Merge remote-tracking branch 'origin/main' into hierarchy_dimensions

NVIDIA · May 3, 2024 · 1bc2270 · 1bc2270
2 parents d425e48 + c910b16
commit 1bc2270
Show file tree

Hide file tree

Showing 979 changed files with 30,112 additions and 15,809 deletions.
diff --git a/.clang-format b/.clang-format
@@ -111,6 +111,9 @@ IncludeCategories:
   - Regex:           '^<(cuda/std/detail/__config|cub/config.cuh|thrust/detail/config.h|thrust/system/cuda/config.h)'
     Priority:            0
     SortPriority:        0
+  - Regex:           '^<cuda/std/detail/libcxx/include/__pragma_push>'
+    Priority:            6
+    SortPriority:        5
   - Regex:           '^<cub/'
     Priority:            2
     SortPriority:        1

diff --git a/.devcontainer/cuda11.1-gcc6/devcontainer.json b/.devcontainer/cuda11.1-gcc6/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda11.1-gcc7/devcontainer.json b/.devcontainer/cuda11.1-gcc7/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda11.1-gcc8/devcontainer.json b/.devcontainer/cuda11.1-gcc8/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda11.1-gcc9/devcontainer.json b/.devcontainer/cuda11.1-gcc9/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda11.1-llvm9/devcontainer.json b/.devcontainer/cuda11.1-llvm9/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda11.8-gcc11/devcontainer.json b/.devcontainer/cuda11.8-gcc11/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc10/devcontainer.json b/.devcontainer/cuda12.4-gcc10/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc11/devcontainer.json b/.devcontainer/cuda12.4-gcc11/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc12/devcontainer.json b/.devcontainer/cuda12.4-gcc12/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc7/devcontainer.json b/.devcontainer/cuda12.4-gcc7/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc8/devcontainer.json b/.devcontainer/cuda12.4-gcc8/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-gcc9/devcontainer.json b/.devcontainer/cuda12.4-gcc9/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm10/devcontainer.json b/.devcontainer/cuda12.4-llvm10/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm11/devcontainer.json b/.devcontainer/cuda12.4-llvm11/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm12/devcontainer.json b/.devcontainer/cuda12.4-llvm12/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm13/devcontainer.json b/.devcontainer/cuda12.4-llvm13/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm14/devcontainer.json b/.devcontainer/cuda12.4-llvm14/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm15/devcontainer.json b/.devcontainer/cuda12.4-llvm15/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm16/devcontainer.json b/.devcontainer/cuda12.4-llvm16/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-llvm9/devcontainer.json b/.devcontainer/cuda12.4-llvm9/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/cuda12.4-oneapi2023.2.0/devcontainer.json b/.devcontainer/cuda12.4-oneapi2023.2.0/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json
@@ -31,7 +31,9 @@
     "vscode": {
       "extensions": [
         "llvm-vs-code-extensions.vscode-clangd",
-        "xaver.clang-format"
+        "xaver.clang-format",
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cmake-tools"
       ],
       "settings": {
         "editor.defaultFormatter": "xaver.clang-format",

diff --git a/.vscode/launch.json b/.vscode/launch.json
@@ -0,0 +1,15 @@
+{
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "(cuda-gdb) Launch",
+            "type": "cuda-gdb",
+            "request": "launch",
+            "program": "${command:cmake.launchTargetPath}",
+            "stopAtEntry": true,
+            "cwd": "${workspaceFolder}",
+            "initCommands": ["set cuda api_failures stop"]
+
+        }
+    ]
+}
diff --git a/CMakePresets.json b/CMakePresets.json
@@ -64,6 +64,16 @@
         "CudaNext_ENABLE_DIALECT_CPP20": true
       }
     },
+    {
+      "name": "all-dev-debug",
+      "displayName": "all-dev debug",
+      "inherits": "all-dev",
+      "cacheVariables":{
+        "CCCL_ENABLE_BENCHMARKS": false,
+        "CMAKE_BUILD_TYPE": "Debug",
+        "CMAKE_CUDA_FLAGS": "-G"
+      }
+    },
     {
       "name": "libcudacxx-codegen",
       "displayName": "libcu++: codegen",
@@ -314,6 +324,10 @@
       "name": "all-dev",
       "configurePreset": "all-dev"
     },
+    {
+      "name": "all-dev-debug",
+      "configurePreset": "all-dev-debug"
+    },
     {
       "name": "libcudacxx-codegen",
       "configurePreset": "libcudacxx-codegen",
@@ -445,6 +459,11 @@
       "configurePreset": "all-dev",
       "inherits": "base"
     },
+    {
+      "name": "all-dev-debug",
+      "configurePreset": "all-dev-debug",
+      "inherits": "all-dev"
+    },
     {
       "name": "libcudacxx-ctest-base",
       "hidden": true,

diff --git a/ci/build_common.sh b/ci/build_common.sh
@@ -167,7 +167,7 @@ function configure_preset()
     local GROUP_NAME="🛠️  CMake Configure ${BUILD_NAME}"
 
     pushd .. > /dev/null
-    run_command "$GROUP_NAME" cmake --preset=$PRESET --log-level=VERBOSE "${GLOBAL_CMAKE_OPTIONS[@]}" $CMAKE_OPTIONS
+    run_command "$GROUP_NAME" cmake --preset=$PRESET --log-level=VERBOSE $CMAKE_OPTIONS "${GLOBAL_CMAKE_OPTIONS[@]}"
     status=$?
     popd > /dev/null
     return $status

diff --git a/ci/update_version.sh b/ci/update_version.sh
@@ -26,7 +26,7 @@ if [ -z "$major" ] || [ -z "$minor" ] || [ -z "$patch" ]; then
 fi
 
 # Version file paths
-CCCL_VERSION_FILE="libcudacxx/include/cuda/std/detail/libcxx/include/__cccl/version.h"
+CCCL_VERSION_FILE="libcudacxx/include/cuda/std/__cccl/version.h"
 THRUST_VERSION_FILE="thrust/thrust/version.h"
 CUB_VERSION_FILE="cub/cub/version.cuh"
 CCCL_CMAKE_VERSION_FILE="lib/cmake/cccl/cccl-config-version.cmake"

diff --git a/cub/cmake/CubInstallRules.cmake b/cub/cmake/CubInstallRules.cmake
@@ -8,6 +8,7 @@ install(DIRECTORY "${CUB_SOURCE_DIR}/cub"
   DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
   FILES_MATCHING
     PATTERN "*.cuh"
+    PATTERN "*.hpp"
 )
 
 install(DIRECTORY "${CUB_SOURCE_DIR}/cub/cmake/"

diff --git a/cub/cmake/header_test.in b/cub/cmake/header_test.in
@@ -46,6 +46,11 @@
 //#define min(...) CUB_MACRO_CHECK('min', windows.h)
 //#define max(...) CUB_MACRO_CHECK('max', windows.h)
 
+#ifdef _WIN32
+// On Windows, make sure any include of Windows.h (e.g. via NVTX) does not define the checked macros
+#  define WIN32_LEAN_AND_MEAN
+#endif // _WIN32
+
 // termios.h conflicts (NVIDIA/thrust#1547)
 #define B0 CUB_MACRO_CHECK("B0", termios.h)
 

diff --git a/cub/cub/cmake/cub-config-version.cmake b/cub/cub/cmake/cub-config-version.cmake
@@ -2,7 +2,7 @@
 include("${CMAKE_CURRENT_LIST_DIR}/cub-header-search.cmake")
 
 set(CUB_VERSION_MAJOR 2)
-set(CUB_VERSION_MINOR 4)
+set(CUB_VERSION_MINOR 5)
 set(CUB_VERSION_PATCH 0)
 set(CUB_VERSION_TWEAK 0)
 set(CUB_VERSION "${CUB_VERSION_MAJOR}.${CUB_VERSION_MINOR}.${CUB_VERSION_PATCH}.${CUB_VERSION_TWEAK}")