpp/bench/prims/distance/masked_nn.cu
rename to cpp/bench/micro/distance/masked_nn.cu
diff --git a/cpp/bench/prims/distance/tune_pairwise/bench.cu b/cpp/bench/micro/distance/tune_pairwise/bench.cu
similarity index 100%
rename from cpp/bench/prims/distance/tune_pairwise/bench.cu
rename to cpp/bench/micro/distance/tune_pairwise/bench.cu
diff --git a/cpp/bench/prims/distance/tune_pairwise/kernel.cu b/cpp/bench/micro/distance/tune_pairwise/kernel.cu
similarity index 100%
rename from cpp/bench/prims/distance/tune_pairwise/kernel.cu
rename to cpp/bench/micro/distance/tune_pairwise/kernel.cu
diff --git a/cpp/bench/prims/distance/tune_pairwise/kernel.cuh b/cpp/bench/micro/distance/tune_pairwise/kernel.cuh
similarity index 100%
rename from cpp/bench/prims/distance/tune_pairwise/kernel.cuh
rename to cpp/bench/micro/distance/tune_pairwise/kernel.cuh
diff --git a/cpp/bench/prims/main.cpp b/cpp/bench/micro/main.cpp
similarity index 100%
rename from cpp/bench/prims/main.cpp
rename to cpp/bench/micro/main.cpp
diff --git a/cpp/bench/prims/matrix/main.cpp b/cpp/bench/micro/matrix/main.cpp
similarity index 100%
rename from cpp/bench/prims/matrix/main.cpp
rename to cpp/bench/micro/matrix/main.cpp
diff --git a/cpp/bench/prims/matrix/select_k.cu b/cpp/bench/micro/matrix/select_k.cu
similarity index 100%
rename from cpp/bench/prims/matrix/select_k.cu
rename to cpp/bench/micro/matrix/select_k.cu
diff --git a/cpp/bench/prims/neighbors/cagra_bench.cuh b/cpp/bench/micro/neighbors/cagra_bench.cuh
similarity index 100%
rename from cpp/bench/prims/neighbors/cagra_bench.cuh
rename to cpp/bench/micro/neighbors/cagra_bench.cuh
diff --git a/cpp/bench/prims/neighbors/knn.cuh b/cpp/bench/micro/neighbors/knn.cuh
similarity index 100%
rename from cpp/bench/prims/neighbors/knn.cuh
rename to cpp/bench/micro/neighbors/knn.cuh
diff --git a/cpp/bench/prims/neighbors/knn/brute_force_float_int64_t.cu b/cpp/bench/micro/neighbors/knn/brute_force_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/brute_force_float_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/brute_force_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/brute_force_float_uint32_t.cu b/cpp/bench/micro/neighbors/knn/brute_force_float_uint32_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/brute_force_float_uint32_t.cu
rename to cpp/bench/micro/neighbors/knn/brute_force_float_uint32_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/cagra_float_uint32_t.cu b/cpp/bench/micro/neighbors/knn/cagra_float_uint32_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/cagra_float_uint32_t.cu
rename to cpp/bench/micro/neighbors/knn/cagra_float_uint32_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_flat_filter_float_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_flat_filter_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_flat_filter_float_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_flat_filter_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_flat_float_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_flat_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_flat_float_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_flat_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_flat_int8_t_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_flat_int8_t_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_flat_int8_t_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_flat_int8_t_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_flat_uint8_t_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_flat_uint8_t_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_flat_uint8_t_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_flat_uint8_t_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_pq_filter_float_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_pq_filter_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_pq_filter_float_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_pq_filter_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_pq_float_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_pq_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_pq_float_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_pq_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_pq_int8_t_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_pq_int8_t_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_pq_int8_t_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_pq_int8_t_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/knn/ivf_pq_uint8_t_int64_t.cu b/cpp/bench/micro/neighbors/knn/ivf_pq_uint8_t_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/knn/ivf_pq_uint8_t_int64_t.cu
rename to cpp/bench/micro/neighbors/knn/ivf_pq_uint8_t_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/refine.cuh b/cpp/bench/micro/neighbors/refine.cuh
similarity index 100%
rename from cpp/bench/prims/neighbors/refine.cuh
rename to cpp/bench/micro/neighbors/refine.cuh
diff --git a/cpp/bench/prims/neighbors/refine_float_int64_t.cu b/cpp/bench/micro/neighbors/refine_float_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/refine_float_int64_t.cu
rename to cpp/bench/micro/neighbors/refine_float_int64_t.cu
diff --git a/cpp/bench/prims/neighbors/refine_uint8_t_int64_t.cu b/cpp/bench/micro/neighbors/refine_uint8_t_int64_t.cu
similarity index 100%
rename from cpp/bench/prims/neighbors/refine_uint8_t_int64_t.cu
rename to cpp/bench/micro/neighbors/refine_uint8_t_int64_t.cu
diff --git a/cpp/bench/prims/CMakeLists.txt b/cpp/bench/prims/CMakeLists.txt
deleted file mode 100644
index fe58453d0..000000000
--- a/cpp/bench/prims/CMakeLists.txt
+++ /dev/null
@@ -1,166 +0,0 @@
-# =============================================================================
-# Copyright (c) 2022-2023, NVIDIA CORPORATION.
-#
-# Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
-# in compliance with the License. You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software distributed under the License
-# is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
-# or implied. See the License for the specific language governing permissions and limitations under
-# the License.
-# =============================================================================
-
-# ##################################################################################################
-# * compiler function -----------------------------------------------------------------------------
-
-function(ConfigureBench)
-
-  set(options OPTIONAL LIB EXPLICIT_INSTANTIATE_ONLY)
-  set(oneValueArgs NAME)
-  set(multiValueArgs PATH TARGETS CONFIGURATIONS)
-
-  cmake_parse_arguments(ConfigureBench "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
-
-  set(BENCH_NAME ${ConfigureBench_NAME})
-
-  add_executable(${BENCH_NAME} ${ConfigureBench_PATH})
-
-  target_link_libraries(
-    ${BENCH_NAME}
-    PRIVATE raft::raft
-            raft_internal
-            $<$<BOOL:${ConfigureBench_LIB}>:raft::compiled>
-            ${RAFT_CTK_MATH_DEPENDENCIES}
-            benchmark::benchmark
-            Threads::Threads
-            $<TARGET_NAME_IF_EXISTS:OpenMP::OpenMP_CXX>
-            $<TARGET_NAME_IF_EXISTS:conda_env>
-  )
-
-  set_target_properties(
-    ${BENCH_NAME}
-    PROPERTIES # set target compile options
-               INSTALL_RPATH "\$ORIGIN/../../../lib"
-               CXX_STANDARD 17
-               CXX_STANDARD_REQUIRED ON
-               CUDA_STANDARD 17
-               CUDA_STANDARD_REQUIRED ON
-               POSITION_INDEPENDENT_CODE ON
-               INTERFACE_POSITION_INDEPENDENT_CODE ON
-  )
-
-  target_compile_options(
-    ${BENCH_NAME} PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${RAFT_CXX_FLAGS}>"
-                          "$<$<COMPILE_LANGUAGE:CUDA>:${RAFT_CUDA_FLAGS}>"
-  )
-
-  if(ConfigureTest_EXPLICIT_INSTANTIATE_ONLY)
-    target_compile_definitions(${BENCH_NAME} PRIVATE "RAFT_EXPLICIT_INSTANTIATE_ONLY")
-  endif()
-
-  target_include_directories(
-    ${BENCH_NAME} PUBLIC "$<BUILD_INTERFACE:${RAFT_SOURCE_DIR}/bench/prims>"
-  )
-
-  install(
-    TARGETS ${BENCH_NAME}
-    COMPONENT testing
-    DESTINATION bin/gbench/prims/libraft
-    EXCLUDE_FROM_ALL
-  )
-
-endfunction()
-
-if(BUILD_PRIMS_BENCH)
-  ConfigureBench(
-    NAME CORE_BENCH PATH bench/prims/core/bitset.cu bench/prims/core/copy.cu bench/prims/main.cpp
-  )
-
-  ConfigureBench(
-    NAME CLUSTER_BENCH PATH bench/prims/cluster/kmeans_balanced.cu bench/prims/cluster/kmeans.cu
-    bench/prims/main.cpp OPTIONAL LIB EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureBench(
-    NAME TUNE_DISTANCE PATH bench/prims/distance/tune_pairwise/kernel.cu
-    bench/prims/distance/tune_pairwise/bench.cu bench/prims/main.cpp
-  )
-
-  ConfigureBench(
-    NAME
-    DISTANCE_BENCH
-    PATH
-    bench/prims/distance/distance_cosine.cu
-    bench/prims/distance/distance_exp_l2.cu
-    bench/prims/distance/distance_l1.cu
-    bench/prims/distance/distance_unexp_l2.cu
-    bench/prims/distance/fused_l2_nn.cu
-    bench/prims/distance/masked_nn.cu
-    bench/prims/distance/kernels.cu
-    bench/prims/main.cpp
-    OPTIONAL
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureBench(
-    NAME
-    LINALG_BENCH
-    PATH
-    bench/prims/linalg/add.cu
-    bench/prims/linalg/map_then_reduce.cu
-    bench/prims/linalg/matrix_vector_op.cu
-    bench/prims/linalg/norm.cu
-    bench/prims/linalg/normalize.cu
-    bench/prims/linalg/reduce_cols_by_key.cu
-    bench/prims/linalg/reduce_rows_by_key.cu
-    bench/prims/linalg/reduce.cu
-    bench/prims/main.cpp
-  )
-
-  ConfigureBench(
-    NAME
-    MATRIX_BENCH
-    PATH
-    bench/prims/matrix/argmin.cu
-    bench/prims/matrix/gather.cu
-    bench/prims/matrix/select_k.cu
-    bench/prims/matrix/main.cpp
-    OPTIONAL
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureBench(
-    NAME RANDOM_BENCH PATH bench/prims/random/make_blobs.cu bench/prims/random/permute.cu
-    bench/prims/random/rng.cu bench/prims/main.cpp
-  )
-
-  ConfigureBench(NAME SPARSE_BENCH PATH bench/prims/sparse/convert_csr.cu bench/prims/main.cpp)
-
-  ConfigureBench(
-    NAME
-    NEIGHBORS_BENCH
-    PATH
-    bench/prims/neighbors/knn/brute_force_float_int64_t.cu
-    bench/prims/neighbors/knn/brute_force_float_uint32_t.cu
-    bench/prims/neighbors/knn/cagra_float_uint32_t.cu
-    bench/prims/neighbors/knn/ivf_flat_filter_float_int64_t.cu
-    bench/prims/neighbors/knn/ivf_flat_float_int64_t.cu
-    bench/prims/neighbors/knn/ivf_flat_int8_t_int64_t.cu
-    bench/prims/neighbors/knn/ivf_flat_uint8_t_int64_t.cu
-    bench/prims/neighbors/knn/ivf_pq_float_int64_t.cu
-    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
-    bench/prims/neighbors/refine_float_int64_t.cu
-    bench/prims/neighbors/refine_uint8_t_int64_t.cu
-    bench/prims/main.cpp
-    OPTIONAL
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
-endif()
diff --git a/cpp/bench/prims/core/bitset.cu b/cpp/bench/prims/core/bitset.cu
deleted file mode 100644
index ce3136bcd..000000000
--- a/cpp/bench/prims/core/bitset.cu
+++ /dev/null
@@ -1,74 +0,0 @@
-/*
- * Copyright (c) 2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/core/bitset.cuh>
-#include <raft/core/device_mdspan.hpp>
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::core {
-
-struct bitset_inputs {
-  uint32_t bitset_len;
-  uint32_t mask_len;
-  uint32_t query_len;
-};  // struct bitset_inputs
-
-template <typename bitset_t, typename index_t>
-struct bitset_bench : public fixture {
-  bitset_bench(const bitset_inputs& p)
-    : params(p),
-      mask{raft::make_device_vector<index_t, index_t>(res, p.mask_len)},
-      queries{raft::make_device_vector<index_t, index_t>(res, p.query_len)},
-      outputs{raft::make_device_vector<bool, index_t>(res, p.query_len)}
-  {
-    raft::random::RngState state{42};
-    raft::random::uniformInt(res, state, mask.view(), index_t{0}, index_t{p.bitset_len});
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      auto my_bitset = raft::core::bitset<bitset_t, index_t>(
-        this->res, raft::make_const_mdspan(mask.view()), params.bitset_len);
-      my_bitset.test(this->res, raft::make_const_mdspan(queries.view()), outputs.view());
-    });
-  }
-
- private:
-  raft::resources res;
-  bitset_inputs params;
-  raft::device_vector<index_t, index_t> mask, queries;
-  raft::device_vector<bool, index_t> outputs;
-};  // struct bitset
-
-const std::vector<bitset_inputs> bitset_input_vecs{
-  {256 * 1024 * 1024, 64 * 1024 * 1024, 256 * 1024 * 1024},    // Standard Bench
-  {256 * 1024 * 1024, 64 * 1024 * 1024, 1024 * 1024 * 1024},   // Extra queries
-  {128 * 1024 * 1024, 1024 * 1024 * 1024, 256 * 1024 * 1024},  // Extra mask to test atomics impact
-};
-
-using Uint8_32  = bitset_bench<uint8_t, uint32_t>;
-using Uint16_64 = bitset_bench<uint16_t, uint32_t>;
-using Uint32_32 = bitset_bench<uint32_t, uint32_t>;
-using Uint32_64 = bitset_bench<uint32_t, uint64_t>;
-
-RAFT_BENCH_REGISTER(Uint8_32, "", bitset_input_vecs);
-RAFT_BENCH_REGISTER(Uint16_64, "", bitset_input_vecs);
-RAFT_BENCH_REGISTER(Uint32_32, "", bitset_input_vecs);
-RAFT_BENCH_REGISTER(Uint32_64, "", bitset_input_vecs);
-
-}  // namespace raft::bench::core
diff --git a/cpp/bench/prims/core/copy.cu b/cpp/bench/prims/core/copy.cu
deleted file mode 100644
index 31ee83b92..000000000
--- a/cpp/bench/prims/core/copy.cu
+++ /dev/null
@@ -1,401 +0,0 @@
-/*
- * Copyright (c) 2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <cstdint>
-#include <raft/core/copy.cuh>
-#include <raft/core/cudart_utils.hpp>
-#include <raft/core/device_mdarray.hpp>
-#include <raft/core/device_resources.hpp>
-#include <raft/core/host_mdarray.hpp>
-#include <raft/core/mdspan_types.hpp>
-#include <raft/core/memory_type.hpp>
-#include <raft/thirdparty/mdspan/include/experimental/mdspan>
-
-namespace raft::bench::core {
-
-template <typename IdxT, std::size_t Rank>
-auto constexpr const default_dims = []() {
-  auto dims = std::array<IdxT, Rank>{};
-  std::fill(dims.begin(), dims.end(), 2);
-  return dims;
-}();
-
-template <typename IdxT>
-auto constexpr const default_dims<IdxT, std::size_t{1}> = std::array<IdxT, 1>{3000000};
-
-template <typename IdxT>
-auto constexpr const default_dims<IdxT, std::size_t{2}> = std::array<IdxT, 2>{1000, 3000};
-
-template <typename IdxT>
-auto constexpr const default_dims<IdxT, std::size_t{3}> = std::array<IdxT, 3>{20, 300, 500};
-
-template <typename T,
-          typename IdxT,
-          typename LayoutPolicy,
-          memory_type MemType,
-          std::size_t Rank,
-          typename = std::make_index_sequence<Rank>>
-struct bench_array_type;
-
-template <typename T,
-          typename IdxT,
-          typename LayoutPolicy,
-          memory_type MemType,
-          std::size_t Rank,
-          std::size_t... S>
-struct bench_array_type<T, IdxT, LayoutPolicy, MemType, Rank, std::index_sequence<S...>> {
-  template <std::size_t>
-  auto static constexpr const extent_type = raft::dynamic_extent;
-
-  using type =
-    std::conditional_t<MemType == memory_type::host,
-                       host_mdarray<T, extents<IdxT, extent_type<S>...>, LayoutPolicy>,
-                       device_mdarray<T, extents<IdxT, extent_type<S>...>, LayoutPolicy>>;
-};
-
-template <typename SrcT,
-          typename DstT,
-          typename IdxT,
-          typename SrcLayoutPolicy,
-          typename DstLayoutPolicy,
-          memory_type SrcMemType,
-          memory_type DstMemType,
-          std::size_t Rank>
-struct params {
-  std::array<IdxT, Rank> dims = default_dims<IdxT, Rank>;
-  using src_array_type =
-    typename bench_array_type<SrcT, IdxT, SrcLayoutPolicy, SrcMemType, Rank>::type;
-  using dst_array_type =
-    typename bench_array_type<DstT, IdxT, DstLayoutPolicy, DstMemType, Rank>::type;
-};
-
-template <typename SrcT,
-          typename DstT,
-          typename IdxT,
-          typename SrcLayoutPolicy,
-          typename DstLayoutPolicy,
-          memory_type SrcMemType,
-          memory_type DstMemType,
-          std::size_t Rank>
-struct CopyBench : public fixture {
-  using params_type =
-    params<SrcT, DstT, IdxT, SrcLayoutPolicy, DstLayoutPolicy, SrcMemType, DstMemType, Rank>;
-  using src_array_type = typename params_type::src_array_type;
-  using dst_array_type = typename params_type::dst_array_type;
-  explicit CopyBench(const params_type& ps)
-    : fixture{true},
-      res_{},
-      params_{ps},
-      src_{
-        res_,
-        typename src_array_type::mapping_type{
-          std::apply([](auto... exts) { return make_extents<IdxT>(exts...); }, ps.dims)},
-        typename src_array_type::container_policy_type{},
-      },
-      dst_{
-        res_,
-        typename dst_array_type::mapping_type{
-          std::apply([](auto... exts) { return make_extents<IdxT>(exts...); }, ps.dims)},
-        typename dst_array_type::container_policy_type{},
-      }
-  {
-    res_.get_cublas_handle();  // initialize cublas handle
-    auto src_data = std::vector<SrcT>(src_.size());
-    std::iota(src_data.begin(), src_data.end(), SrcT{});
-    raft::copy(src_.data_handle(), src_data.data(), src_.size(), res_.get_stream());
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() { raft::copy(res_, dst_.view(), src_.view()); });
-  }
-
- private:
-  raft::device_resources res_;
-  params_type params_;
-  src_array_type src_;
-  dst_array_type dst_;
-};
-
-template <typename ParamsT>
-auto static const inputs = std::vector<ParamsT>{ParamsT{}};
-
-#define COPY_REGISTER(BenchT) \
-  RAFT_BENCH_REGISTER(BenchT, "BenchT", inputs<typename BenchT::params_type>)
-
-using copy_bench_device_device_1d_same_dtype_same_layout        = CopyBench<int,
-                                                                     int,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_c_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     1>;
-using copy_bench_device_device_1d_same_dtype_diff_layout        = CopyBench<int,
-                                                                     int,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_f_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     1>;
-using copy_bench_device_device_1d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                     double,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_f_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     1>;
-using copy_bench_device_device_2d_same_dtype_diff_layout        = CopyBench<int,
-                                                                     int,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_f_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     2>;
-using copy_bench_device_device_2d_same_dtype_diff_layout_cublas = CopyBench<float,
-                                                                            float,
-                                                                            int,
-                                                                            layout_c_contiguous,
-                                                                            layout_f_contiguous,
-                                                                            memory_type::device,
-                                                                            memory_type::device,
-                                                                            2>;
-using copy_bench_device_device_3d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                     double,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_f_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     3>;
-using copy_bench_device_device_3d_diff_dtype_same_layout        = CopyBench<float,
-                                                                     double,
-                                                                     int,
-                                                                     layout_c_contiguous,
-                                                                     layout_c_contiguous,
-                                                                     memory_type::device,
-                                                                     memory_type::device,
-                                                                     3>;
-
-using copy_bench_host_host_1d_same_dtype_same_layout             = CopyBench<int,
-                                                                 int,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_c_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 1>;
-using copy_bench_host_host_1d_same_dtype_diff_layout             = CopyBench<int,
-                                                                 int,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_f_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 1>;
-using copy_bench_host_host_1d_diff_dtype_diff_layout             = CopyBench<float,
-                                                                 double,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_f_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 1>;
-using copy_bench_host_host_2d_same_dtype_diff_layout             = CopyBench<int,
-                                                                 int,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_f_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 2>;
-using copy_bench_host_host_2d_same_dtype_diff_layout_float_float = CopyBench<float,
-                                                                             float,
-                                                                             int,
-                                                                             layout_c_contiguous,
-                                                                             layout_f_contiguous,
-                                                                             memory_type::host,
-                                                                             memory_type::host,
-                                                                             2>;
-using copy_bench_host_host_3d_diff_dtype_same_layout             = CopyBench<float,
-                                                                 double,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_c_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 3>;
-using copy_bench_host_host_3d_diff_dtype_diff_layout             = CopyBench<float,
-                                                                 double,
-                                                                 int,
-                                                                 layout_c_contiguous,
-                                                                 layout_f_contiguous,
-                                                                 memory_type::host,
-                                                                 memory_type::host,
-                                                                 3>;
-
-using copy_bench_device_host_1d_same_dtype_same_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_c_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   1>;
-using copy_bench_device_host_1d_same_dtype_diff_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   1>;
-using copy_bench_device_host_1d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   1>;
-using copy_bench_device_host_2d_same_dtype_diff_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   2>;
-using copy_bench_device_host_2d_same_dtype_diff_layout_cublas = CopyBench<float,
-                                                                          float,
-                                                                          int,
-                                                                          layout_c_contiguous,
-                                                                          layout_f_contiguous,
-                                                                          memory_type::device,
-                                                                          memory_type::host,
-                                                                          2>;
-using copy_bench_device_host_3d_diff_dtype_same_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_c_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   3>;
-using copy_bench_device_host_3d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::device,
-                                                                   memory_type::host,
-                                                                   3>;
-
-using copy_bench_host_device_1d_same_dtype_same_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_c_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   1>;
-using copy_bench_host_device_1d_same_dtype_diff_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   1>;
-using copy_bench_host_device_1d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   1>;
-using copy_bench_host_device_2d_same_dtype_diff_layout        = CopyBench<int,
-                                                                   int,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   2>;
-using copy_bench_host_device_2d_same_dtype_diff_layout_cublas = CopyBench<float,
-                                                                          float,
-                                                                          int,
-                                                                          layout_c_contiguous,
-                                                                          layout_f_contiguous,
-                                                                          memory_type::host,
-                                                                          memory_type::device,
-                                                                          2>;
-using copy_bench_host_device_3d_diff_dtype_diff_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_f_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   3>;
-using copy_bench_host_device_3d_diff_dtype_same_layout        = CopyBench<float,
-                                                                   double,
-                                                                   int,
-                                                                   layout_c_contiguous,
-                                                                   layout_c_contiguous,
-                                                                   memory_type::host,
-                                                                   memory_type::device,
-                                                                   3>;
-
-// COPY_REGISTER(copy_bench_same_dtype_1d_host_host);
-COPY_REGISTER(copy_bench_device_device_1d_same_dtype_same_layout);
-COPY_REGISTER(copy_bench_device_device_1d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_device_1d_diff_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_device_2d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_device_2d_same_dtype_diff_layout_cublas);
-COPY_REGISTER(copy_bench_device_device_3d_diff_dtype_same_layout);
-COPY_REGISTER(copy_bench_device_device_3d_diff_dtype_diff_layout);
-
-COPY_REGISTER(copy_bench_host_host_1d_same_dtype_same_layout);
-COPY_REGISTER(copy_bench_host_host_1d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_host_1d_diff_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_host_2d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_host_2d_same_dtype_diff_layout_float_float);
-COPY_REGISTER(copy_bench_host_host_3d_diff_dtype_same_layout);
-COPY_REGISTER(copy_bench_host_host_3d_diff_dtype_diff_layout);
-
-COPY_REGISTER(copy_bench_device_host_1d_same_dtype_same_layout);
-COPY_REGISTER(copy_bench_device_host_1d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_host_1d_diff_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_host_2d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_device_host_2d_same_dtype_diff_layout_cublas);
-COPY_REGISTER(copy_bench_device_host_3d_diff_dtype_same_layout);
-COPY_REGISTER(copy_bench_device_host_3d_diff_dtype_diff_layout);
-
-COPY_REGISTER(copy_bench_host_device_1d_same_dtype_same_layout);
-COPY_REGISTER(copy_bench_host_device_1d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_device_1d_diff_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_device_2d_same_dtype_diff_layout);
-COPY_REGISTER(copy_bench_host_device_2d_same_dtype_diff_layout_cublas);
-COPY_REGISTER(copy_bench_host_device_3d_diff_dtype_same_layout);
-COPY_REGISTER(copy_bench_host_device_3d_diff_dtype_diff_layout);
-
-}  // namespace raft::bench::core
diff --git a/cpp/bench/prims/linalg/add.cu b/cpp/bench/prims/linalg/add.cu
deleted file mode 100644
index 456214ad7..000000000
--- a/cpp/bench/prims/linalg/add.cu
+++ /dev/null
@@ -1,51 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/add.cuh>
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-struct add_inputs {
-  int len;
-};  // struct add_inputs
-
-template <typename T>
-struct add : public fixture {
-  add(const add_inputs& p) : params(p), ptr0(p.len, stream), ptr1(p.len, stream) {}
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::linalg::add(ptr0.data(), ptr0.data(), ptr1.data(), params.len, stream);
-    });
-  }
-
- private:
-  add_inputs params;
-  rmm::device_uvector<T> ptr0, ptr1;
-};  // struct add
-
-const std::vector<add_inputs> add_input_vecs{
-  {256 * 1024 * 1024}, {256 * 1024 * 1024 + 2}, {256 * 1024 * 1024 + 1}
-
-};
-
-RAFT_BENCH_REGISTER(add<float>, "", add_input_vecs);
-RAFT_BENCH_REGISTER(add<double>, "", add_input_vecs);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/map_then_reduce.cu b/cpp/bench/prims/linalg/map_then_reduce.cu
deleted file mode 100644
index 84aebd85b..000000000
--- a/cpp/bench/prims/linalg/map_then_reduce.cu
+++ /dev/null
@@ -1,64 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/map_then_reduce.cuh>
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-struct map_then_reduce_inputs {
-  int len;
-};
-
-template <typename Type>
-struct Identity {
-  HDI Type operator()(Type a) { return a; }
-};
-
-template <typename T>
-struct map_then_reduce : public fixture {
-  map_then_reduce(const map_then_reduce_inputs& p) : params(p), in(p.len, stream), out(1, stream) {}
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::linalg::mapThenSumReduce(out.data(), params.len, Identity<T>(), stream, in.data());
-    });
-  }
-
- private:
-  map_then_reduce_inputs params;
-  rmm::device_uvector<T> out, in;
-};  // struct MapThenReduce
-
-const std::vector<map_then_reduce_inputs> map_then_reduce_input_vecs{
-  {1024 * 1024},
-  {32 * 1024 * 1024},
-  {1024 * 1024 * 1024},
-  {1024 * 1024 + 2},
-  {32 * 1024 * 1024 + 2},
-  {1024 * 1024 * 1024 + 2},
-  {1024 * 1024 + 1},
-  {32 * 1024 * 1024 + 1},
-  {1024 * 1024 * 1024 + 1},
-
-};
-
-RAFT_BENCH_REGISTER(map_then_reduce<float>, "", map_then_reduce_input_vecs);
-RAFT_BENCH_REGISTER(map_then_reduce<double>, "", map_then_reduce_input_vecs);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/matrix_vector_op.cu b/cpp/bench/prims/linalg/matrix_vector_op.cu
deleted file mode 100644
index d1fbaee79..000000000
--- a/cpp/bench/prims/linalg/matrix_vector_op.cu
+++ /dev/null
@@ -1,156 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/matrix_vector_op.cuh>
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-template <typename IdxT>
-struct mat_vec_op_inputs {
-  IdxT rows, cols;
-  bool rowMajor, bcastAlongRows;
-  IdxT inAlignOffset, outAlignOffset;
-};  // struct mat_vec_op_inputs
-
-template <typename IdxT>
-inline auto operator<<(std::ostream& os, const mat_vec_op_inputs<IdxT>& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols << "#" << p.rowMajor << "#" << p.bcastAlongRows << "#"
-     << p.inAlignOffset << "#" << p.outAlignOffset;
-  return os;
-}
-
-template <typename OpT, typename T, typename IdxT>
-struct mat_vec_op : public fixture {
-  mat_vec_op(const mat_vec_op_inputs<IdxT>& p)
-    : params(p),
-      out(p.rows * p.cols + params.outAlignOffset, stream),
-      in(p.rows * p.cols + params.inAlignOffset, stream),
-      vec1(p.bcastAlongRows ? p.cols : p.rows, stream),
-      vec2(p.bcastAlongRows ? p.cols : p.rows, stream)
-  {
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      if constexpr (OpT::useTwoVectors) {
-        raft::linalg::matrixVectorOp(out.data() + params.outAlignOffset,
-                                     in.data() + params.inAlignOffset,
-                                     vec1.data(),
-                                     vec2.data(),
-                                     params.cols,
-                                     params.rows,
-                                     params.rowMajor,
-                                     params.bcastAlongRows,
-                                     OpT{},
-                                     stream);
-      } else {
-        raft::linalg::matrixVectorOp(out.data() + params.outAlignOffset,
-                                     in.data() + params.inAlignOffset,
-                                     vec1.data(),
-                                     params.cols,
-                                     params.rows,
-                                     params.rowMajor,
-                                     params.bcastAlongRows,
-                                     OpT{},
-                                     stream);
-      }
-    });
-  }
-
- private:
-  mat_vec_op_inputs<IdxT> params;
-  rmm::device_uvector<T> out, in, vec1, vec2;
-};  // struct MatVecOp
-
-template <typename IdxT>
-std::vector<mat_vec_op_inputs<IdxT>> get_mv_inputs()
-{
-  std::vector<mat_vec_op_inputs<IdxT>> out;
-
-  // Scalability benchmark with round dimensions
-  std::vector<IdxT> rows = {1000, 100000, 1000000};
-  std::vector<IdxT> cols = {8, 64, 256, 1024};
-  for (bool rowMajor : {true, false}) {
-    for (bool alongRows : {true, false}) {
-      for (IdxT rows_ : rows) {
-        for (IdxT cols_ : cols) {
-          out.push_back({rows_, cols_, rowMajor, alongRows, 0, 0});
-        }
-      }
-    }
-  }
-
-  // Odd dimensions, misalignment
-  std::vector<std::tuple<IdxT, IdxT>> rowcols = {
-    {44739207, 7},
-    {44739207, 15},
-    {44739207, 16},
-    {44739207, 17},
-    {2611236, 256},
-    {2611236, 257},
-    {2611236, 263},
-  };
-  for (bool rowMajor : {true, false}) {
-    for (bool alongRows : {true, false}) {
-      for (auto rc : rowcols) {
-        for (IdxT inAlignOffset : {0, 1}) {
-          for (IdxT outAlignOffset : {0, 1}) {
-            out.push_back({std::get<0>(rc),
-                           std::get<1>(rc),
-                           rowMajor,
-                           alongRows,
-                           inAlignOffset,
-                           outAlignOffset});
-          }
-        }
-      }
-    }
-  }
-  return out;
-}
-
-const std::vector<mat_vec_op_inputs<int>> mv_input_i32     = get_mv_inputs<int>();
-const std::vector<mat_vec_op_inputs<int64_t>> mv_input_i64 = get_mv_inputs<int64_t>();
-
-template <typename T>
-struct Add1Vec {
-  static constexpr bool useTwoVectors = false;
-  HDI T operator()(T a, T b) const { return a + b; };
-};
-template <typename T>
-struct Add2Vec {
-  static constexpr bool useTwoVectors = true;
-  HDI T operator()(T a, T b, T c) const { return a + b + c; };
-};
-
-RAFT_BENCH_REGISTER((mat_vec_op<Add1Vec<float>, float, int>), "", mv_input_i32);
-RAFT_BENCH_REGISTER((mat_vec_op<Add1Vec<double>, double, int>), "", mv_input_i32);
-RAFT_BENCH_REGISTER((mat_vec_op<Add2Vec<float>, float, int>), "", mv_input_i32);
-RAFT_BENCH_REGISTER((mat_vec_op<Add2Vec<double>, double, int>), "", mv_input_i32);
-RAFT_BENCH_REGISTER((mat_vec_op<Add1Vec<float>, float, int64_t>), "", mv_input_i64);
-RAFT_BENCH_REGISTER((mat_vec_op<Add1Vec<double>, double, int64_t>), "", mv_input_i64);
-RAFT_BENCH_REGISTER((mat_vec_op<Add2Vec<float>, float, int64_t>), "", mv_input_i64);
-RAFT_BENCH_REGISTER((mat_vec_op<Add2Vec<double>, double, int64_t>), "", mv_input_i64);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/norm.cu b/cpp/bench/prims/linalg/norm.cu
deleted file mode 100644
index 1db23e4ca..000000000
--- a/cpp/bench/prims/linalg/norm.cu
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/matrix_vector_op.cuh>
-#include <raft/linalg/norm.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/cuda_utils.cuh>
-#include <raft/util/itertools.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-template <typename IdxT>
-struct norm_input {
-  IdxT rows, cols;
-};
-
-template <typename IdxT>
-inline auto operator<<(std::ostream& os, const norm_input<IdxT>& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols;
-  return os;
-}
-
-template <typename T, typename IdxT>
-struct rowNorm : public fixture {
-  rowNorm(const norm_input<IdxT>& p) : params(p), in(p.rows * p.cols, stream), dots(p.rows, stream)
-  {
-    raft::random::RngState rng{1234};
-    raft::random::uniform(handle, rng, in.data(), p.rows * p.cols, (T)-10.0, (T)10.0);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      auto input_view = raft::make_device_matrix_view<const T, IdxT, raft::row_major>(
-        in.data(), params.rows, params.cols);
-      auto output_view =
-        raft::make_device_vector_view<T, IdxT, raft::row_major>(dots.data(), params.rows);
-      raft::linalg::norm(handle,
-                         input_view,
-                         output_view,
-                         raft::linalg::L2Norm,
-                         raft::linalg::Apply::ALONG_ROWS,
-                         raft::sqrt_op());
-    });
-  }
-
- private:
-  norm_input<IdxT> params;
-  rmm::device_uvector<T> in, dots;
-};  // struct rowNorm
-
-const std::vector<norm_input<int>> norm_inputs_i32 =
-  raft::util::itertools::product<norm_input<int>>({10, 100, 1000, 10000, 100000},
-                                                  {16, 32, 64, 128, 256, 512, 1024});
-const std::vector<norm_input<int64_t>> norm_inputs_i64 =
-  raft::util::itertools::product<norm_input<int64_t>>({10, 100, 1000, 10000, 100000},
-                                                      {16, 32, 64, 128, 256, 512, 1024});
-
-RAFT_BENCH_REGISTER((rowNorm<float, int>), "", norm_inputs_i32);
-RAFT_BENCH_REGISTER((rowNorm<double, int>), "", norm_inputs_i32);
-RAFT_BENCH_REGISTER((rowNorm<float, int64_t>), "", norm_inputs_i64);
-RAFT_BENCH_REGISTER((rowNorm<double, int64_t>), "", norm_inputs_i64);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/normalize.cu b/cpp/bench/prims/linalg/normalize.cu
deleted file mode 100644
index 91319e774..000000000
--- a/cpp/bench/prims/linalg/normalize.cu
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/normalize.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/itertools.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-template <typename IdxT>
-struct normalize_input {
-  IdxT rows, cols;
-};
-
-template <typename IdxT>
-inline auto operator<<(std::ostream& os, const normalize_input<IdxT>& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols;
-  return os;
-}
-
-template <typename T, typename IdxT>
-struct rowNormalize : public fixture {
-  rowNormalize(const normalize_input<IdxT>& p)
-    : params(p), in(p.rows * p.cols, stream), out(p.rows * p.cols, stream)
-  {
-    raft::random::RngState rng{1234};
-    raft::random::uniform(handle, rng, in.data(), p.rows * p.cols, (T)-10.0, (T)10.0);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      auto input_view = raft::make_device_matrix_view<const T, IdxT, raft::row_major>(
-        in.data(), params.rows, params.cols);
-      auto output_view = raft::make_device_matrix_view<T, IdxT, raft::row_major>(
-        out.data(), params.rows, params.cols);
-      raft::linalg::row_normalize(handle, input_view, output_view, raft::linalg::L2Norm);
-    });
-  }
-
- private:
-  normalize_input<IdxT> params;
-  rmm::device_uvector<T> in, out;
-};  // struct rowNormalize
-
-const std::vector<normalize_input<int>> normalize_inputs_i32 =
-  raft::util::itertools::product<normalize_input<int>>(
-    {10, 100, 1000, 10000, 100000}, {8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384});
-const std::vector<normalize_input<int64_t>> normalize_inputs_i64 =
-  raft::util::itertools::product<normalize_input<int64_t>>(
-    {10, 100, 1000, 10000, 100000}, {8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384});
-
-RAFT_BENCH_REGISTER((rowNormalize<float, int>), "", normalize_inputs_i32);
-RAFT_BENCH_REGISTER((rowNormalize<double, int>), "", normalize_inputs_i32);
-RAFT_BENCH_REGISTER((rowNormalize<float, int64_t>), "", normalize_inputs_i64);
-RAFT_BENCH_REGISTER((rowNormalize<double, int64_t>), "", normalize_inputs_i64);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/reduce.cu b/cpp/bench/prims/linalg/reduce.cu
deleted file mode 100644
index cf41c5916..000000000
--- a/cpp/bench/prims/linalg/reduce.cu
+++ /dev/null
@@ -1,62 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/reduce.cuh>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-struct input_size {
-  int rows, cols;
-  bool along_rows;
-};
-
-template <typename T>
-struct reduce : public fixture {
-  reduce(bool along_rows, const input_size& p)
-    : input_size(p), along_rows(along_rows), in(p.rows * p.cols, stream), out(p.rows, stream)
-  {
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::linalg::reduce(
-        out.data(), in.data(), input_size.cols, input_size.rows, T(0.f), true, along_rows, stream);
-    });
-  }
-
- private:
-  bool along_rows;
-  input_size input_size;
-  rmm::device_uvector<T> in, out;
-};  // struct reduce
-
-const std::vector<input_size> kInputSizes{{8 * 1024, 1024},
-                                          {1024, 8 * 1024},
-                                          {8 * 1024, 8 * 1024},
-                                          {32 * 1024, 1024},
-                                          {1024, 32 * 1024},
-                                          {32 * 1024, 32 * 1024}};
-
-const std::vector<bool> kAlongRows{false, true};
-
-RAFT_BENCH_REGISTER(reduce<float>, "", kAlongRows, kInputSizes);
-RAFT_BENCH_REGISTER(reduce<double>, "", kAlongRows, kInputSizes);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/reduce_cols_by_key.cu b/cpp/bench/prims/linalg/reduce_cols_by_key.cu
deleted file mode 100644
index 1b584e80c..000000000
--- a/cpp/bench/prims/linalg/reduce_cols_by_key.cu
+++ /dev/null
@@ -1,78 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/reduce_cols_by_key.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/itertools.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-template <typename IdxT>
-struct rcbk_params {
-  IdxT rows, cols;
-  IdxT keys;
-};
-
-template <typename IdxT>
-inline auto operator<<(std::ostream& os, const rcbk_params<IdxT>& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols << "#" << p.keys;
-  return os;
-}
-
-template <typename T, typename KeyT, typename IdxT>
-struct reduce_cols_by_key : public fixture {
-  reduce_cols_by_key(const rcbk_params<IdxT>& p)
-    : params(p), in(p.rows * p.cols, stream), out(p.rows * p.keys, stream), keys(p.cols, stream)
-  {
-    raft::random::RngState rng{42};
-    raft::random::uniformInt(handle, rng, keys.data(), p.cols, (KeyT)0, (KeyT)p.keys);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      raft::linalg::reduce_cols_by_key(
-        in.data(), keys.data(), out.data(), params.rows, params.cols, params.keys, stream, false);
-    });
-  }
-
- protected:
-  rcbk_params<IdxT> params;
-  rmm::device_uvector<T> in, out;
-  rmm::device_uvector<KeyT> keys;
-};  // struct reduce_cols_by_key
-
-const std::vector<rcbk_params<int>> rcbk_inputs_i32 =
-  raft::util::itertools::product<rcbk_params<int>>(
-    {1, 10, 100, 1000}, {1000, 10000, 100000}, {8, 32, 128, 512, 2048});
-const std::vector<rcbk_params<int64_t>> rcbk_inputs_i64 =
-  raft::util::itertools::product<rcbk_params<int64_t>>(
-    {1, 10, 100, 1000}, {1000, 10000, 100000}, {8, 32, 128, 512, 2048});
-
-RAFT_BENCH_REGISTER((reduce_cols_by_key<float, uint32_t, int>), "", rcbk_inputs_i32);
-RAFT_BENCH_REGISTER((reduce_cols_by_key<double, uint32_t, int>), "", rcbk_inputs_i32);
-RAFT_BENCH_REGISTER((reduce_cols_by_key<float, uint32_t, int64_t>), "", rcbk_inputs_i64);
-RAFT_BENCH_REGISTER((reduce_cols_by_key<double, uint32_t, int64_t>), "", rcbk_inputs_i64);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/linalg/reduce_rows_by_key.cu b/cpp/bench/prims/linalg/reduce_rows_by_key.cu
deleted file mode 100644
index b68cefc27..000000000
--- a/cpp/bench/prims/linalg/reduce_rows_by_key.cu
+++ /dev/null
@@ -1,88 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/linalg/reduce_rows_by_key.cuh>
-#include <raft/random/rng.cuh>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::linalg {
-
-struct rrbk_params {
-  int64_t rows, cols;
-  int64_t keys;
-};
-
-template <typename T, typename KeyT>
-struct reduce_rows_by_key : public fixture {
-  reduce_rows_by_key(const rrbk_params& p)
-    : params(p),
-      in(p.rows * p.cols, stream),
-      out(p.keys * p.cols, stream),
-      keys(p.rows, stream),
-      workspace(p.rows, stream)
-  {
-    raft::random::RngState rng{42};
-    raft::random::uniformInt(handle, rng, keys.data(), p.rows, (KeyT)0, (KeyT)p.keys);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::linalg::reduce_rows_by_key(in.data(),
-                                       params.cols,
-                                       keys.data(),
-                                       workspace.data(),
-                                       params.rows,
-                                       params.cols,
-                                       params.keys,
-                                       out.data(),
-                                       stream,
-                                       false);
-    });
-  }
-
- protected:
-  rrbk_params params;
-  rmm::device_uvector<T> in, out;
-  rmm::device_uvector<KeyT> keys;
-  rmm::device_uvector<char> workspace;
-};  // struct reduce_rows_by_key
-
-const std::vector<rrbk_params> kInputSizes{
-  {10000, 128, 64},
-  {100000, 128, 64},
-  {1000000, 128, 64},
-  {10000000, 128, 64},
-  {10000, 128, 256},
-  {100000, 128, 256},
-  {1000000, 128, 256},
-  {10000000, 128, 256},
-  {10000, 128, 1024},
-  {100000, 128, 1024},
-  {1000000, 128, 1024},
-  {10000000, 128, 1024},
-  {10000, 128, 4096},
-  {100000, 128, 4096},
-  {1000000, 128, 4096},
-  {10000000, 128, 4096},
-};
-
-RAFT_BENCH_REGISTER((reduce_rows_by_key<float, uint32_t>), "", kInputSizes);
-RAFT_BENCH_REGISTER((reduce_rows_by_key<double, uint32_t>), "", kInputSizes);
-
-}  // namespace raft::bench::linalg
diff --git a/cpp/bench/prims/matrix/argmin.cu b/cpp/bench/prims/matrix/argmin.cu
deleted file mode 100644
index afee81aa0..000000000
--- a/cpp/bench/prims/matrix/argmin.cu
+++ /dev/null
@@ -1,67 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/core/resource/cuda_stream.hpp>
-#include <raft/matrix/argmin.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/itertools.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::matrix {
-
-template <typename IdxT>
-struct ArgminParams {
-  IdxT rows, cols;
-};
-
-template <typename T, typename OutT, typename IdxT>
-struct Argmin : public fixture {
-  Argmin(const ArgminParams<IdxT>& p) : params(p), matrix(this->handle), indices(this->handle) {}
-
-  void allocate_data(const ::benchmark::State& state) override
-  {
-    matrix  = raft::make_device_matrix<T, IdxT>(handle, params.rows, params.cols);
-    indices = raft::make_device_vector<OutT, IdxT>(handle, params.rows);
-
-    raft::random::RngState rng{1234};
-    raft::random::uniform(
-      handle, rng, matrix.data_handle(), params.rows * params.cols, T(-1), T(1));
-    resource::sync_stream(handle, stream);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::matrix::argmin(handle, raft::make_const_mdspan(matrix.view()), indices.view());
-    });
-  }
-
- private:
-  ArgminParams<IdxT> params;
-  raft::device_matrix<T, IdxT> matrix;
-  raft::device_vector<OutT, IdxT> indices;
-};  // struct Argmin
-
-const std::vector<ArgminParams<int64_t>> argmin_inputs_i64 =
-  raft::util::itertools::product<ArgminParams<int64_t>>({1000, 10000, 100000, 1000000, 10000000},
-                                                        {64, 128, 256, 512, 1024});
-
-RAFT_BENCH_REGISTER((Argmin<float, uint32_t, int64_t>), "", argmin_inputs_i64);
-RAFT_BENCH_REGISTER((Argmin<double, uint32_t, int64_t>), "", argmin_inputs_i64);
-
-}  // namespace raft::bench::matrix
diff --git a/cpp/bench/prims/matrix/gather.cu b/cpp/bench/prims/matrix/gather.cu
deleted file mode 100644
index 00a145ffa..000000000
--- a/cpp/bench/prims/matrix/gather.cu
+++ /dev/null
@@ -1,102 +0,0 @@
-/*
- * Copyright (c) 2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/core/resource/cuda_stream.hpp>
-#include <raft/matrix/gather.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/itertools.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::matrix {
-
-template <typename IdxT>
-struct GatherParams {
-  IdxT rows, cols, map_length;
-};
-
-template <typename IdxT>
-inline auto operator<<(std::ostream& os, const GatherParams<IdxT>& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols << "#" << p.map_length;
-  return os;
-}
-
-template <typename T, typename MapT, typename IdxT, bool Conditional = false>
-struct Gather : public fixture {
-  Gather(const GatherParams<IdxT>& p)
-    : params(p), matrix(this->handle), map(this->handle), out(this->handle), stencil(this->handle)
-  {
-  }
-
-  void allocate_data(const ::benchmark::State& state) override
-  {
-    matrix  = raft::make_device_matrix<T, IdxT>(handle, params.rows, params.cols);
-    map     = raft::make_device_vector<MapT, IdxT>(handle, params.map_length);
-    out     = raft::make_device_matrix<T, IdxT>(handle, params.map_length, params.cols);
-    stencil = raft::make_device_vector<T, IdxT>(handle, Conditional ? params.map_length : IdxT(0));
-
-    raft::random::RngState rng{1234};
-    raft::random::uniform(
-      handle, rng, matrix.data_handle(), params.rows * params.cols, T(-1), T(1));
-    raft::random::uniformInt(
-      handle, rng, map.data_handle(), params.map_length, (MapT)0, (MapT)params.rows);
-    if constexpr (Conditional) {
-      raft::random::uniform(handle, rng, stencil.data_handle(), params.map_length, T(-1), T(1));
-    }
-    resource::sync_stream(handle, stream);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      auto matrix_const_view = raft::make_const_mdspan(matrix.view());
-      auto map_const_view    = raft::make_const_mdspan(map.view());
-      if constexpr (Conditional) {
-        auto stencil_const_view = raft::make_const_mdspan(stencil.view());
-        auto pred_op            = raft::plug_const_op(T(0.0), raft::greater_op());
-        raft::matrix::gather_if(
-          handle, matrix_const_view, out.view(), map_const_view, stencil_const_view, pred_op);
-      } else {
-        raft::matrix::gather(handle, matrix_const_view, map_const_view, out.view());
-      }
-    });
-  }
-
- private:
-  GatherParams<IdxT> params;
-  raft::device_matrix<T, IdxT> matrix, out;
-  raft::device_vector<T, IdxT> stencil;
-  raft::device_vector<MapT, IdxT> map;
-};  // struct Gather
-
-template <typename T, typename MapT, typename IdxT>
-using GatherIf = Gather<T, MapT, IdxT, true>;
-
-const std::vector<GatherParams<int64_t>> gather_inputs_i64 =
-  raft::util::itertools::product<GatherParams<int64_t>>(
-    {1000000}, {10, 20, 50, 100, 200, 500}, {1000, 10000, 100000, 1000000});
-
-RAFT_BENCH_REGISTER((Gather<float, uint32_t, int64_t>), "", gather_inputs_i64);
-RAFT_BENCH_REGISTER((Gather<double, uint32_t, int64_t>), "", gather_inputs_i64);
-RAFT_BENCH_REGISTER((GatherIf<float, uint32_t, int64_t>), "", gather_inputs_i64);
-RAFT_BENCH_REGISTER((GatherIf<double, uint32_t, int64_t>), "", gather_inputs_i64);
-}  // namespace raft::bench::matrix
diff --git a/cpp/bench/prims/random/make_blobs.cu b/cpp/bench/prims/random/make_blobs.cu
deleted file mode 100644
index f43d914cf..000000000
--- a/cpp/bench/prims/random/make_blobs.cu
+++ /dev/null
@@ -1,87 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/random/make_blobs.cuh>
-#include <rmm/device_uvector.hpp>
-#include <vector>
-
-namespace raft::bench::random {
-struct make_blobs_inputs {
-  int rows, cols, clusters;
-  bool row_major;
-};  // struct make_blobs_inputs
-
-inline auto operator<<(std::ostream& os, const make_blobs_inputs& p) -> std::ostream&
-{
-  os << p.rows << "#" << p.cols << "#" << p.clusters << "#" << p.row_major;
-  return os;
-}
-
-template <typename T>
-struct make_blobs : public fixture {
-  make_blobs(const make_blobs_inputs& p)
-    : params(p), data(p.rows * p.cols, stream), labels(p.rows, stream)
-  {
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    std::ostringstream label_stream;
-    label_stream << params;
-    state.SetLabel(label_stream.str());
-
-    loop_on_state(state, [this]() {
-      raft::random::make_blobs(data.data(),
-                               labels.data(),
-                               params.rows,
-                               params.cols,
-                               params.clusters,
-                               this->stream,
-                               params.row_major);
-    });
-  }
-
- private:
-  make_blobs_inputs params;
-  rmm::device_uvector<T> data;
-  rmm::device_uvector<int> labels;
-};  // struct MakeBlobs
-
-static std::vector<make_blobs_inputs> get_make_blobs_input_vecs()
-{
-  std::vector<make_blobs_inputs> out;
-  make_blobs_inputs p;
-  for (auto rows : std::vector<int>{100000, 1000000}) {
-    for (auto cols : std::vector<int>{10, 100}) {
-      for (auto clusters : std::vector<int>{2, 10, 100}) {
-        p.rows      = rows;
-        p.cols      = cols;
-        p.clusters  = clusters;
-        p.row_major = true;
-        out.push_back(p);
-        p.row_major = false;
-        out.push_back(p);
-      }
-    }
-  }
-  return out;
-}
-
-RAFT_BENCH_REGISTER(make_blobs<float>, "", get_make_blobs_input_vecs());
-RAFT_BENCH_REGISTER(make_blobs<double>, "", get_make_blobs_input_vecs());
-
-}  // namespace raft::bench::random
diff --git a/cpp/bench/prims/random/permute.cu b/cpp/bench/prims/random/permute.cu
deleted file mode 100644
index 829cf4272..000000000
--- a/cpp/bench/prims/random/permute.cu
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/random/permute.cuh>
-#include <raft/random/rng.cuh>
-#include <raft/util/cudart_utils.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::random {
-
-struct permute_inputs {
-  int rows, cols;
-  bool needPerms, needShuffle, rowMajor;
-};  // struct permute_inputs
-
-template <typename T>
-struct permute : public fixture {
-  permute(const permute_inputs& p)
-    : params(p),
-      perms(p.needPerms ? p.rows : 0, stream),
-      out(p.rows * p.cols, stream),
-      in(p.rows * p.cols, stream)
-  {
-    raft::random::RngState r(123456ULL);
-    uniform(handle, r, in.data(), p.rows, T(-1.0), T(1.0));
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    raft::random::RngState r(123456ULL);
-    loop_on_state(state, [this, &r]() {
-      raft::random::permute(
-        perms.data(), out.data(), in.data(), params.cols, params.rows, params.rowMajor, stream);
-    });
-  }
-
- private:
-  raft::device_resources handle;
-  permute_inputs params;
-  rmm::device_uvector<T> out, in;
-  rmm::device_uvector<int> perms;
-};  // struct permute
-
-const std::vector<permute_inputs> permute_input_vecs = {
-  {32 * 1024, 128, true, true, true},
-  {1024 * 1024, 128, true, true, true},
-  {32 * 1024, 128 + 2, true, true, true},
-  {1024 * 1024, 128 + 2, true, true, true},
-  {32 * 1024, 128 + 1, true, true, true},
-  {1024 * 1024, 128 + 1, true, true, true},
-
-  {32 * 1024, 128, true, true, false},
-  {1024 * 1024, 128, true, true, false},
-  {32 * 1024, 128 + 2, true, true, false},
-  {1024 * 1024, 128 + 2, true, true, false},
-  {32 * 1024, 128 + 1, true, true, false},
-  {1024 * 1024, 128 + 1, true, true, false},
-
-};
-
-RAFT_BENCH_REGISTER(permute<float>, "", permute_input_vecs);
-RAFT_BENCH_REGISTER(permute<double>, "", permute_input_vecs);
-
-}  // namespace raft::bench::random
diff --git a/cpp/bench/prims/random/rng.cu b/cpp/bench/prims/random/rng.cu
deleted file mode 100644
index d15c9441d..000000000
--- a/cpp/bench/prims/random/rng.cu
+++ /dev/null
@@ -1,120 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <common/benchmark.hpp>
-#include <raft/random/rng.cuh>
-#include <raft/util/cudart_utils.hpp>
-
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::random {
-
-enum RandomType {
-  RNG_Normal,
-  RNG_LogNormal,
-  RNG_Uniform,
-  RNG_Gumbel,
-  RNG_Logistic,
-  RNG_Exp,
-  RNG_Rayleigh,
-  RNG_Laplace,
-  RNG_Fill
-};  // enum RandomType
-
-template <typename T>
-struct rng_inputs {
-  int len;
-  RandomType type;
-  raft::random::GeneratorType gtype;
-  T start, end;
-};  // struct rng_inputs
-
-template <typename T>
-struct rng : public fixture {
-  rng(const rng_inputs<T>& p) : params(p), ptr(p.len, stream) {}
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    raft::random::RngState r(123456ULL, params.gtype);
-    loop_on_state(state, [this, &r]() {
-      switch (params.type) {
-        case RNG_Normal: normal(handle, r, ptr.data(), params.len, params.start, params.end); break;
-        case RNG_LogNormal:
-          lognormal(handle, r, ptr.data(), params.len, params.start, params.end);
-          break;
-        case RNG_Uniform:
-          uniform(handle, r, ptr.data(), params.len, params.start, params.end);
-          break;
-        case RNG_Gumbel: gumbel(handle, r, ptr.data(), params.len, params.start, params.end); break;
-        case RNG_Logistic:
-          logistic(handle, r, ptr.data(), params.len, params.start, params.end);
-          break;
-        case RNG_Exp: exponential(handle, r, ptr.data(), params.len, params.start); break;
-        case RNG_Rayleigh: rayleigh(handle, r, ptr.data(), params.len, params.start); break;
-        case RNG_Laplace:
-          laplace(handle, r, ptr.data(), params.len, params.start, params.end);
-          break;
-        case RNG_Fill: fill(handle, r, ptr.data(), params.len, params.start); break;
-      };
-    });
-  }
-
- private:
-  rng_inputs<T> params;
-  rmm::device_uvector<T> ptr;
-};  // struct RngBench
-
-template <typename T>
-static std::vector<rng_inputs<T>> get_rng_input_vecs()
-{
-  using namespace raft::random;
-  return {
-    {1024 * 1024, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 + 2, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 2, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 2, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 + 1, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 1, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 1, RNG_Uniform, GenPhilox, T(-1.0), T(1.0)},
-
-    {1024 * 1024, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {1024 * 1024 + 2, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 2, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 2, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {1024 * 1024 + 1, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 1, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 1, RNG_Uniform, GenPC, T(-1.0), T(1.0)},
-
-    {1024 * 1024, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 + 2, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 2, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 2, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 + 1, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {32 * 1024 * 1024 + 1, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-    {1024 * 1024 * 1024 + 1, RNG_Fill, GenPhilox, T(-1.0), T(1.0)},
-  };
-}
-
-RAFT_BENCH_REGISTER(rng<float>, "", get_rng_input_vecs<float>());
-RAFT_BENCH_REGISTER(rng<double>, "", get_rng_input_vecs<double>());
-
-}  // namespace raft::bench::random
diff --git a/cpp/bench/prims/sparse/convert_csr.cu b/cpp/bench/prims/sparse/convert_csr.cu
deleted file mode 100644
index 634c749a5..000000000
--- a/cpp/bench/prims/sparse/convert_csr.cu
+++ /dev/null
@@ -1,137 +0,0 @@
-/*
- * Copyright (c) 2022-2023, NVIDIA CORPORATION.
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#include <stdio.h>
-
-#include <common/benchmark.hpp>
-#include <raft/sparse/convert/csr.cuh>
-#include <rmm/device_uvector.hpp>
-
-namespace raft::bench::sparse {
-
-template <typename index_t>
-struct bench_param {
-  index_t num_cols;
-  index_t num_rows;
-  index_t divisor;
-};
-
-template <typename index_t>
-RAFT_KERNEL init_adj_kernel(bool* adj, index_t num_rows, index_t num_cols, index_t divisor)
-{
-  index_t r = blockDim.y * blockIdx.y + threadIdx.y;
-  index_t c = blockDim.x * blockIdx.x + threadIdx.x;
-
-  for (; r < num_rows; r += gridDim.y * blockDim.y) {
-    for (; c < num_cols; c += gridDim.x * blockDim.x) {
-      adj[r * num_cols + c] = c % divisor == 0;
-    }
-  }
-}
-
-template <typename index_t>
-void init_adj(bool* adj, index_t num_rows, index_t num_cols, index_t divisor, cudaStream_t stream)
-{
-  // adj matrix: element a_ij is set to one if j is divisible by divisor.
-  dim3 block(32, 32);
-  const index_t max_y_grid_dim = 65535;
-  dim3 grid(num_cols / 32 + 1, (int)min(num_rows / 32 + 1, max_y_grid_dim));
-  init_adj_kernel<index_t><<<grid, block, 0, stream>>>(adj, num_rows, num_cols, divisor);
-  RAFT_CHECK_CUDA(stream);
-}
-
-template <typename index_t>
-struct bench_base : public fixture {
-  bench_base(const bench_param<index_t>& p)
-    : params(p),
-      handle(stream),
-      adj(p.num_rows * p.num_cols, stream),
-      row_ind(p.num_rows, stream),
-      row_ind_host(p.num_rows),
-      row_counters(p.num_rows, stream),
-      // col_ind is over-dimensioned because nnz is unknown at this point
-      col_ind(p.num_rows * p.num_cols, stream)
-  {
-    init_adj(adj.data(), p.num_rows, p.num_cols, p.divisor, stream);
-
-    std::vector<index_t> row_ind_host(p.num_rows);
-    for (size_t i = 0; i < row_ind_host.size(); ++i) {
-      size_t nnz_per_row = raft::ceildiv(p.num_cols, p.divisor);
-      row_ind_host[i]    = nnz_per_row * i;
-    }
-    raft::update_device(row_ind.data(), row_ind_host.data(), row_ind.size(), stream);
-  }
-
-  void run_benchmark(::benchmark::State& state) override
-  {
-    loop_on_state(state, [this]() {
-      raft::sparse::convert::adj_to_csr<index_t>(handle,
-                                                 adj.data(),
-                                                 row_ind.data(),
-                                                 params.num_rows,
-                                                 params.num_cols,
-                                                 row_counters.data(),
-                                                 col_ind.data());
-    });
-
-    // Estimate bandwidth:
-    index_t num_entries = params.num_rows * params.num_cols;
-    index_t bytes_read  = num_entries * sizeof(bool);
-    index_t bytes_write = num_entries / params.divisor * sizeof(index_t);
-
-    state.counters["BW"]      = benchmark::Counter(bytes_read + bytes_write,
-                                              benchmark::Counter::kIsIterationInvariantRate,
-                                              benchmark::Counter::OneK::kIs1024);
-    state.counters["BW read"] = benchmark::Counter(
-      bytes_read, benchmark::Counter::kIsIterationInvariantRate, benchmark::Counter::OneK::kIs1024);
-    state.counters["BW write"] = benchmark::Counter(bytes_write,
-                                                    benchmark::Counter::kIsIterationInvariantRate,
-                                                    benchmark::Counter::OneK::kIs1024);
-
-    state.counters["Fraction nz"] = benchmark::Counter(100.0 / ((double)params.divisor));
-    state.counters["Columns"]     = benchmark::Counter(params.num_cols);
-    state.counters["Rows"]        = benchmark::Counter(params.num_rows);
-  }
-
- protected:
-  raft::device_resources handle;
-  bench_param<index_t> params;
-  rmm::device_uvector<bool> adj;
-  rmm::device_uvector<index_t> row_ind;
-  std::vector<index_t> row_ind_host;
-  rmm::device_uvector<index_t> row_counters;
-  rmm::device_uvector<index_t> col_ind;
-};  // struct bench_base
-
-const int64_t num_cols = 1 << 30;
-
-const std::vector<bench_param<int64_t>> bench_params = {
-  {num_cols, 1, 8},
-  {num_cols >> 3, 1 << 3, 8},
-  {num_cols >> 6, 1 << 6, 8},
-
-  {num_cols, 1, 64},
-  {num_cols >> 3, 1 << 3, 64},
-  {num_cols >> 6, 1 << 6, 64},
-
-  {num_cols, 1, 2048},
-  {num_cols >> 3, 1 << 3, 2048},
-  {num_cols >> 6, 1 << 6, 2048},
-};
-
-RAFT_BENCH_REGISTER(bench_base<int64_t>, "", bench_params);
-
-}  // namespace raft::bench::sparse
diff --git a/cpp/cmake/thirdparty/get_raft.cmake b/cpp/cmake/thirdparty/get_raft.cmake
new file mode 100644
index 000000000..6128b5c43
--- /dev/null
+++ b/cpp/cmake/thirdparty/get_raft.cmake
@@ -0,0 +1,63 @@
+# =============================================================================
+# Copyright (c) 2023, NVIDIA CORPORATION.
+#
+# Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software distributed under the License
+# is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
+# or implied. See the License for the specific language governing permissions and limitations under
+# the License.
+
+# Use RAPIDS_VERSION from cmake/thirdparty/fetch_rapids.cmake
+set(RAFT_VERSION "${RAPIDS_VERSION}")
+set(RAFT_FORK "rapidsai")
+set(RAFT_PINNED_TAG "branch-${RAPIDS_VERSION}")
+
+function(find_and_configure_raft)
+    set(oneValueArgs VERSION FORK PINNED_TAG COMPILE_LIBRARY ENABLE_NVTX ENABLE_MNMG_DEPENDENCIES)
+    cmake_parse_arguments(PKG "${options}" "${oneValueArgs}"
+            "${multiValueArgs}" ${ARGN} )
+
+    set(RAFT_COMPONENTS "")
+    if(PKG_COMPILE_LIBRARY)
+        string(APPEND RAFT_COMPONENTS " compiled")
+    endif()
+
+    if(PKG_ENABLE_MNMG_DEPENDENCIES)
+        string(APPEND RAFT_COMPONENTS " distributed")
+    endif()
+
+    #-----------------------------------------------------
+    # Invoke CPM find_package()
+    #-----------------------------------------------------
+    rapids_cpm_find(raft ${PKG_VERSION}
+            GLOBAL_TARGETS      raft::raft
+            BUILD_EXPORT_SET    raft-template-exports
+            INSTALL_EXPORT_SET  raft-template-exports
+            COMPONENTS          ${RAFT_COMPONENTS}
+            CPM_ARGS
+            GIT_REPOSITORY https://github.com/${PKG_FORK}/raft.git
+            GIT_TAG        ${PKG_PINNED_TAG}
+            SOURCE_SUBDIR  cpp
+            OPTIONS
+            "BUILD_TESTS OFF"
+            "BUILD_PRIMS_BENCH OFF"
+            "BUILD_ANN_BENCH OFF"
+            "RAFT_NVTX   ${ENABLE_NVTX}"
+            "RAFT_COMPILE_LIBRARY ${PKG_COMPILE_LIBRARY}"
+            )
+endfunction()
+
+# Change pinned tag here to test a commit in CI
+# To use a different RAFT locally, set the CMake variable
+# CPM_raft_SOURCE=/path/to/local/raft
+find_and_configure_raft(VERSION  ${RAFT_VERSION}.00
+        FORK                     ${RAFT_FORK}
+        PINNED_TAG               ${RAFT_PINNED_TAG}
+        COMPILE_LIBRARY          ON
+        ENABLE_MNMG_DEPENDENCIES OFF
+        ENABLE_NVTX              OFF
+)
diff --git a/cpp/doxygen/Doxyfile b/cpp/doxygen/Doxyfile
index 3eb0763ea..0a2c7f8f6 100644
--- a/cpp/doxygen/Doxyfile
+++ b/cpp/doxygen/Doxyfile
@@ -32,7 +32,7 @@ DOXYFILE_ENCODING      = UTF-8
 # title of most generated pages and in a few other places.
 # The default value is: My Project.
 
-PROJECT_NAME           = "RAFT C++ API"
+PROJECT_NAME           = "cuVS C++ API"
 
 # The PROJECT_NUMBER tag can be used to enter a project or revision number. This
 # could be handy for archiving the generated documentation or if some version
@@ -879,27 +879,7 @@ RECURSIVE              = YES
 # Note that relative paths are relative to the directory from which doxygen is
 # run.
 
-EXCLUDE                = ../include/raft/sparse/linalg/symmetrize.hpp \
-                         ../include/raft/cache \
-                         ../include/raft/common \
-                         ../include/raft/lap \
-                         ../include/raft/sparse/selection \
-                         ../include/raft/sparse/csr.hpp \
-                         ../include/raft/linalg/lanczos.cuh \
-                         ../include/raft/linalg/lanczos.hpp \
-                         ../include/raft/util/cuda_utils.cuh \
-                         ../include/raft/util/cudart_utils.hpp \
-                         ../include/raft/util/device_atomics.cuh \
-                         ../include/raft/util/device_utils.cuh \
-                         ../include/raft/core/error.hpp \
-                         ../include/raft/core/handle.hpp \
-                         ../include/raft/util/integer_utils.hpp \
-                         ../include/raft/util/pow2_utils.cuh \
-                         ../include/raft/util/vectorized.cuh \
-                         ../include/raft/raft.hpp \
-                         ../include/raft/core/cudart_utils.hpp \
-                         ../include/raft/matrix/math.cuh \
-                         ../include/raft/matrix/matrix.cuh
+EXCLUDE                = ../include/cuvs/sparse/selection 
 
 # The EXCLUDE_SYMLINKS tag can be used to select whether or not files or
 # directories that are symbolic links (a Unix file system feature) are excluded
diff --git a/cpp/doxygen/main_page.md b/cpp/doxygen/main_page.md
index ff0c7820c..d002df001 100644
--- a/cpp/doxygen/main_page.md
+++ b/cpp/doxygen/main_page.md
@@ -1,22 +1,3 @@
-# libraft
+# libcuvs
 
-RAFT contains fundamental widely-used algorithms and primitives for data science, graph and machine learning. The algorithms are CUDA-accelerated and form building-blocks for rapidly composing analytics.
-
-By taking a primitives-based approach to algorithm development, RAFT
-- accelerates algorithm construction time,
-- reduces the maintenance burden by maximizing reuse across projects, and
-- centralizes core reusable computations, allowing future optimizations to benefit all algorithms that use them.
-
-While not exhaustive, the following general categories help summarize the accelerated functions in RAFT:
-
-#####
-| Category | Examples |
-| --- | --- |
-| **Data Formats** | sparse & dense, conversions, data generation |
-| **Dense Linear Algebra** | matrix arithmetic, norms, factorization, least squares, svd & eigenvalue problems |
-| **Spatial** | pairwise distances, nearest neighbors, neighborhood graph construction |
-| **Sparse Operations** | linear algebra, eigenvalue problems, slicing, symmetrization, labeling |
-| **Basic Clustering** | spectral clustering, hierarchical clustering, k-means |
-| **Solvers** | combinatorial optimization, iterative solvers |
-| **Statistics** | sampling, moments and summary statistics, metrics |
-| **Distributed Tools** | multi-node multi-gpu infrastructure |
+cuVS is a library for vector search on the GPU
\ No newline at end of file
diff --git a/cpp/src/distance/distance.cu b/cpp/src/distance/distance.cu
index 8c9460831..0a35a335e 100644
--- a/cpp/src/distance/distance.cu
+++ b/cpp/src/distance/distance.cu
@@ -14,8 +14,8 @@
  * limitations under the License.
  */
 
-#include <raft/distance/detail/kernels/rbf_fin_op.cuh>  // rbf_fin_op
-#include <raft/distance/distance-inl.cuh>
+#include <cuvs/distance/detail/kernels/rbf_fin_op.cuh>  // rbf_fin_op
+#include <cuvs/distance/distance-inl.cuh>
 
 /*
  * Hierarchy of instantiations:
diff --git a/cpp/src/matrix/detail/select_k_double_int64_t.cu b/cpp/src/matrix/detail/select_k_double_int64_t.cu
index c75a5b526..c3949149f 100644
--- a/cpp/src/matrix/detail/select_k_double_int64_t.cu
+++ b/cpp/src/matrix/detail/select_k_double_int64_t.cu
@@ -16,7 +16,7 @@
 
 #include <raft/matrix/detail/select_k-inl.cuh>
 
-#define instantiate_raft_matrix_detail_select_k(T, IdxT)                            \
+#define instantiate_cuvs_matrix_detail_select_k(T, IdxT)                            \
   template void raft::matrix::detail::select_k(raft::resources const& handle,       \
                                                const T* in_val,                     \
                                                const IdxT* in_idx,                  \
@@ -29,6 +29,6 @@
                                                rmm::mr::device_memory_resource* mr, \
                                                bool sorted)
 
-instantiate_raft_matrix_detail_select_k(double, int64_t);
+instantiate_cuvs_matrix_detail_select_k(double, int64_t);
 
-#undef instantiate_raft_matrix_detail_select_k
+#undef instantiate_cuvs_matrix_detail_select_k
diff --git a/cpp/test/CMakeLists.txt b/cpp/test/CMakeLists.txt
index 6c03da8d7..1e01d59b2 100644
--- a/cpp/test/CMakeLists.txt
+++ b/cpp/test/CMakeLists.txt
@@ -25,40 +25,38 @@ function(ConfigureTest)
   set(oneValueArgs NAME GPUS PERCENT)
   set(multiValueArgs PATH TARGETS CONFIGURATIONS)
 
-  cmake_parse_arguments(_RAFT_TEST "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
-  if(NOT DEFINED _RAFT_TEST_GPUS AND NOT DEFINED _RAFT_TEST_PERCENT)
-    set(_RAFT_TEST_GPUS 1)
-    set(_RAFT_TEST_PERCENT 30)
+  cmake_parse_arguments(_CUVS_TEST "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
+  if(NOT DEFINED _CUVS_TEST_GPUS AND NOT DEFINED _CUVS_TEST_PERCENT)
+    set(_CUVS_TEST_GPUS 1)
+    set(_CUVS_TEST_PERCENT 30)
   endif()
-  if(NOT DEFINED _RAFT_TEST_GPUS)
-    set(_RAFT_TEST_GPUS 1)
+  if(NOT DEFINED _CUVS_TEST_GPUS)
+    set(_CUVS_TEST_GPUS 1)
   endif()
-  if(NOT DEFINED _RAFT_TEST_PERCENT)
-    set(_RAFT_TEST_PERCENT 100)
+  if(NOT DEFINED _CUVS_TEST_PERCENT)
+    set(_CUVS_TEST_PERCENT 100)
   endif()
 
-  if(_RAFT_TEST_NOCUDA)
-    set(TEST_NAME "${_RAFT_TEST_NAME}_NOCUDA")
+  if(_CUVS_TEST_NOCUDA)
+    set(TEST_NAME "${_CUVS_TEST_NAME}_NOCUDA")
   else()
-    set(TEST_NAME ${_RAFT_TEST_NAME})
+    set(TEST_NAME ${_CUVS_TEST_NAME})
   endif()
 
-  add_executable(${TEST_NAME} ${_RAFT_TEST_PATH})
+  add_executable(${TEST_NAME} ${_CUVS_TEST_PATH})
   target_link_libraries(
     ${TEST_NAME}
-    PRIVATE raft
-            raft_internal
-            $<$<BOOL:${_RAFT_TEST_LIB}>:raft::compiled>
+    PRIVATE raft::raft
             GTest::gtest
             GTest::gtest_main
             Threads::Threads
-            ${RAFT_CTK_MATH_DEPENDENCIES}
+            ${CUVS_CTK_MATH_DEPENDENCIES}
             $<TARGET_NAME_IF_EXISTS:OpenMP::OpenMP_CXX>
             $<TARGET_NAME_IF_EXISTS:conda_env>
   )
   set_target_properties(
     ${TEST_NAME}
-    PROPERTIES RUNTIME_OUTPUT_DIRECTORY "$<BUILD_INTERFACE:${RAFT_BINARY_DIR}/gtests>"
+    PROPERTIES RUNTIME_OUTPUT_DIRECTORY "$<BUILD_INTERFACE:${CUVS_BINARY_DIR}/gtests>"
                INSTALL_RPATH "\$ORIGIN/../../../lib"
                CXX_STANDARD 17
                CXX_STANDARD_REQUIRED ON
@@ -66,23 +64,23 @@ function(ConfigureTest)
                CUDA_STANDARD_REQUIRED ON
   )
   target_compile_options(
-    ${TEST_NAME} PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${RAFT_CXX_FLAGS}>"
-                         "$<$<COMPILE_LANGUAGE:CUDA>:${RAFT_CUDA_FLAGS}>"
+    ${TEST_NAME} PRIVATE "$<$<COMPILE_LANGUAGE:CXX>:${CUVS_CXX_FLAGS}>"
+                         "$<$<COMPILE_LANGUAGE:CUDA>:${CUVS_CUDA_FLAGS}>"
   )
-  if(_RAFT_TEST_EXPLICIT_INSTANTIATE_ONLY)
-    target_compile_definitions(${TEST_NAME} PRIVATE "RAFT_EXPLICIT_INSTANTIATE_ONLY")
+  if(_CUVS_TEST_EXPLICIT_INSTANTIATE_ONLY)
+    target_compile_definitions(${TEST_NAME} PRIVATE "CUVS_EXPLICIT_INSTANTIATE_ONLY")
   endif()
-  if(_RAFT_TEST_NOCUDA)
-    target_compile_definitions(${TEST_NAME} PRIVATE "RAFT_DISABLE_CUDA")
+  if(_CUVS_TEST_NOCUDA)
+    target_compile_definitions(${TEST_NAME} PRIVATE "CUVS_DISABLE_CUDA")
   endif()
 
-  target_include_directories(${TEST_NAME} PUBLIC "$<BUILD_INTERFACE:${RAFT_SOURCE_DIR}/test>")
+  target_include_directories(${TEST_NAME} PUBLIC "$<BUILD_INTERFACE:${CUVS_SOURCE_DIR}/test>")
 
   rapids_test_add(
     NAME ${TEST_NAME}
     COMMAND ${TEST_NAME}
-    GPUS ${_RAFT_TEST_GPUS}
-    PERCENT ${_RAFT_TEST_PERCENT}
+    GPUS ${_CUVS_TEST_GPUS}
+    PERCENT ${_CUVS_TEST_PERCENT}
     INSTALL_COMPONENT_SET testing
   )
 endfunction()
@@ -108,43 +106,6 @@ if(BUILD_TESTS)
     EXPLICIT_INSTANTIATE_ONLY
   )
 
-  ConfigureTest(
-    NAME
-    CORE_TEST
-    PATH
-    test/core/bitset.cu
-    test/core/device_resources_manager.cpp
-    test/core/device_setter.cpp
-    test/core/logger.cpp
-    test/core/math_device.cu
-    test/core/math_host.cpp
-    test/core/operators_device.cu
-    test/core/operators_host.cpp
-    test/core/handle.cpp
-    test/core/interruptible.cu
-    test/core/nvtx.cpp
-    test/core/mdarray.cu
-    test/core/mdspan_copy.cpp
-    test/core/mdspan_copy.cu
-    test/core/mdspan_utils.cu
-    test/core/numpy_serializer.cu
-    test/core/memory_type.cpp
-    test/core/sparse_matrix.cu
-    test/core/sparse_matrix.cpp
-    test/core/span.cpp
-    test/core/span.cu
-    test/core/stream_view.cpp
-    test/core/temporary_device_buffer.cu
-    test/test.cpp
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureTest(
-    NAME CORE_TEST PATH test/core/stream_view.cpp test/core/mdspan_copy.cpp LIB
-    EXPLICIT_INSTANTIATE_ONLY NOCUDA
-  )
-
   ConfigureTest(
     NAME
     DISTANCE_TEST
@@ -174,168 +135,6 @@ if(BUILD_TESTS)
     EXPLICIT_INSTANTIATE_ONLY
   )
 
-  list(
-    APPEND
-    EXT_HEADER_TEST_SOURCES
-    test/ext_headers/raft_neighbors_brute_force.cu
-    test/ext_headers/raft_distance_distance.cu
-    test/ext_headers/raft_distance_detail_pairwise_matrix_dispatch.cu
-    test/ext_headers/raft_matrix_detail_select_k.cu
-    test/ext_headers/raft_neighbors_ball_cover.cu
-    test/ext_headers/raft_spatial_knn_detail_fused_l2_knn.cu
-    test/ext_headers/raft_distance_fused_l2_nn.cu
-    test/ext_headers/raft_neighbors_ivf_pq.cu
-    test/ext_headers/raft_util_memory_pool.cpp
-    test/ext_headers/raft_neighbors_ivf_flat.cu
-    test/ext_headers/raft_core_logger.cpp
-    test/ext_headers/raft_neighbors_refine.cu
-    test/ext_headers/raft_neighbors_detail_ivf_flat_search.cu
-    test/ext_headers/raft_neighbors_detail_selection_faiss.cu
-    test/ext_headers/raft_linalg_detail_coalesced_reduction.cu
-    test/ext_headers/raft_spatial_knn_detail_ball_cover_registers.cu
-    test/ext_headers/raft_neighbors_detail_ivf_flat_interleaved_scan.cu
-    test/ext_headers/raft_neighbors_detail_ivf_pq_compute_similarity.cu
-  )
-
-  # Test that the split headers compile in isolation with:
-  #
-  # * EXT_HEADERS_TEST_COMPILED_EXPLICIT: RAFT_COMPILED, RAFT_EXPLICIT_INSTANTIATE_ONLY defined
-  # * EXT_HEADERS_TEST_COMPILED_IMPLICIT: RAFT_COMPILED defined
-  # * EXT_HEADERS_TEST_IMPLICIT:          no macros defined.
-  ConfigureTest(
-    NAME EXT_HEADERS_TEST_COMPILED_EXPLICIT PATH ${EXT_HEADER_TEST_SOURCES} LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-  ConfigureTest(NAME EXT_HEADERS_TEST_COMPILED_IMPLICIT PATH ${EXT_HEADER_TEST_SOURCES} LIB)
-  ConfigureTest(NAME EXT_HEADERS_TEST_IMPLICIT PATH ${EXT_HEADER_TEST_SOURCES})
-
-  ConfigureTest(NAME LABEL_TEST PATH test/label/label.cu test/label/merge_labels.cu)
-
-  ConfigureTest(
-    NAME
-    LINALG_TEST
-    PATH
-    test/linalg/add.cu
-    test/linalg/axpy.cu
-    test/linalg/binary_op.cu
-    test/linalg/cholesky_r1.cu
-    test/linalg/coalesced_reduction.cu
-    test/linalg/divide.cu
-    test/linalg/dot.cu
-    test/linalg/eig.cu
-    test/linalg/eig_sel.cu
-    test/linalg/gemm_layout.cu
-    test/linalg/gemv.cu
-    test/linalg/map.cu
-    test/linalg/map_then_reduce.cu
-    test/linalg/matrix_vector.cu
-    test/linalg/matrix_vector_op.cu
-    test/linalg/mean_squared_error.cu
-    test/linalg/multiply.cu
-    test/linalg/norm.cu
-    test/linalg/normalize.cu
-    test/linalg/power.cu
-    test/linalg/randomized_svd.cu
-    test/linalg/reduce.cu
-    test/linalg/reduce_cols_by_key.cu
-    test/linalg/reduce_rows_by_key.cu
-    test/linalg/rsvd.cu
-    test/linalg/sqrt.cu
-    test/linalg/strided_reduction.cu
-    test/linalg/subtract.cu
-    test/linalg/svd.cu
-    test/linalg/ternary_op.cu
-    test/linalg/transpose.cu
-    test/linalg/unary_op.cu
-  )
-
-  ConfigureTest(
-    NAME
-    MATRIX_TEST
-    PATH
-    test/matrix/argmax.cu
-    test/matrix/argmin.cu
-    test/matrix/columnSort.cu
-    test/matrix/diagonal.cu
-    test/matrix/gather.cu
-    test/matrix/scatter.cu
-    test/matrix/eye.cu
-    test/matrix/linewise_op.cu
-    test/matrix/math.cu
-    test/matrix/matrix.cu
-    test/matrix/norm.cu
-    test/matrix/reverse.cu
-    test/matrix/slice.cu
-    test/matrix/triangular.cu
-    test/sparse/spectral_matrix.cu
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureTest(NAME MATRIX_SELECT_TEST PATH test/matrix/select_k.cu LIB EXPLICIT_INSTANTIATE_ONLY)
-
-  ConfigureTest(
-    NAME MATRIX_SELECT_LARGE_TEST PATH test/matrix/select_large_k.cu LIB EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureTest(
-    NAME
-    RANDOM_TEST
-    PATH
-    test/random/make_blobs.cu
-    test/random/make_regression.cu
-    test/random/multi_variable_gaussian.cu
-    test/random/rng_pcg_host_api.cu
-    test/random/permute.cu
-    test/random/rng.cu
-    test/random/rng_discrete.cu
-    test/random/rng_int.cu
-    test/random/rmat_rectangular_generator.cu
-    test/random/sample_without_replacement.cu
-  )
-
-  ConfigureTest(
-    NAME SOLVERS_TEST PATH test/cluster/cluster_solvers_deprecated.cu test/linalg/eigen_solvers.cu
-    test/lap/lap.cu test/sparse/mst.cu LIB EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureTest(
-    NAME
-    SPARSE_TEST
-    PATH
-    test/sparse/add.cu
-    test/sparse/convert_coo.cu
-    test/sparse/convert_csr.cu
-    test/sparse/csr_row_slice.cu
-    test/sparse/csr_to_dense.cu
-    test/sparse/csr_transpose.cu
-    test/sparse/degree.cu
-    test/sparse/filter.cu
-    test/sparse/norm.cu
-    test/sparse/normalize.cu
-    test/sparse/reduce.cu
-    test/sparse/row_op.cu
-    test/sparse/sort.cu
-    test/sparse/spgemmi.cu
-    test/sparse/symmetrize.cu
-  )
-
-  ConfigureTest(
-    NAME SPARSE_DIST_TEST PATH test/sparse/dist_coo_spmv.cu test/sparse/distance.cu
-    test/sparse/gram.cu LIB EXPLICIT_INSTANTIATE_ONLY
-  )
-
-  ConfigureTest(
-    NAME
-    SPARSE_NEIGHBORS_TEST
-    PATH
-    test/sparse/neighbors/cross_component_nn.cu
-    test/sparse/neighbors/brute_force.cu
-    test/sparse/neighbors/knn_graph.cu
-    LIB
-    EXPLICIT_INSTANTIATE_ONLY
-  )
-
   ConfigureTest(
     NAME
     NEIGHBORS_TEST
@@ -421,9 +220,6 @@ if(BUILD_TESTS)
     NAME
     STATS_TEST
     PATH
-    test/stats/accuracy.cu
-    test/stats/adjusted_rand_index.cu
-    test/stats/completeness_score.cu
     test/stats/contingencyMatrix.cu
     test/stats/cov.cu
     test/stats/dispersion.cu
@@ -432,38 +228,15 @@ if(BUILD_TESTS)
     test/stats/homogeneity_score.cu
     test/stats/information_criterion.cu
     test/stats/kl_divergence.cu
-    test/stats/mean.cu
-    test/stats/meanvar.cu
-    test/stats/mean_center.cu
-    test/stats/minmax.cu
     test/stats/mutual_info_score.cu
     test/stats/neighborhood_recall.cu
     test/stats/r2_score.cu
     test/stats/rand_index.cu
-    test/stats/regression_metrics.cu
     test/stats/silhouette_score.cu
-    test/stats/stddev.cu
-    test/stats/sum.cu
     test/stats/trustworthiness.cu
-    test/stats/weighted_mean.cu
-    test/stats/v_measure.cu
     LIB
     EXPLICIT_INSTANTIATE_ONLY
   )
-
-  ConfigureTest(
-    NAME
-    UTILS_TEST
-    PATH
-    test/core/seive.cu
-    test/util/bitonic_sort.cu
-    test/util/cudart_utils.cpp
-    test/util/device_atomics.cu
-    test/util/integer_utils.cpp
-    test/util/integer_utils.cu
-    test/util/pow2_utils.cu
-    test/util/reduction.cu
-  )
 endif()
 
 # ##################################################################################################
diff --git a/dependencies.yaml b/dependencies.yaml
index 8be7a24dd..7d6c318d6 100644
--- a/dependencies.yaml
+++ b/dependencies.yaml
@@ -81,16 +81,16 @@ files:
       - test_python_common
       - test_cuvs
       - cupy
-  py_build_cuda_ann_bench:
+  py_build_cuvs_bench:
     output: pyproject
-    pyproject_dir: python/cuda-ann-bench
+    pyproject_dir: python/cuvs-bench
     extras:
       table: build-system
     includes:
       - build_wheels
-  py_run_cuda_ann_bench:
+  py_run_cuvs_bench:
     output: pyproject
-    pyproject_dir: python/cuda-ann-bench
+    pyproject_dir: python/cuvs-bench
     extras:
       table: project
     includes:
diff --git a/python/cuda-ann-bench/LICENSE b/python/cuvs-bench/LICENSE
similarity index 100%
rename from python/cuda-ann-bench/LICENSE
rename to python/cuvs-bench/LICENSE
diff --git a/python/cuda-ann-bench/pyproject.toml b/python/cuvs-bench/pyproject.toml
similarity index 97%
rename from python/cuda-ann-bench/pyproject.toml
rename to python/cuvs-bench/pyproject.toml
index ad2306179..508a9f94c 100644
--- a/python/cuda-ann-bench/pyproject.toml
+++ b/python/cuvs-bench/pyproject.toml
@@ -10,7 +10,7 @@ requires = [
 [project]
 name = "cuvs-ann-bench"
 version = "24.02.00"
-description = "CUDA ANN benchmarks"
+description = "cuVS benchmarks"
 authors = [
     { name = "NVIDIA Corporation" },
 ]
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/__init__.py b/python/cuvs-bench/src/cuvs-bench/__init__.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/__init__.py
rename to python/cuvs-bench/src/cuvs-bench/__init__.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/constraints/__init__.py b/python/cuvs-bench/src/cuvs-bench/constraints/__init__.py
similarity index 87%
rename from python/cuda-ann-bench/src/cuda-ann-bench/constraints/__init__.py
rename to python/cuvs-bench/src/cuvs-bench/constraints/__init__.py
index d827c3d1e..2de3ac98e 100644
--- a/python/cuda-ann-bench/src/cuda-ann-bench/constraints/__init__.py
+++ b/python/cuvs-bench/src/cuvs-bench/constraints/__init__.py
@@ -16,13 +16,13 @@
 DTYPE_SIZES = {"float": 4, "half": 2, "fp8": 1}
 
 
-def raft_ivf_pq_build_constraints(params, dims):
+def cuvs_ivf_pq_build_constraints(params, dims):
     if "pq_dim" in params:
         return params["pq_dim"] <= dims
     return True
 
 
-def raft_ivf_pq_search_constraints(params, build_params, k, batch_size):
+def cuvs_ivf_pq_search_constraints(params, build_params, k, batch_size):
     ret = True
     if "internalDistanceDtype" in params and "smemLutDtype" in params:
         ret = (
@@ -35,7 +35,7 @@ def raft_ivf_pq_search_constraints(params, build_params, k, batch_size):
     return ret
 
 
-def raft_cagra_search_constraints(params, build_params, k, batch_size):
+def cuvs_cagra_search_constraints(params, build_params, k, batch_size):
     if "itopk" in params:
         return params["itopk"] >= k
 
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/data_export/__main__.py b/python/cuvs-bench/src/cuvs-bench/data_export/__main__.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/data_export/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/data_export/__main__.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/generate_groundtruth/__main__.py b/python/cuvs-bench/src/cuvs-bench/generate_groundtruth/__main__.py
similarity index 99%
rename from python/cuda-ann-bench/src/cuda-ann-bench/generate_groundtruth/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/generate_groundtruth/__main__.py
index f4d97edea..1e3dc9859 100644
--- a/python/cuda-ann-bench/src/cuda-ann-bench/generate_groundtruth/__main__.py
+++ b/python/cuvs-bench/src/cuvs-bench/generate_groundtruth/__main__.py
@@ -22,9 +22,10 @@
 import numpy as np
 import rmm
 from pylibraft.common import DeviceResources
-from pylibraft.neighbors.brute_force import knn
 from rmm.allocators.cupy import rmm_cupy_allocator
 
+from cuvs.neighbors.brute_force import knn
+
 from .utils import memmap_bin_file, suffix_from_dtype, write_bin
 
 
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/generate_groundtruth/utils.py b/python/cuvs-bench/src/cuvs-bench/generate_groundtruth/utils.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/generate_groundtruth/utils.py
rename to python/cuvs-bench/src/cuvs-bench/generate_groundtruth/utils.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/__main__.py b/python/cuvs-bench/src/cuvs-bench/get_dataset/__main__.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/get_dataset/__main__.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/fbin_to_f16bin.py b/python/cuvs-bench/src/cuvs-bench/get_dataset/fbin_to_f16bin.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/fbin_to_f16bin.py
rename to python/cuvs-bench/src/cuvs-bench/get_dataset/fbin_to_f16bin.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/hdf5_to_fbin.py b/python/cuvs-bench/src/cuvs-bench/get_dataset/hdf5_to_fbin.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/get_dataset/hdf5_to_fbin.py
rename to python/cuvs-bench/src/cuvs-bench/get_dataset/hdf5_to_fbin.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/plot/__main__.py b/python/cuvs-bench/src/cuvs-bench/plot/__main__.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/plot/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/plot/__main__.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/__main__.py b/python/cuvs-bench/src/cuvs-bench/run/__main__.py
similarity index 99%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/run/__main__.py
index 4611f3926..20d5010c1 100644
--- a/python/cuda-ann-bench/src/cuda-ann-bench/run/__main__.py
+++ b/python/cuvs-bench/src/cuvs-bench/run/__main__.py
@@ -325,7 +325,7 @@ def main():
     parser.add_argument(
         "--algo-groups",
         help='add comma separated <algorithm>.<group> to run. \
-              Example usage: "--algo-groups=raft_cagra.large,hnswlib.large"',
+              Example usage: "--algo-groups=cuvs_cagra.large,hnswlib.large"',
     )
     parser.add_argument(
         "-f",
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/algos.yaml b/python/cuvs-bench/src/cuvs-bench/run/algos.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/algos.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/algos.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/faiss_gpu_ivf_flat.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/faiss_gpu_ivf_flat.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/faiss_gpu_ivf_flat.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/faiss_gpu_ivf_flat.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/faiss_gpu_ivf_pq.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/faiss_gpu_ivf_pq.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/faiss_gpu_ivf_pq.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/faiss_gpu_ivf_pq.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/hnswlib.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/hnswlib.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/hnswlib.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/hnswlib.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_cagra.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_cagra.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_cagra.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_cagra.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_cagra_hnswlib.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_cagra_hnswlib.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_cagra_hnswlib.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_cagra_hnswlib.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_ivf_flat.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_ivf_flat.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_ivf_flat.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_ivf_flat.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_ivf_pq.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_ivf_pq.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/algos/raft_ivf_pq.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/algos/raft_ivf_pq.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/bigann-100M.json b/python/cuvs-bench/src/cuvs-bench/run/conf/bigann-100M.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/bigann-100M.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/bigann-100M.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/datasets.yaml b/python/cuvs-bench/src/cuvs-bench/run/conf/datasets.yaml
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/datasets.yaml
rename to python/cuvs-bench/src/cuvs-bench/run/conf/datasets.yaml
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-100M.json b/python/cuvs-bench/src/cuvs-bench/run/conf/deep-100M.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-100M.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/deep-100M.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-1B.json b/python/cuvs-bench/src/cuvs-bench/run/conf/deep-1B.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-1B.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/deep-1B.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-image-96-inner.json b/python/cuvs-bench/src/cuvs-bench/run/conf/deep-image-96-inner.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/deep-image-96-inner.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/deep-image-96-inner.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/fashion-mnist-784-euclidean.json b/python/cuvs-bench/src/cuvs-bench/run/conf/fashion-mnist-784-euclidean.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/fashion-mnist-784-euclidean.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/fashion-mnist-784-euclidean.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/gist-960-euclidean.json b/python/cuvs-bench/src/cuvs-bench/run/conf/gist-960-euclidean.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/gist-960-euclidean.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/gist-960-euclidean.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-100-angular.json b/python/cuvs-bench/src/cuvs-bench/run/conf/glove-100-angular.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-100-angular.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/glove-100-angular.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-100-inner.json b/python/cuvs-bench/src/cuvs-bench/run/conf/glove-100-inner.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-100-inner.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/glove-100-inner.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-50-angular.json b/python/cuvs-bench/src/cuvs-bench/run/conf/glove-50-angular.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-50-angular.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/glove-50-angular.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-50-inner.json b/python/cuvs-bench/src/cuvs-bench/run/conf/glove-50-inner.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/glove-50-inner.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/glove-50-inner.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/lastfm-65-angular.json b/python/cuvs-bench/src/cuvs-bench/run/conf/lastfm-65-angular.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/lastfm-65-angular.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/lastfm-65-angular.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/mnist-784-euclidean.json b/python/cuvs-bench/src/cuvs-bench/run/conf/mnist-784-euclidean.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/mnist-784-euclidean.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/mnist-784-euclidean.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/nytimes-256-angular.json b/python/cuvs-bench/src/cuvs-bench/run/conf/nytimes-256-angular.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/nytimes-256-angular.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/nytimes-256-angular.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/nytimes-256-inner.json b/python/cuvs-bench/src/cuvs-bench/run/conf/nytimes-256-inner.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/nytimes-256-inner.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/nytimes-256-inner.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/sift-128-euclidean.json b/python/cuvs-bench/src/cuvs-bench/run/conf/sift-128-euclidean.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/sift-128-euclidean.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/sift-128-euclidean.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_10M.json b/python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_10M.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_10M.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_10M.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_1M.json b/python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_1M.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_1M.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_1M.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_88M.json b/python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_88M.json
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/run/conf/wiki_all_88M.json
rename to python/cuvs-bench/src/cuvs-bench/run/conf/wiki_all_88M.json
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/split_groundtruth/__main__.py b/python/cuvs-bench/src/cuvs-bench/split_groundtruth/__main__.py
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/split_groundtruth/__main__.py
rename to python/cuvs-bench/src/cuvs-bench/split_groundtruth/__main__.py
diff --git a/python/cuda-ann-bench/src/cuda-ann-bench/split_groundtruth/split_groundtruth.pl b/python/cuvs-bench/src/cuvs-bench/split_groundtruth/split_groundtruth.pl
similarity index 100%
rename from python/cuda-ann-bench/src/cuda-ann-bench/split_groundtruth/split_groundtruth.pl
rename to python/cuvs-bench/src/cuvs-bench/split_groundtruth/split_groundtruth.pl