Merge branch 'develop' of github.com:UoB-HPC/BabelStream into develop

CHANGELOG.md

@@ -2,6 +2,11 @@
 All notable changes to this project will be documented in this file.
 
 ## Unreleased
+### Added
+- Ability to build Kokkos and RAJA versions against existing packages.
+
+### Changed
+- RAJA CUDA CMake build issues resolved.
 - Fix CUDA memory limit check.
 - Use long double for `check_solution` in case of large problem size.
 

CMakeLists.txt

@@ -125,7 +125,7 @@ message(STATUS "Default ${CMAKE_BUILD_TYPE} flags are `${DEFAULT_${BUILD_TYPE}_F
 # setup common build flag defaults if there are no overrides
 if (NOT DEFINED ${BUILD_TYPE}_FLAGS)
     set(ACTUAL_${BUILD_TYPE}_FLAGS ${DEFAULT_${BUILD_TYPE}_FLAGS})
-elseif ()
+else ()
     set(ACTUAL_${BUILD_TYPE}_FLAGS ${${BUILD_TYPE}_FLAGS})
 endif ()
 

README.md

@@ -100,7 +100,7 @@ The source for each model's implementations are located in `./src/<model>`.
 
 Currently available models are:
 ```
-omp;ocl;std;std20;hip;cuda;kokkos;sycl;sycl2020;acc;raja;tbb;thrust
+omp;ocl;std-data;std-indices;std-ranges;hip;cuda;kokkos;sycl;sycl2020;acc;raja;tbb;thrust
 ```
 
 #### Overriding default flags
@@ -163,11 +163,11 @@ Pull Requests should be made against the `develop` branch.
 
 Please cite BabelStream via this reference:
 
-Deakin T, Price J, Martineau M, McIntosh-Smith S. GPU-STREAM v2.0: Benchmarking the achievable memory bandwidth of many-core processors across diverse parallel programming models. 2016. Paper presented at P^3MA Workshop at ISC High Performance, Frankfurt, Germany. DOI: 10.1007/978- 3-319-46079-6_34
+Deakin T, Price J, Martineau M, McIntosh-Smith S. Evaluating attainable memory bandwidth of parallel programming models via BabelStream. International Journal of Computational Science and Engineering. Special issue. Vol. 17, No. 3, pp. 247–262. 2018. DOI: 10.1504/IJCSE.2018.095847
 
 ### Other BabelStream publications
 
-* Deakin T, Price J, Martineau M, McIntosh-Smith S. Evaluating attainable memory bandwidth of parallel programming models via BabelStream. International Journal of Computational Science and Engineering. Special issue. Vol. 17, No. 3, pp. 247–262. 2018.DOI: 10.1504/IJCSE.2018.095847
+* Deakin T, Price J, Martineau M, McIntosh-Smith S. GPU-STREAM v2.0: Benchmarking the achievable memory bandwidth of many-core processors across diverse parallel programming models. 2016. Paper presented at P^3MA Workshop at ISC High Performance, Frankfurt, Germany. DOI: 10.1007/978- 3-319-46079-6_34
 
 * Deakin T, McIntosh-Smith S. GPU-STREAM: Benchmarking the achievable memory bandwidth of Graphics Processing Units. 2015. Poster session presented at IEEE/ACM SuperComputing, Austin, United States.
   You can view the [Poster and Extended Abstract](http://sc15.supercomputing.org/sites/all/themes/SC15images/tech_poster/tech_poster_pages/post150.html).

src/kokkos/model.cmake

@@ -1,16 +1,17 @@
-
 register_flag_optional(CMAKE_CXX_COMPILER
         "Any CXX compiler that is supported by CMake detection and RAJA.
          See https://github.com/kokkos/kokkos#primary-tested-compilers-on-x86-are"
         "c++")
 
-register_flag_required(KOKKOS_IN_TREE
+register_flag_optional(KOKKOS_IN_TREE
         "Absolute path to the *source* distribution directory of Kokkos.
          Remember to append Kokkos specific flags as well, for example:
-
              -DKOKKOS_IN_TREE=... -DKokkos_ENABLE_OPENMP=ON -DKokkos_ARCH_ZEN=ON ...
+         See https://github.com/kokkos/kokkos/blob/master/BUILD.md for all available options" "")
 
-         See https://github.com/kokkos/kokkos/blob/master/BUILD.md for all available options")
+register_flag_optional(KOKKOS_IN_PACKAGE
+        "Use if Kokkos is part of a package dependency: 
+        Path to package R-Path containing Kokkos libs" "")
 
 # compiler vendor and arch specific flags
 set(KOKKOS_FLAGS_CPU_INTEL -qopt-streaming-stores=always)
@@ -20,13 +21,18 @@ macro(setup)
     set(CMAKE_CXX_STANDARD 14)
     cmake_policy(SET CMP0074 NEW) #see https://github.com/kokkos/kokkos/blob/master/BUILD.md
 
-    message(STATUS "Building using in-tree Kokkos source at `${KOKKOS_IN_TREE}`")
 
     if (EXISTS "${KOKKOS_IN_TREE}")
+        message(STATUS "Building using in-tree Kokkos source at `${KOKKOS_IN_TREE}`")
         add_subdirectory(${KOKKOS_IN_TREE} ${CMAKE_BINARY_DIR}/kokkos)
         register_link_library(Kokkos::kokkos)
+    elseif (EXISTS "${KOKKOS_IN_PACKAGE}")
+        message(STATUS "Building using packaged Kokkos at `${KOKKOS_IN_PACKAGE}`")
+        set (Kokkos_DIR "${KOKKOS_IN_PACKAGE}/lib64/cmake/Kokkos")
+        find_package(Kokkos REQUIRED)
+        register_link_library(Kokkos::kokkos) 
     else()
-        message(FATAL_ERROR "`${KOKKOS_IN_TREE}` does not exist")
+        message(FATAL_ERROR "Neither `${KOKKOS_IN_TREE}`, or `${KOKKOS_IN_PACKAGE}` exists")
     endif ()
 
     register_append_compiler_and_arch_specific_cxx_flags(
@@ -36,5 +42,3 @@ macro(setup)
     )
 
 endmacro()
-
-

src/raja/model.cmake

@@ -1,18 +1,19 @@
-
 register_flag_optional(CMAKE_CXX_COMPILER
         "Any CXX compiler that is supported by CMake detection and RAJA.
          See https://raja.readthedocs.io/en/main/getting_started.html#build-and-install"
         "c++")
 
-register_flag_required(RAJA_IN_TREE
+register_flag_optional(RAJA_IN_TREE
         "Absolute path to the *source* distribution directory of RAJA.
          Make sure to use the release version of RAJA or clone RAJA recursively with submodules.
          Remember to append RAJA specific flags as well, for example:
-
              -DRAJA_IN_TREE=... -DENABLE_OPENMP=ON -DENABLE_CUDA=ON ...
-
          See https://github.com/LLNL/RAJA/blob/08cbbafd2d21589ebf341f7275c229412d0fe903/CMakeLists.txt#L44 for all available options
-")
+" "")
+
+register_flag_optional(RAJA_IN_PACKAGE
+        "Use if Raja is part of a package dependency: 
+        Path to installation" "")
 
 register_flag_optional(TARGET
         "Target offload device, implemented values are CPU, NVIDIA"
@@ -76,16 +77,22 @@ macro(setup)
         register_link_library(RAJA)
         # RAJA's cmake screws with where the binary will end up, resetting it here:
         set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR})
+    
+    elseif (EXISTS "${RAJA_IN_PACKAGE}")
+        message(STATUS "Building using packaged Raja at `${RAJA_IN_PACKAGE}`")
+        find_package(RAJA REQUIRED)
+        register_link_library(RAJA)
+    
     else ()
-        message(FATAL_ERROR "`${RAJA_IN_TREE}` does not exist")
+        message(FATAL_ERROR "Neither `${RAJA_IN_TREE}` or `${RAJA_IN_PACKAGE}` exists")
     endif ()
 
 
     if (ENABLE_CUDA)
         # RAJA needs the codebase to be compiled with nvcc, so we tell cmake to treat sources as *.cu
         enable_language(CUDA)
-        set_source_files_properties(RAJAStream.cpp PROPERTIES LANGUAGE CUDA)
+        set_source_files_properties(src/raja/RAJAStream.cpp PROPERTIES LANGUAGE CUDA)
-        set_source_files_properties(main.cpp PROPERTIES LANGUAGE CUDA)
+        set_source_files_properties(src/main.cpp PROPERTIES LANGUAGE CUDA)
     endif ()