GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-07-18 10:54:26 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix failing tests on arm/ppc

Browse Source
This commit is contained in:
Antonio Sanchez 2025-02-28 13:15:33 -08:00
parent ea37d9e73e
commit c53002f5fb
5 changed files with 53 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
Eigen/src/Core/util/Meta.h
View File

@ -166,10 +166,8 @@ template<> struct make_unsigned<signed int> { typedef unsigned int type; }
template<> struct make_unsigned<unsigned int> { typedef unsigned int type; }; template<> struct make_unsigned<unsigned int> { typedef unsigned int type; };
template<> struct make_unsigned<signed long> { typedef unsigned long type; }; template<> struct make_unsigned<signed long> { typedef unsigned long type; };
template<> struct make_unsigned<unsigned long> { typedef unsigned long type; }; template<> struct make_unsigned<unsigned long> { typedef unsigned long type; };
#if EIGEN_COMP_MSVC template<> struct make_unsigned<signed long long> { typedef unsigned long type; };
template<> struct make_unsigned<signed __int64> { typedef unsigned __int64 type; }; template<> struct make_unsigned<unsigned long long> { typedef unsigned long type; };
template<> struct make_unsigned<unsigned __int64> { typedef unsigned __int64 type; };
#endif
#endif #endif
template <typename T> struct add_const { typedef const T type; }; template <typename T> struct add_const { typedef const T type; };
@ -527,7 +525,7 @@ template<typename T, typename U> struct scalar_product_traits
} // end namespace internal } // end namespace internal
namespace numext { namespace numext {
#if defined(__CUDA_ARCH__) #if defined(__CUDA_ARCH__)
template<typename T> EIGEN_DEVICE_FUNC void swap(T &a, T &b) { T tmp = b; b = a; a = tmp; } template<typename T> EIGEN_DEVICE_FUNC void swap(T &a, T &b) { T tmp = b; b = a; a = tmp; }
#else #else

2
blas/CMakeLists.txt
View File

@ -7,7 +7,7 @@ workaround_9220(Fortran EIGEN_Fortran_COMPILER_WORKS)
if(EIGEN_Fortran_COMPILER_WORKS) if(EIGEN_Fortran_COMPILER_WORKS)
enable_language(Fortran OPTIONAL) enable_language(Fortran OPTIONAL)
if(NOT CMAKE_Fortran_COMPILER) if(CMAKE_CROSSCOMPILING OR NOT CMAKE_Fortran_COMPILER)
set(EIGEN_Fortran_COMPILER_WORKS OFF) set(EIGEN_Fortran_COMPILER_WORKS OFF)
endif() endif()
endif() endif()

8
cmake/EigenTesting.cmake
View File

@ -71,7 +71,7 @@ macro(ei_add_test_internal testname testname_with_suffix)
endif(${ARGC} GREATER 2) endif(${ARGC} GREATER 2)
if(EIGEN_TEST_CUSTOM_CXX_FLAGS) if(EIGEN_TEST_CUSTOM_CXX_FLAGS)
ei_add_target_property(${targetname} COMPILE_FLAGS "${EIGEN_TEST_CUSTOM_CXX_FLAGS}") target_compile_options(${targetname} PRIVATE ${EIGEN_TEST_CUSTOM_CXX_FLAGS})
endif() endif()
if(EIGEN_STANDARD_LIBRARIES_TO_LINK_TO) if(EIGEN_STANDARD_LIBRARIES_TO_LINK_TO)
@ -145,10 +145,10 @@ macro(ei_add_test_internal_sycl testname testname_with_suffix)
endif() endif()
if(EIGEN_NO_ASSERTION_CHECKING) if(EIGEN_NO_ASSERTION_CHECKING)
ei_add_target_property(${targetname} COMPILE_FLAGS "-DEIGEN_NO_ASSERTION_CHECKING=1") target_compile_options(${targetname} PRIVATE "-DEIGEN_NO_ASSERTION_CHECKING=1")
else(EIGEN_NO_ASSERTION_CHECKING) else(EIGEN_NO_ASSERTION_CHECKING)
if(EIGEN_DEBUG_ASSERTS) if(EIGEN_DEBUG_ASSERTS)
ei_add_target_property(${targetname} COMPILE_FLAGS "-DEIGEN_DEBUG_ASSERTS=1") target_compile_options(${targetname} PRIVATE "-DEIGEN_DEBUG_ASSERTS=1")
endif(EIGEN_DEBUG_ASSERTS) endif(EIGEN_DEBUG_ASSERTS)
endif(EIGEN_NO_ASSERTION_CHECKING) endif(EIGEN_NO_ASSERTION_CHECKING)
@ -166,7 +166,7 @@ macro(ei_add_test_internal_sycl testname testname_with_suffix)
endif(${ARGC} GREATER 2) endif(${ARGC} GREATER 2)
if(EIGEN_TEST_CUSTOM_CXX_FLAGS) if(EIGEN_TEST_CUSTOM_CXX_FLAGS)
ei_add_target_property(${targetname} COMPILE_FLAGS "${EIGEN_TEST_CUSTOM_CXX_FLAGS}") target_compile_options(${targetname} PRIVATE ${EIGEN_TEST_CUSTOM_CXX_FLAGS})
endif() endif()
if(EIGEN_STANDARD_LIBRARIES_TO_LINK_TO) if(EIGEN_STANDARD_LIBRARIES_TO_LINK_TO)

14
test/CMakeLists.txt
View File

@ -1,6 +1,6 @@
# generate split test header file only if it does not yet exist # generate split test header file only if it does not yet exist
# in order to prevent a rebuild everytime cmake is configured # in order to prevent a rebuild everytime cmake is configured
if(NOT EXISTS ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h) if(NOT EXISTS ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h)
file(WRITE ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h "") file(WRITE ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h "")
foreach(i RANGE 1 999) foreach(i RANGE 1 999)
file(APPEND ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h file(APPEND ${CMAKE_CURRENT_BINARY_DIR}/split_test_helper.h
@ -340,7 +340,7 @@ endif()
# boost MP unit test # boost MP unit test
find_package(Boost) find_package(Boost)
if(Boost_FOUND) if(Boost_FOUND AND EIGEN_TEST_CXX11)
include_directories(${Boost_INCLUDE_DIRS}) include_directories(${Boost_INCLUDE_DIRS})
ei_add_test(boostmultiprec "" "${Boost_LIBRARIES}") ei_add_test(boostmultiprec "" "${Boost_LIBRARIES}")
ei_add_property(EIGEN_TESTED_BACKENDS "Boost.Multiprecision, ") ei_add_property(EIGEN_TESTED_BACKENDS "Boost.Multiprecision, ")
@ -361,9 +361,9 @@ if(EIGEN_TEST_CUDA)
find_package(CUDA 5.0) find_package(CUDA 5.0)
if(CUDA_FOUND) if(CUDA_FOUND)
set(CUDA_PROPAGATE_HOST_FLAGS OFF) set(CUDA_PROPAGATE_HOST_FLAGS OFF)
if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "Clang") if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "Clang")
set(CUDA_NVCC_FLAGS "-ccbin ${CMAKE_C_COMPILER}" CACHE STRING "nvcc flags" FORCE) set(CUDA_NVCC_FLAGS "-ccbin ${CMAKE_C_COMPILER}" CACHE STRING "nvcc flags" FORCE)
endif() endif()
if(EIGEN_TEST_CUDA_CLANG) if(EIGEN_TEST_CUDA_CLANG)
@ -371,9 +371,9 @@ if(CUDA_FOUND)
endif() endif()
cuda_include_directories(${CMAKE_CURRENT_BINARY_DIR}) cuda_include_directories(${CMAKE_CURRENT_BINARY_DIR})
set(EIGEN_ADD_TEST_FILENAME_EXTENSION "cu") set(EIGEN_ADD_TEST_FILENAME_EXTENSION "cu")
ei_add_test(cuda_basic) ei_add_test(cuda_basic)
unset(EIGEN_ADD_TEST_FILENAME_EXTENSION) unset(EIGEN_ADD_TEST_FILENAME_EXTENSION)
endif(CUDA_FOUND) endif(CUDA_FOUND)
@ -381,7 +381,7 @@ endif(CUDA_FOUND)
endif(EIGEN_TEST_CUDA) endif(EIGEN_TEST_CUDA)
file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/failtests) file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/failtests)
add_test(NAME failtests WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/failtests COMMAND ${CMAKE_COMMAND} ${Eigen_SOURCE_DIR} -G "${CMAKE_GENERATOR}" -DEIGEN_FAILTEST=ON) add_test(NAME failtests WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/failtests COMMAND ${CMAKE_COMMAND} ${Eigen_SOURCE_DIR} -G "${CMAKE_GENERATOR}" -DEIGEN_FAILTEST=ON)
option(EIGEN_TEST_BUILD_DOCUMENTATION "Test building the doxygen documentation" OFF) option(EIGEN_TEST_BUILD_DOCUMENTATION "Test building the doxygen documentation" OFF)

62
unsupported/test/CMakeLists.txt
View File

@ -215,18 +215,6 @@ if(CUDA_FOUND AND EIGEN_TEST_CUDA)
message(STATUS "Flags used to compile cuda code: " ${CMAKE_CXX_FLAGS}) message(STATUS "Flags used to compile cuda code: " ${CMAKE_CXX_FLAGS})
if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "Clang")
set(CUDA_NVCC_FLAGS "-ccbin ${CMAKE_C_COMPILER}" CACHE STRING "nvcc flags" FORCE)
endif()
if(EIGEN_TEST_CUDA_CLANG)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 --cuda-gpu-arch=sm_${EIGEN_CUDA_COMPUTE_ARCH}")
endif()
set(EIGEN_CUDA_RELAXED_CONSTEXPR "--expt-relaxed-constexpr")
if (${CUDA_VERSION} STREQUAL "7.0")
set(EIGEN_CUDA_RELAXED_CONSTEXPR "--relaxed-constexpr")
endif()
if( (NOT EIGEN_TEST_CXX11) OR (CMAKE_VERSION VERSION_LESS 3.3)) if( (NOT EIGEN_TEST_CXX11) OR (CMAKE_VERSION VERSION_LESS 3.3))
set(EIGEN_CUDA_CXX11_FLAG "-std=c++11") set(EIGEN_CUDA_CXX11_FLAG "-std=c++11")
else() else()
@ -234,8 +222,39 @@ if(CUDA_FOUND AND EIGEN_TEST_CUDA)
set(EIGEN_CUDA_CXX11_FLAG "") set(EIGEN_CUDA_CXX11_FLAG "")
endif() endif()
set(CUDA_NVCC_FLAGS "${EIGEN_CUDA_CXX11_FLAG} ${EIGEN_CUDA_RELAXED_CONSTEXPR} -arch compute_${EIGEN_CUDA_COMPUTE_ARCH} -Xcudafe \"--display_error_number\" ${CUDA_NVCC_FLAGS}") if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "Clang")
cuda_include_directories("${CMAKE_CURRENT_BINARY_DIR}" "${CUDA_TOOLKIT_ROOT_DIR}/include") set(CUDA_NVCC_FLAGS "-ccbin ${CMAKE_C_COMPILER}" CACHE STRING "nvcc flags" FORCE)
endif()
if(EIGEN_TEST_CUDA_CLANG)
string(APPEND CMAKE_CXX_FLAGS " --cuda-path=${CUDA_TOOLKIT_ROOT_DIR} ${EIGEN_CUDA_CXX11_FLAG}")
foreach(ARCH IN LISTS EIGEN_CUDA_COMPUTE_ARCH)
string(APPEND CMAKE_CXX_FLAGS " --cuda-gpu-arch=sm_${ARCH}")
endforeach()
else()
set(CUDA_PROPAGATE_HOST_FLAGS OFF)
set(NVCC_ARCH_FLAGS)
# Define an -arch=sm_<arch>, otherwise if GPU does not exactly match one of
# those in the arch list for -gencode, the kernels will fail to run with
# cudaErrorNoKernelImageForDevice
# This can happen with newer cards (e.g. sm_75) and compiling with older
# versions of nvcc (e.g. 9.2) that do not support their specific arch.
list(LENGTH EIGEN_CUDA_COMPUTE_ARCH EIGEN_CUDA_COMPUTE_ARCH_SIZE)
if(EIGEN_CUDA_COMPUTE_ARCH_SIZE)
list(GET EIGEN_CUDA_COMPUTE_ARCH 0 EIGEN_CUDA_COMPUTE_DEFAULT)
set(NVCC_ARCH_FLAGS " -arch=sm_${EIGEN_CUDA_COMPUTE_DEFAULT}")
endif()
foreach(ARCH IN LISTS EIGEN_CUDA_COMPUTE_ARCH)
string(APPEND NVCC_ARCH_FLAGS " -gencode arch=compute_${ARCH},code=sm_${ARCH}")
endforeach()
set(CUDA_NVCC_FLAGS "--expt-relaxed-constexpr -Xcudafe \"--display_error_number\" ${NVCC_ARCH_FLAGS} ${CUDA_NVCC_FLAGS} ${EIGEN_CUDA_CXX11_FLAG}")
cuda_include_directories("${CMAKE_CURRENT_BINARY_DIR}" "${CUDA_TOOLKIT_ROOT_DIR}/include")
endif()
set(EIGEN_CUDA_RELAXED_CONSTEXPR "--expt-relaxed-constexpr")
if (${CUDA_VERSION} STREQUAL "7.0")
set(EIGEN_CUDA_RELAXED_CONSTEXPR "--relaxed-constexpr")
endif()
set(EIGEN_ADD_TEST_FILENAME_EXTENSION "cu") set(EIGEN_ADD_TEST_FILENAME_EXTENSION "cu")
ei_add_test(cxx11_tensor_complex_cuda) ei_add_test(cxx11_tensor_complex_cuda)
@ -246,18 +265,13 @@ if(CUDA_FOUND AND EIGEN_TEST_CUDA)
ei_add_test(cxx11_tensor_scan_cuda) ei_add_test(cxx11_tensor_scan_cuda)
# Contractions require arch 3.0 or higher # Contractions require arch 3.0 or higher
if (${EIGEN_CUDA_COMPUTE_ARCH} GREATER 29) ei_add_test(cxx11_tensor_device)
ei_add_test(cxx11_tensor_device) ei_add_test(cxx11_tensor_cuda)
ei_add_test(cxx11_tensor_cuda) ei_add_test(cxx11_tensor_contract_cuda)
ei_add_test(cxx11_tensor_contract_cuda) ei_add_test(cxx11_tensor_of_float16_cuda)
ei_add_test(cxx11_tensor_of_float16_cuda)
endif()
# The random number generation code requires arch 3.5 or greater. # The random number generation code requires arch 3.5 or greater.
if (${EIGEN_CUDA_COMPUTE_ARCH} GREATER 34) ei_add_test(cxx11_tensor_random_cuda)
ei_add_test(cxx11_tensor_random_cuda)
endif()
unset(EIGEN_ADD_TEST_FILENAME_EXTENSION) unset(EIGEN_ADD_TEST_FILENAME_EXTENSION)
endif() endif()