cjolivier01 closed pull request #9362: Fix CUDA auto-detect and cxx flags
URL: https://github.com/apache/incubator-mxnet/pull/9362
This is a PR merged from a forked repository.
As GitHub hides the original diff on merge, it is displayed below for
the sake of provenance:
As this is a foreign pull request (from a fork), the diff is supplied
below (as it won't show otherwise due to GitHub magic):
diff --git a/cmake/FirstClassLangCuda.cmake b/cmake/FirstClassLangCuda.cmake
index 2d2662d616..4d89fc1d04 100644
--- a/cmake/FirstClassLangCuda.cmake
+++ b/cmake/FirstClassLangCuda.cmake
@@ -19,6 +19,9 @@
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag("-std=c++11" SUPPORT_CXX11)
+if(USE_CXX14_IF_AVAILABLE)
+ check_cxx_compiler_flag("-std=c++14" SUPPORT_CXX14)
+endif()
################################################################################################
# Short command for cuDNN detection. Believe it soon will be a part of CUDA
toolkit distribution.
@@ -73,9 +76,9 @@ function(mshadow_detect_installed_gpus out_variable)
"}\n")
enable_language(CUDA)
- try_run(__nvcc_res __compile_result ${PROJECT_BINARY_DIR} ${file}
- COMPILE_OUTPUT_VARIABLE __compile_out
- RUN_OUTPUT_VARIABLE __nvcc_out)
+ try_run(__nvcc_res __compile_result ${PROJECT_BINARY_DIR} ${__cufile}
+ COMPILE_OUTPUT_VARIABLE __compile_out
+ RUN_OUTPUT_VARIABLE __nvcc_out)
if(__nvcc_res EQUAL 0 AND __compile_result)
# nvcc outputs text containing line breaks when building with MSVC.
@@ -90,8 +93,8 @@ function(mshadow_detect_installed_gpus out_variable)
endif()
if(NOT CUDA_gpu_detect_output)
- message(WARNING "Automatic GPU detection failed. Building for all known
architectures (${mshadow_known_gpu_archs}).")
- set(${out_variable} ${mshadow_known_gpu_archs} PARENT_SCOPE)
+ message(WARNING "Automatic GPU detection failed. Building for all known
architectures (${mxnet_known_gpu_archs}).")
+ set(${out_variable} ${mxnet_known_gpu_archs} PARENT_SCOPE)
else()
set(${out_variable} ${CUDA_gpu_detect_output} PARENT_SCOPE)
endif()
@@ -121,12 +124,12 @@ endif ()
# Usage:
# mshadow_select_nvcc_arch_flags(out_variable)
function(mshadow_select_nvcc_arch_flags out_variable)
-
- set(CUDA_ARCH_LIST "Common" CACHE STRING "Select target NVIDIA GPU
achitecture.")
+
+ set(CUDA_ARCH_LIST "Auto" CACHE STRING "Select target NVIDIA GPU
achitecture.")
set_property( CACHE CUDA_ARCH_LIST PROPERTY STRINGS "" "All" "Common"
${CUDA_KNOWN_GPU_ARCHITECTURES} )
mark_as_advanced(CUDA_ARCH_NAME)
-
-
+
+
if("X${CUDA_ARCH_LIST}" STREQUAL "X" )
set(CUDA_ARCH_LIST "All")
endif()
@@ -134,11 +137,13 @@ function(mshadow_select_nvcc_arch_flags out_variable)
set(cuda_arch_bin)
set(cuda_arch_ptx)
+ message(STATUS " CUDA_ARCH_LIST: ${CUDA_ARCH_LIST}")
if("${CUDA_ARCH_LIST}" STREQUAL "All")
set(CUDA_ARCH_LIST ${CUDA_KNOWN_GPU_ARCHITECTURES})
elseif("${CUDA_ARCH_LIST}" STREQUAL "Common")
set(CUDA_ARCH_LIST ${CUDA_COMMON_GPU_ARCHITECTURES})
- elseif("${CUDA_ARCH_LIST}" STREQUAL "Auto")
+ elseif("${CUDA_ARCH_LIST}" STREQUAL "Auto" OR "${CUDA_ARCH_LIST}" STREQUAL
"")
+ set(mxnet_known_gpu_archs ${CUDA_COMMON_GPU_ARCHITECTURES})
mshadow_detect_installed_gpus(CUDA_ARCH_LIST)
message(STATUS "Autodetected CUDA architecture(s): ${CUDA_ARCH_LIST}")
endif()
@@ -204,7 +209,7 @@ function(mshadow_select_nvcc_arch_flags out_variable)
if(cuda_arch_ptx)
list(REMOVE_DUPLICATES cuda_arch_ptx)
endif()
-
+
message(STATUS "cuda arch bin: ${cuda_arch_bin}")
message(STATUS "cuda arch ptx: ${cuda_arch_ptx}")
set(nvcc_flags "")
@@ -229,6 +234,25 @@ function(mshadow_select_nvcc_arch_flags out_variable)
list(APPEND nvcc_archs_readable compute_${arch})
endforeach()
+ if(NOT MSVC)
+ if(SUPPORT_CXX14)
+ list(APPEND nvcc_flags "-std=c++14")
+ elseif(SUPPORT_CXX11)
+ list(APPEND nvcc_flags "-std=c++11")
+ endif()
+ endif()
+
+ string (REPLACE " " ";" CMAKE_CXX_FLAGS_STR "${CMAKE_CXX_FLAGS}")
+ foreach(_flag ${CMAKE_CXX_FLAGS_STR})
+ # Remove -std=c++XX flags
+ if(NOT "${_flag}" MATCHES "-std=.+")
+ # Remove link flags
+ if(NOT "${_flag}" MATCHES "-Wl,.+")
+ list(APPEND nvcc_flags "-Xcompiler ${_flag}")
+ endif()
+ endif()
+ endforeach()
+
string(REPLACE ";" " " nvcc_archs_readable "${nvcc_archs_readable}")
set(${out_variable} ${nvcc_flags} PARENT_SCOPE)
set(${out_variable}_readable ${nvcc_archs_readable} PARENT_SCOPE)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
