Merge pull request #7 from SC-SGS/reduce_data

Reduce data

Author: breyerml

.jenkins/Jenkinsfile-multigpu-tests

@@ -56,7 +56,7 @@ pipeline {
                             \\"state\\": \\"pending\\",
                             \\"context\\": \\"jenkins-ctest-multigpu\\",
                             \\"description\\": \\"Jenkins CI Job: jenkins-ctest-multigpu\\",
-                            \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Multibranch-Github/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
+                            \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Github-Multigpu/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
                     }"
                 '''
             }
@@ -127,7 +127,7 @@ pipeline {
                                             srun -w argon-gtx -N 1 -n 1 -t 01:00:00 -D /scratch/jenkins/plssvm/${BUILD_TAG}/PLSSVM --gres=gpu:2 bash -c "\
                                             module load cuda &&\
                                             cd build/Release &&\
-                                            ctest -j4 --no-compress-output -T Test --timeout 600; \
+                                            ctest -j4 --no-compress-output -T Test --timeout 1200; \
                                             returncode=$? && \
                                             cp -r Testing /data/argon-fs/sgs/jenkins/workspace/$(basename ${WORKSPACE})/${BUILD_TAG}/Testing &&\
                                             exit $returncode"
@@ -168,7 +168,7 @@ pipeline {
                                             module use /home/breyerml/.modulefiles/ &&\
                                             module load pcsgs05/hipsycl &&\
                                             cd build/Release_hip &&\
-                                            ctest -j4 --no-compress-output -T Test --timeout 600; \
+                                            ctest -j4 --no-compress-output -T Test --timeout 1200; \
                                             returncode=$? && \
                                             cp -r Testing /data/argon-fs/sgs/jenkins/workspace/$(basename ${WORKSPACE})/${BUILD_TAG}/Testing_hip && \
                                             exit $returncode"
@@ -209,7 +209,7 @@ pipeline {
                                             module use /home/breyerml/.modulefiles/ &&\
                                             module load pcsgs05/dpcpp_rt &&\
                                             cd build/Release_dpcpp &&\
-                                            ctest -j4 --no-compress-output -T Test --timeout 600; \
+                                            ctest -j4 --no-compress-output -T Test --timeout 1200; \
                                             returncode=$? && \
                                             cp -r Testing /data/argon-fs/sgs/jenkins/workspace/$(basename ${WORKSPACE})/${BUILD_TAG}/Testing_dpcpp &&\
                                             exit $returncode"
@@ -261,7 +261,7 @@ pipeline {
                         \\"state\\": \\"success\\",
                         \\"context\\": \\"jenkins-ctest-multigpu\\",
                         \\"description\\": \\"Jenkins CI Job: jenkins-ctest-multigpu\\",
-                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Multibranch-Github/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
+                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Github-Multigpu/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
                 }"
             '''
         }
@@ -280,7 +280,7 @@ pipeline {
                         \\"state\\": \\"failure\\",
                         \\"context\\": \\"jenkins-ctest-multigpu\\",
                         \\"description\\": \\"Jenkins CI Job: jenkins-ctest-multigpu\\",
-                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Multibranch-Github/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
+                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Github-Multigpu/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
                 }"
             '''
         }
@@ -299,7 +299,7 @@ pipeline {
                         \\"state\\": \\"error\\",
                         \\"context\\": \\"jenkins-ctest-multigpu\\",
                         \\"description\\": \\"Jenkins CI Job: jenkins-ctest-multigpu\\",
-                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Multibranch-Github/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
+                        \\"target_url\\": \\"https://simsgs.informatik.uni-stuttgart.de/jenkins/view/PLSSVM/job/PLSSVM/job/Github-Multigpu/job/${BRANCH_NAME}/$BUILD_NUMBER\\"
                 }"
             '''
         }

CMakeLists.txt

@@ -6,7 +6,7 @@
 
 cmake_minimum_required(VERSION 3.18)
 
-project(ParallelLeastSquaresSupportVectorMachine
+project("PLSSVM - Parallel Least-Squares Support Vector Machine"
         VERSION 0.1.0
         LANGUAGES CXX
         DESCRIPTION "A Support Vector Machine implementation using different backends.")
@@ -27,10 +27,18 @@ configure_file(
 ########################################################################################################################
 ## set base sources
 set(PLSSVM_BASE_SOURCES
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/detail/execution_range.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/detail/file_reader.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/detail/string_utility.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/exceptions/exceptions.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/exceptions/source_location.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/backend_types.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/kernel_types.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/csvm.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/parameter.cpp
-    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/parameter_train.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/parameter_predict.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/parameter_train.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/src/plssvm/target_platforms.cpp
 )
 
 ## create base library: linked against all backend libraries
@@ -337,7 +345,7 @@ if(DEFINED PLSSVM_THREAD_BLOCK_SIZE)
     if (PLSSVM_THREAD_BLOCK_SIZE MATCHES "^[0-9]+$" AND PLSSVM_THREAD_BLOCK_SIZE GREATER 0)
         message(STATUS "Set THREAD_BLOCK_SIZE to ${PLSSVM_THREAD_BLOCK_SIZE}.")
         # add target definition
-        target_compile_definitions(${PLSSVM_EXECUTABLE_NAME} PUBLIC PLSSVM_THREAD_BLOCK_SIZE=${PLSSVM_THREAD_BLOCK_SIZE})
+        target_compile_definitions(${PLSSVM_ALL_LIBRARY_NAME} PUBLIC PLSSVM_THREAD_BLOCK_SIZE=${PLSSVM_THREAD_BLOCK_SIZE})
     else()
         message(FATAL_ERROR "PLSSVM_THREAD_BLOCK_SIZE must be an integer greater than 0 but is \"${PLSSVM_THREAD_BLOCK_SIZE}\"!")
     endif()
@@ -351,12 +359,26 @@ if(DEFINED PLSSVM_INTERNAL_BLOCK_SIZE)
     if (PLSSVM_INTERNAL_BLOCK_SIZE MATCHES "^[0-9]+$" AND PLSSVM_INTERNAL_BLOCK_SIZE GREATER 0)
         message(STATUS "Set INTERNAL_BLOCK_SIZE to ${PLSSVM_INTERNAL_BLOCK_SIZE}.")
         # add target definition
-        target_compile_definitions(${PLSSVM_EXECUTABLE_NAME} PUBLIC PLSSVM_INTERNAL_BLOCK_SIZE=${PLSSVM_INTERNAL_BLOCK_SIZE})
+        target_compile_definitions(${PLSSVM_ALL_LIBRARY_NAME} PUBLIC PLSSVM_INTERNAL_BLOCK_SIZE=${PLSSVM_INTERNAL_BLOCK_SIZE})
     else()
         message(FATAL_ERROR "PLSSVM_INTERNAL_BLOCK_SIZE must be an integer greater than 0 but is \"${PLSSVM_INTERNAL_BLOCK_SIZE}\"!")
     endif()
 endif()
 
+## set specific internal block sizes of requested
+if(DEFINED ENV{PLSSVM_OPENMP_BLOCK_SIZE})
+    set(PLSSVM_OPENMP_BLOCK_SIZE $ENV{PLSSVM_OPENMP_BLOCK_SIZE} CACHE STRING "The used block size for the OpenMP kernel." FORCE)
+endif()
+if(DEFINED PLSSVM_OPENMP_BLOCK_SIZE)
+    if (PLSSVM_OPENMP_BLOCK_SIZE MATCHES "^[0-9]+$" AND PLSSVM_OPENMP_BLOCK_SIZE GREATER 0)
+        message(STATUS "Set PLSSVM_OPENMP_BLOCK_SIZE to ${PLSSVM_OPENMP_BLOCK_SIZE}.")
+        # add target definition
+        target_compile_definitions(${PLSSVM_ALL_LIBRARY_NAME} PUBLIC PLSSVM_OPENMP_BLOCK_SIZE=${PLSSVM_OPENMP_BLOCK_SIZE})
+    else()
+        message(FATAL_ERROR "PLSSVM_OPENMP_BLOCK_SIZE must be an integer greater than 0 but is \"${PLSSVM_OPENMP_BLOCK_SIZE}\"!")
+    endif()
+endif()
+
 ## change executable floating points from double precision to single precision
 option(PLSSVM_EXECUTABLES_USE_SINGLE_PRECISION "Build the svm-train and svm-predict executables with single precision instead of double precision." OFF)
 if(PLSSVM_EXECUTABLES_USE_SINGLE_PRECISION)
@@ -402,6 +424,20 @@ if(PLSSVM_ENABLE_TESTING)
 endif()
 
 
+
+########################################################################################################################
+##                                        enable timing generation via a script                                       ##
+########################################################################################################################
+option(PLSSVM_GENERATE_TIMING_SCRIPT "Generate a timing script used for performance measurement." OFF)
+if(PLSSVM_GENERATE_TIMING_SCRIPT)
+    configure_file(
+            ${CMAKE_CURRENT_SOURCE_DIR}/cmake/time.sh.in
+            ${CMAKE_BINARY_DIR}/time.sh
+            @ONLY
+    )
+endif()
+
+
 ########################################################################################################################
 ##                                            print short (backend) summary                                           ##
 ########################################################################################################################
@@ -469,19 +505,3 @@ install(FILES "${PROJECT_BINARY_DIR}/plssvmConfig.cmake"
         "${PROJECT_BINARY_DIR}/plssvmConfigVersion.cmake"
         DESTINATION ${CMAKE_INSTALL_DATAROOTDIR}/plssvm/cmake
         )
-
-
-
-
-
-# TODO: separate repo?
-## generate scripts
-option(PLSSVM_GENERATE_TIMINGSCRIPT "Generate script for timings." ON)
-if(PLSSVM_GENERATE_TIMINGSCRIPT)
-    configure_file(
-            ${CMAKE_CURRENT_SOURCE_DIR}/cmake/time.sh.in
-            ${CMAKE_BINARY_DIR}/time.sh
-            @ONLY
-    )
-endif()
-

README.md

@@ -3,38 +3,38 @@
 
 Implementation of a parallel [least-squares support-vector machine](https://en.wikipedia.org/wiki/Least-squares_support-vector_machine) using multiple different backends.
 The currently available backends are:
-- [OpenMP](https://www.openmp.org/)
-- [CUDA](https://developer.nvidia.com/cuda-zone)
-- [OpenCL](https://www.khronos.org/opencl/)
-- [SYCL](https://www.khronos.org/sycl/)
+  - [OpenMP](https://www.openmp.org/)
+  - [CUDA](https://developer.nvidia.com/cuda-zone)
+  - [OpenCL](https://www.khronos.org/opencl/)
+  - [SYCL](https://www.khronos.org/sycl/)
 
 ## Getting Started
 
 ### Dependencies
 
-General dependencies:
-- a C++17 capable compiler (e.g. [`gcc`](https://gcc.gnu.org/) or [`clang`](https://clang.llvm.org/))
-- [CMake](https://cmake.org/) 3.18 or newer
-- [cxxopts](https://github.com/jarro2783/cxxopts), [fast_float](https://github.com/fastfloat/fast_float) and [{fmt}](https://github.com/fmtlib/fmt) (all three are automatically build during the CMake configuration if they couldn't be found using the respective `find_package` call)
-- [GoogleTest](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
-- [OpenMP](https://www.openmp.org/) 4.0 or newer (optional) to speed-up file parsing
+General dependencies: 
+  - a C++17 capable compiler (e.g. [`gcc`](https://gcc.gnu.org/) or [`clang`](https://clang.llvm.org/))
+  - [CMake](https://cmake.org/) 3.18 or newer
+  - [cxxopts](https://github.com/jarro2783/cxxopts), [fast_float](https://github.com/fastfloat/fast_float) and [{fmt}](https://github.com/fmtlib/fmt) (all three are automatically build during the CMake configuration if they couldn't be found using the respective `find_package` call)
+  - [GoogleTest](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
+  - [OpenMP](https://www.openmp.org/) 4.0 or newer (optional) to speed-up file parsing
 
 Additional dependencies for the OpenMP backend:
-- compiler with OpenMP support
+  - compiler with OpenMP support
 
 Additional dependencies for the CUDA backend:
-- CUDA SDK
-- either NVIDIA [`nvcc`](https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html) or [`clang` with CUDA support enabled](https://llvm.org/docs/CompileCudaWithLLVM.html)
+  - CUDA SDK
+  - either NVIDIA [`nvcc`](https://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html) or [`clang` with CUDA support enabled](https://llvm.org/docs/CompileCudaWithLLVM.html)
 
 Additional dependencies for the OpenCL backend:
-- OpenCL runtime and header files
+  - OpenCL runtime and header files
 
 Additional dependencies for the SYCL backend:
-- the code must be compiled with a SYCL capable compiler; currently tested with [DPC++](https://github.com/intel/llvm) and [hipSYCL](https://github.com/illuhad/hipSYCL)
+  - the code must be compiled with a SYCL capable compiler; currently tested with [DPC++](https://github.com/intel/llvm) and [hipSYCL](https://github.com/illuhad/hipSYCL)
 
 Additional dependencies if `PLSSVM_ENABLE_TESTING` and `PLSSVM_GENERATE_TEST_FILE` are both set to `ON`:
-- [Python3](https://www.python.org/) with the [`argparse`](https://docs.python.org/3/library/argparse.html) and [`sklearn`](https://scikit-learn.org/stable/) modules
+  - [Python3](https://www.python.org/) with the [`argparse`](https://docs.python.org/3/library/argparse.html) and [`sklearn`](https://scikit-learn.org/stable/) modules
 
 ### Building
 
@@ -52,10 +52,10 @@ Building the library can be done using the normal CMake approach:
 
 The **required** CMake option `PLSSVM_TARGET_PLATFORMS` is used to determine for which targets the backends should be compiled.
 Valid targets are:
-- `cpu`: compile for the CPU; **no** architectural specifications  is allowed
-- `nvidia`: compile for NVIDIA GPUs; **at least one** architectural specification is necessary, e.g. `nvidia:sm_86,sm_70`
-- `amd`: compile for AMD GPUs; **at least one** architectural specification is necessary, e.g. `amd:gfx906`
-- `intel`: compile for Intel GPUs; **no** architectural specification is allowed
+  - `cpu`: compile for the CPU; **no** architectural specifications  is allowed
+  - `nvidia`: compile for NVIDIA GPUs; **at least one** architectural specification is necessary, e.g. `nvidia:sm_86,sm_70`
+  - `amd`: compile for AMD GPUs; **at least one** architectural specification is necessary, e.g. `amd:gfx906`
+  - `intel`: compile for Intel GPUs; **no** architectural specification is allowed
 
 At least one of the above targets must be present.
 
@@ -74,52 +74,53 @@ optional arguments:
 Example invocations:
 
 ```bash
-> python3 utility/gpu_name_to_arch.py --name "GeForce RTX 3080"
+> python3 utility_scripts/gpu_name_to_arch.py --name "GeForce RTX 3080"
 sm_86
-> python3 utility/gpu_name_to_arch.py --name "Radeon VII"
+> python3 utility_scripts/gpu_name_to_arch.py --name "Radeon VII"
 gfx906
 ```
 
 If no GPU name is provided, the script tries to automatically detect any NVIDIA or AMD GPU
 (requires the Python3 dependencies [`GPUtil`](https://pypi.org/project/GPUtil/) and [`pyamdgpuinfo`](https://pypi.org/project/pyamdgpuinfo/)).
 
 If the architectural information for the requested GPU could not be retrieved, one option would be to have a look at:
-- for NVIDIA GPUs:  [Your GPU Compute Capability](https://developer.nvidia.com/cuda-gpus)
-- for AMD GPUs: [ROCm Documentation](https://github.com/RadeonOpenCompute/ROCm_Documentation/blob/master/ROCm_Compiler_SDK/ROCm-Native-ISA.rst)
+  - for NVIDIA GPUs:  [Your GPU Compute Capability](https://developer.nvidia.com/cuda-gpus)
+  - for AMD GPUs: [ROCm Documentation](https://github.com/RadeonOpenCompute/ROCm_Documentation/blob/master/ROCm_Compiler_SDK/ROCm-Native-ISA.rst)
 
 #### Optional CMake Options
 
 The `[optional_options]` can be one or multiple of:
 
-- `PLSSVM_ENABLE_OPENMP_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
+  - `PLSSVM_ENABLE_OPENMP_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
     - `ON`: check for the OpenMP backend and fail if not available
     - `AUTO`: check for the OpenMP backend but **do not** fail if not available
     - `OFF`: do not check for the OpenMP backend
-- `PLSSVM_ENABLE_CUDA_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
+  - `PLSSVM_ENABLE_CUDA_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
     - `ON`: check for the CUDA backend and fail if not available
     - `AUTO`: check for the CUDA backend but **do not** fail if not available
     - `OFF`: do not check for the CUDA backend
-- `PLSSVM_ENABLE_OPENCL_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
+  - `PLSSVM_ENABLE_OPENCL_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
     - `ON`: check for the OpenCL backend and fail if not available
     - `AUTO`: check for the OpenCL backend but **do not** fail if not available
     - `OFF`: do not check for the OpenCL backend
-- `PLSSVM_ENABLE_SYCL_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
-  - `ON`: check for the SYCL backend and fail if not available
-  - `AUTO`: check for the SYCL backend but **do not** fail if not available
-  - `OFF`: do not check for the SYCL backend
+  - `PLSSVM_ENABLE_SYCL_BACKEND=ON|OFF|AUTO` (default: `AUTO`):
+    - `ON`: check for the SYCL backend and fail if not available
+    - `AUTO`: check for the SYCL backend but **do not** fail if not available
+    - `OFF`: do not check for the SYCL backend
 
 **Attention:** at least one backend must be enabled and available!
 
-- `PLSSVM_ENABLE_ASSERTS=ON|OFF` (default: `OFF`): enables custom assertions regardless whether the `DEBUG` macro is defined or not
-- `PLSSVM_THREAD_BLOCK_SIZE` (default: `16`): set a specific thread block size used in the GPU kernels (for fine-tuning optimizations)
-- `PLSSVM_INTERNAL_BLOCK_SIZE` (default: `6`: set a specific internal block size used in the GPU kernels (for fine-tuning optimizations)
-- `PLSSVM_EXECUTABLES_USE_SINGLE_PRECISION` (default: `OFF`): enables single precision calculations instead of double precision for the `svm-train` and `svm-predict` executables
-- `PLSSVM_ENABLE_LTO=ON|OFF` (default: `ON`): enable interprocedural optimization (IPO/LTO) if supported by the compiler
-- `PLSSVM_ENABLE_DOCUMENTATION=ON|OFF` (default: `OFF`): enable the `doc` target using doxygen
-- `PLSSVM_ENABLE_TESTING=ON|OFF` (default: ON): enable testing using GoogleTest and ctest
+  - `PLSSVM_ENABLE_ASSERTS=ON|OFF` (default: `OFF`): enables custom assertions regardless whether the `DEBUG` macro is defined or not
+  - `PLSSVM_THREAD_BLOCK_SIZE` (default: `16`): set a specific thread block size used in the GPU kernels (for fine-tuning optimizations)
+  - `PLSSVM_INTERNAL_BLOCK_SIZE` (default: `6`: set a specific internal block size used in the GPU kernels (for fine-tuning optimizations)
+  - `PLSSVM_EXECUTABLES_USE_SINGLE_PRECISION` (default: `OFF`): enables single precision calculations instead of double precision for the `svm-train` and `svm-predict` executables
+  - `PLSSVM_ENABLE_LTO=ON|OFF` (default: `ON`): enable interprocedural optimization (IPO/LTO) if supported by the compiler
+  - `PLSSVM_ENABLE_DOCUMENTATION=ON|OFF` (default: `OFF`): enable the `doc` target using doxygen
+  - `PLSSVM_ENABLE_TESTING=ON|OFF` (default: `ON`): enable testing using GoogleTest and ctest
+  - `PLSSVM_GENERATE_TIMING_SCRIPT=ON|OFF` (default: `OFF`): configure a timing script usable for performance measurement
 
 If `PLSSVM_ENABLE_TESTING` is set to `ON`, the following options can also be set:
-- `PLSSVM_GENERATE_TEST_FILE=ON|OFF` (default: `ON`): automatically generate test files
+  - `PLSSVM_GENERATE_TEST_FILE=ON|OFF` (default: `ON`): automatically generate test files
     - `PLSSVM_TEST_FILE_NUM_DATA_POINTS` (default: `5000`): the number of data points in the test file
     - `PLSSVM_TEST_FILE_NUM_FEATURES` (default: `2000`): the number of features per data point