Merge remote-tracking branch 'origin/shared_memory' into 2D_tiling

SC-SGS · Nov 20, 2023 · caf75a0 · caf75a0
2 parents e4f9f14 + 0c5f5cb
commit caf75a0
Show file tree

Hide file tree

Showing 303 changed files with 22,588 additions and 11,724 deletions.
diff --git a/.clang-tidy b/.clang-tidy
@@ -0,0 +1,47 @@
+# Generated from CLion Inspection settings
+---
+Checks: '-*,
+cert-*,
+misc-*,
+mpi-*,
+cppcoreguidelines-*,
+-cppcoreguidelines-avoid-magic-numbers,
+-cppcoreguidelines-pro-bounds-pointer-arithmetic,
+google-default-arguments,
+google-runtime-operator,
+google-explicit-constructor,
+hicpp-multiway-paths-covered,
+hicpp-exception-baseclass,
+modernize-*,
+-modernize-use-trailing-return-type,
+-modernize-avoid-c-arrays,
+-modernize-use-using,
+-modernize-use-default-member-init,
+-modernize-macro-to-enum,
+portability-simd-intrinsics,
+readability-*,
+-readability-redundant-preprocessor,
+-readability-named-parameter,
+-readability-function-size,
+-readability-simplify-boolean-expr,
+-readability-identifier-length,
+-readability-duplicate-include,
+-readability-magic-numbers,
+-readability-braces-around-statements,
+-readability-redundant-member-init,
+-readability-suspicious-call-argument,
+-readability-qualified-auto,
+-readability-isolate-declaration,
+-readability-uppercase-literal-suffix,
+-readability-container-data-pointer,
+-readability-else-after-return,
+-readability-redundant-access-specifiers,
+-readability-function-cognitive-complexity,
+-readability-implicit-bool-conversion,
+-readability-container-contains,
+-readability-identifier-naming,
+bugprone-*'
+
+CheckOptions:
+  - key:    misc-non-private-member-variables-in-classes.IgnoreClassesWithAllMemberVariablesBeingPublic
+    value:  true
diff --git a/.github/workflows/msvc_windows.yml b/.github/workflows/msvc_windows.yml
@@ -1,29 +1,31 @@
-name: Windows CPU
+name: Windows MSVC CPU
 on: push
 jobs:
   MSVC-Test:
     runs-on: windows-latest
     steps:
-      - name: "Install MSVC 14.29"
-        uses: ilammy/[email protected]
-        with:
-          toolset: 14.29
-      - name: "Install cmake 3.22.2"
-        uses: lukka/[email protected]
-      - name: "Clone Step"
+      - name: "Install latest MSVC compiler"
+        uses: ilammy/[email protected]
+      - name: "Install cmake 3.25.2"
+        uses: lukka/[email protected]
+      - name: "Clone the PLSSVM repository into PLSSVM/"
         uses: actions/checkout@v2
         with:
           path: PLSSVM
-      - name: "CMake Step"
+      - name: "Install Python dependencies"
+        uses: py-actions/py-dependency-install@v4
+        with:
+          path: "PLSSVM/.github/workflows/requirements.txt"
+      - name: "Configure PLSSVM using CMake"
         run: |
           mkdir PLSSVM/build
           cd PLSSVM/build
-          cmake -DCMAKE_BUILD_TYPE=Debug -DPLSSVM_TARGET_PLATFORMS="cpu" -DPLSSVM_ENABLE_TESTING=ON -DPLSSVM_GENERATE_TEST_FILE=OFF -DPLSSVM_ENABLE_LTO=OFF -DPLSSVM_ENABLE_ASSERTS=ON ..
-      - name: "Build Step"
+          cmake -DCMAKE_BUILD_TYPE=Release -DPLSSVM_TARGET_PLATFORMS="cpu" -DPLSSVM_ENABLE_TESTING=ON -DPLSSVM_TEST_FILE_NUM_DATA_POINTS=50 -DPLSSVM_TEST_FILE_NUM_FEATURES=20 -DPLSSVM_ENABLE_LTO=OFF -DPLSSVM_ENABLE_ASSERTS=ON ..
+      - name: "Build PLSSVM"
         run: |
           cd PLSSVM/build
           cmake --build .
-      - name: "Test Step"
+      - name: "Run tests"
         run: |
           cd PLSSVM/build
-          ctest --output-on-failure -C Debug
+          ctest -j 2 -E ".*Exceptions\.|SourceLocation" --output-on-failure -C Release
diff --git a/.github/workflows/requirements.txt b/.github/workflows/requirements.txt
@@ -0,0 +1,5 @@
+### required python packages for the Windows MSVC tests
+argparse
+scikit-learn
+numpy
+humanize
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -85,21 +85,28 @@ target_include_directories(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
 )
 
 ## set library cxx standard
+if ($ENV{CLION_IDE})
+    # somehow necessary for CLion to detect the C++ standard as 17
+    set(CMAKE_CXX_STANDARD 17)
+endif ()
 target_compile_features(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC cxx_std_17)
 ## additional base library compile options
 target_compile_options(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
         $<$<COMPILE_LANGUAGE:CXX>:$<$<CXX_COMPILER_ID:GNU,Clang>:-Wall -Wextra -Wdouble-promotion -fno-common -Wshadow -Wcast-qual
-        -Wnull-dereference -Wctor-dtor-privacy -Wnon-virtual-dtor -Wextra-semi -Wunreachable-code -Wuninitialized
+        -Wnull-dereference -Wctor-dtor-privacy -Wnon-virtual-dtor -Wextra-semi -Wunreachable-code -Wuninitialized -Wno-ctor-dtor-privacy
         -fPIC>
         $<$<COMPILE_LANG_AND_ID:CXX,GNU>:-Wsuggest-override -Wstrict-null-sentinel -Wlogical-op -Wduplicated-branches -Wimplicit-fallthrough=5>
         $<$<COMPILE_LANG_AND_ID:CXX,Clang>:-Wdocumentation -Wmost>
-        $<$<COMPILE_LANG_AND_ID:CXX,MSVC>:/W4 /bigobj /wd4459>>
+        $<$<COMPILE_LANG_AND_ID:CXX,MSVC>:/W4 /bigobj /wd4459 /Zc:lambda>>
 )
 
 ## nvcc doesn't recognize -Werror=??? option, so only set it when using a CXX compiler
 target_compile_options(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
         $<$<COMPILE_LANG_AND_ID:CXX,GNU,Clang>:-Werror=switch -fstrict-enums>
-        $<$<COMPILE_LANG_AND_ID:CXX,MSVC>:/we4062 /wd4005>
+        $<$<COMPILE_LANG_AND_ID:CXX,MSVC>:/we4062 /wd4005 /wd4702 /wd4849 /wd4127>
+        # /wd4702: prevent warning for unreachable code in tests (the tests that currently skip the cg_streaming and cg_implicit tests)
+        # /wd4849: ignore "OpenMP 'reduction' clause ignored in 'simd' directive" -> no SIMD clause currently effective in MSVC
+        # /wd4127: ignore "conditional expression is constant" from {fmt} ranges.h header
 )
 ## enable additional optimization flags only in RELEASE mode
 target_compile_options(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
@@ -110,13 +117,14 @@ target_compile_definitions(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
         $<$<COMPILE_LANG_AND_ID:CXX,MSVC>:NOMINMAX>
 )
 
-## enable standard library debug modes if compiled in debug
-if (CMAKE_BUILD_TYPE MATCHES Debug)
+option(PLSSVM_ENABLE_STL_DEBUG_MODE "Enables the debug modes for the STL implementations. Note: changes the ABI!" OFF)
+# GCC standard library (libstdc++): _GLIBCXX_DEBUG
+# LLVM standard library (libc++): LIBCXX_ENABLE_DEBUG_MODE
+# MSVC standard library: _ITERATOR_DEBUG_LEVEL
+set(PLSSVM_STL_DEBUG_MODE_FLAGS "-D_GLIBCXX_DEBUG -DLIBCXX_ENABLE_DEBUG_MODE -D_ITERATOR_DEBUG_LEVEL")
+if (PLSSVM_ENABLE_STL_DEBUG_MODE)
     message(STATUS "Enable standard library debug modes.")
-    # GCC standard library (libstdc++): _GLIBCXX_DEBUG
-    # LLVM standard library (libc++): LIBCXX_ENABLE_DEBUG_MODE
-    # MSVC standard library: _ITERATOR_DEBUG_LEVEL
-    target_compile_definitions(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC _GLIBCXX_DEBUG LIBCXX_ENABLE_DEBUG_MODE _ITERATOR_DEBUG_LEVEL)
+    target_compile_definitions(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC ${PLSSVM_STL_DEBUG_MODE_FLAGS})
 endif ()
 
 
@@ -233,23 +241,27 @@ else ()
 endif ()
 
 ## try finding fmt
-find_package(fmt 10.0.0 QUIET)
+find_package(fmt 10.1.1 QUIET)
 if (fmt_FOUND)
     message(STATUS "Found package fmt.")
 else ()
-    message(STATUS "Couldn't find package fmt. Building from source ...")
-    set(PLSSVM_fmt_VERSION 10.0.0)
+    message(STATUS "Couldn't find package fmt. Building from source ...")  # TODO: change to next release; bug fix commit
+    set(PLSSVM_fmt_VERSION d9063baf227882da0f48c761abcbb08247eb1296)
+    if (PLSSVM_ENABLE_STL_DEBUG_MODE)
+        set(CMAKE_CXX_FLAGS_OLD "${CMAKE_CXX_FLAGS}")
+        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${PLSSVM_STL_DEBUG_MODE_FLAGS}")
+    endif ()
     # fetch string formatting library fmt
     FetchContent_Declare(fmt
             GIT_REPOSITORY https://github.com/fmtlib/fmt.git
             GIT_TAG ${PLSSVM_fmt_VERSION}
-            GIT_SHALLOW TRUE
+#            GIT_SHALLOW TRUE
             QUIET
 
-            set (FMT_TEST OFF CACHE BOOL "" FORCE)
-            set (FMT_DOC OFF CACHE BOOL "" FORCE)
-            set (FMT_INSTALL OFF CACHE BOOL "" FORCE)
-            set (FMT_SYSTEM_HEADERS ON CACHE BOOL "" FORCE)
+#            set (FMT_TEST OFF CACHE BOOL "" FORCE)
+#            set (FMT_DOC OFF CACHE BOOL "" FORCE)
+#            set (FMT_INSTALL OFF CACHE BOOL "" FORCE)
+#            set (FMT_SYSTEM_HEADERS ON CACHE BOOL "" FORCE)
     )
     FetchContent_MakeAvailable(fmt)
     set_property(TARGET fmt PROPERTY POSITION_INDEPENDENT_CODE ON)
@@ -261,6 +273,9 @@ else ()
     # append fmt to installed targets if build from source
     list(APPEND PLSSVM_TARGETS_TO_INSTALL "fmt")
     message(STATUS "Installing {fmt} version ${PLSSVM_fmt_VERSION}.")
+    if (PLSSVM_ENABLE_STL_DEBUG_MODE)
+        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS_OLD}")
+    endif ()
 endif ()
 target_link_libraries(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC fmt::fmt)
 
@@ -688,7 +703,7 @@ choose the SYCL implementation to be used in the SYCL backend: ${PLSSVM_SYCL_BAC
     set(PLSSVM_SYCL_KERNEL_INVOCATION_TYPE_MANPAGE_ENTRY "
 .TP
 .B --sycl_kernel_invocation_type
-choose the kernel invocation type when using SYCL as backend: automatic|nd_range|hierarchical (default: automatic)
+choose the kernel invocation type when using SYCL as backend: automatic|nd_range (default: automatic)
 ")
 endif ()
 set(PLSSVM_SYCL_MANPAGE_ENTRY "${PLSSVM_SYCL_KERNEL_INVOCATION_TYPE_MANPAGE_ENTRY}${PLSSVM_SYCL_IMPLEMENTATION_TYPE_MANPAGE_ENTRY}")

diff --git a/README.md b/README.md
@@ -53,10 +53,10 @@ General dependencies:
 
 - a C++17 capable compiler (e.g. [`gcc`](https://gcc.gnu.org/) or [`clang`](https://clang.llvm.org/))
 - [CMake](https://cmake.org/) 3.21 or newer
-- [cxxopts ≥ v3.0.0](https://github.com/jarro2783/cxxopts), [fast_float](https://github.com/fastfloat/fast_float), [{fmt} ≥ v8.1.1](https://github.com/fmtlib/fmt), and [igor](https://github.com/bluescarni/igor) (all four are automatically build during the CMake configuration if they couldn't be found using the respective `find_package` call)
+- [cxxopts ≥ v3.1.1](https://github.com/jarro2783/cxxopts), [fast_float ≥ v3.10.0](https://github.com/fastfloat/fast_float), [{fmt} ≥ v10.1.1](https://github.com/fmtlib/fmt), and [igor](https://github.com/bluescarni/igor) (all four are automatically build during the CMake configuration if they couldn't be found using the respective `find_package` call)
 - [GoogleTest ≥ v1.11.0](https://github.com/google/googletest) if testing is enabled (automatically build during the CMake configuration if `find_package(GTest)` wasn't successful)
 - [doxygen](https://www.doxygen.nl/index.html) if documentation generation is enabled
-- [Pybind11 ≥ v2.10.3](https://github.com/pybind/pybind11) if Python bindings are enabled
+- [Pybind11 ≥ v2.11.1](https://github.com/pybind/pybind11) if Python bindings are enabled
 - [OpenMP](https://www.openmp.org/) 4.0 or newer (optional) to speed-up library utilities (like file parsing)
 - multiple Python modules used in the utility scripts, to install all modules use `pip install --user -r install/python_requirements.txt`
 
@@ -88,6 +88,12 @@ Additional dependencies if `PLSSVM_ENABLE_TESTING` and `PLSSVM_GENERATE_TEST_FIL
 
 ### Building
 
+Download and install all *necessary* Python3 dependencies:
+
+```bash
+pip install -r install/python_requirements.txt
+```
+
 Building the library can be done using the normal CMake approach:
 
 ```bash
@@ -198,6 +204,7 @@ The `[optional_options]` can be one or multiple of:
 - `PLSSVM_ENABLE_PERFORMANCE_TRACKING`: enable gathering performance characteristics for the three executables using YAML files; example Python3 scripts to perform performance measurements and to process the resulting YAML files can be found in the `utility_scripts/` directory (requires the Python3 modules [wrapt-timeout-decorator](https://pypi.org/project/wrapt-timeout-decorator/), [`pyyaml`](https://pyyaml.org/), and [`pint`](https://pint.readthedocs.io/en/stable/))
 - `PLSSVM_ENABLE_TESTING=ON|OFF` (default: `ON`): enable testing using GoogleTest and ctest
 - `PLSSVM_ENABLE_LANGUAGE_BINDINGS=ON|OFF` (default: `OFF`): enable language bindings
+- `PLSSVM_STL_DEBUG_MODE_FLAGS=ON|OFF` (default: `OFF`): enable STL debug modes (**note**: changes the ABI!)
 
 If `PLSSVM_ENABLE_TESTING` is set to `ON`, the following options can also be set:
 
@@ -229,9 +236,9 @@ To use DPC++ for SYCL simply set the `CMAKE_CXX_COMPILER` to the respective DPC+
 
 If the SYCL implementation is DPC++ the following additional options are available:
 
-- `PLSSVM_SYCL_BACKEND_DPCPP_USE_LEVEL_ZERO` (default: `ON`): use DPC++'s Level-Zero backend instead of its OpenCL backend
-- `PLSSVM_SYCL_BACKEND_DPCPP_GPU_AMD_USE_HIP` (default: `ON`): use DPC++'s HIP backend instead of its OpenCL backend for AMD GPUs
 - `PLSSVM_SYCL_BACKEND_DPCPP_ENABLE_AOT` (default: `ON`): enable Ahead-of-Time (AOT) compilation for the specified target platforms
+- `PLSSVM_SYCL_BACKEND_DPCPP_USE_LEVEL_ZERO` (default: `ON`): use DPC++'s Level-Zero backend instead of its OpenCL backend **(only available if a CPU or Intel GPU is targeted)**
+- `PLSSVM_SYCL_BACKEND_DPCPP_GPU_AMD_USE_HIP` (default: `ON`): use DPC++'s HIP backend instead of its OpenCL backend for AMD GPUs **(only available if an AMD GPU is targeted)**
 
 If the SYCL implementation is hipSYCL the following additional option is available:
 
@@ -345,12 +352,12 @@ Usage:
   -c, --cost arg                set the parameter C (default: 1)
   -e, --epsilon arg             set the tolerance of termination criterion (default: 0.001)
   -i, --max_iter arg            set the maximum number of CG iterations (default: num_features)
-  -s, --solver arg              choose the solver: automatic|cg_explicit|cg_streaming|cg_implicit (default: automatic)
+  -l, --solver arg              choose the solver: automatic|cg_explicit|cg_streaming|cg_implicit (default: automatic)
   -a, --classification arg      the classification strategy to use for multi-class classification: oaa|oao (default: oaa)
   -b, --backend arg             choose the backend: automatic|openmp|cuda|hip|opencl|sycl (default: automatic)
   -p, --target_platform arg     choose the target platform: automatic|cpu|gpu_nvidia|gpu_amd|gpu_intel (default: automatic)
       --sycl_kernel_invocation_type arg
-                                choose the kernel invocation type when using SYCL as backend: automatic|nd_range|hierarchical (default: automatic)
+                                choose the kernel invocation type when using SYCL as backend: automatic|nd_range (default: automatic)
       --sycl_implementation_type arg
                                 choose the SYCL implementation to be used in the SYCL backend: automatic|dpcpp|hipsycl (default: automatic)
       --performance_tracking arg
@@ -404,7 +411,7 @@ The `--target_platform=automatic` option works for the different backends as fol
 - `SYCL`: tries to find available devices in the following order: NVIDIA GPUs 🠦 AMD GPUs 🠦 Intel GPUs 🠦 CPU
 
 The `--sycl_kernel_invocation_type` and `--sycl_implementation_type` flags are only used if the `--backend` is `sycl`, otherwise a warning is emitted on `stderr`.
-If the `--sycl_kernel_invocation_type` is `automatic`, the `nd_range` invocation type is always used, except for hipSYCL on CPUs where the hierarchical formulation is used instead (if hipSYCL wasn't build with `omp.accelerated`).
+If the `--sycl_kernel_invocation_type` is `automatic`, the `nd_range` invocation type is currently always used.
 If the `--sycl_implementation_type` is `automatic`, the used SYCL implementation is determined by the `PLSSVM_SYCL_BACKEND_PREFERRED_IMPLEMENTATION` cmake flag.
 
 ### Predicting

diff --git a/bindings/Python/CMakeLists.txt b/bindings/Python/CMakeLists.txt
@@ -14,19 +14,19 @@ if (pybind11_FOUND)
     message(STATUS "Found package pybind11.")
 else ()
     message(STATUS "Couldn't find package pybind11. Building from source ...")
-    set(PLSSVM_pybind11_VERSION v2.10.3)
+    set(PLSSVM_pybind11_VERSION v2.11.1)
     # fetch pybind11 library for creating Python bindings
     FetchContent_Declare(pybind11
             GIT_REPOSITORY https://github.com/pybind/pybind11.git
             GIT_TAG ${PLSSVM_pybind11_VERSION}
             GIT_SHALLOW TRUE
             QUIET
-            )
+    )
     FetchContent_MakeAvailable(pybind11)
     target_include_directories(${PLSSVM_BASE_LIBRARY_NAME} PUBLIC
             $<BUILD_INTERFACE:${pybind11_SOURCE_DIR}/include>
             $<INSTALL_INTERFACE:include>
-            )
+    )
     message(STATUS "Installed pybind11 version ${PLSSVM_pybind11_VERSION}.")
 endif ()
 
@@ -53,7 +53,7 @@ set(PLSSVM_PYTHON_BINDINGS_SOURCES
         ${CMAKE_CURRENT_LIST_DIR}/sklearn.cpp
 
         ${CMAKE_CURRENT_LIST_DIR}/main.cpp
-        )
+)
 # set backends specific source files if the respective backend is used
 if (TARGET ${PLSSVM_OPENMP_BACKEND_LIBRARY_NAME})
     list(APPEND PLSSVM_PYTHON_BINDINGS_SOURCES ${CMAKE_CURRENT_LIST_DIR}/backends/openmp_csvm.cpp)