Skip to content

O
opencv
Commit c4246bc5 authored by Vladislav Vinogradov's avatar Vladislav Vinogradov
Browse files
Options

update FindCUDA CMake module to the latest version from upstream

parent fd6ef87c
Show whitespace changes
Inline Side-by-side
Showing with 110 additions and 240 deletions
cmake/FindCUDA.cmake
View file @ c4246bc5
......@@ -31,10 +31,8 @@
# The following variables affect the behavior of the macros in the
# script (in alphebetical order). Note that any of these flags can be
# changed multiple times in the same directory before calling
# CUDA_ADD_EXECUTABLE, CUDA_ADD_LIBRARY, CUDA_COMPILE, CUDA_COMPILE_PTX
# or CUDA_WRAP_SRCS.
#
# ::
# CUDA_ADD_EXECUTABLE, CUDA_ADD_LIBRARY, CUDA_COMPILE, CUDA_COMPILE_PTX,
# CUDA_COMPILE_FATBIN, CUDA_COMPILE_CUBIN or CUDA_WRAP_SRCS::
#
# CUDA_64_BIT_DEVICE_CODE (Default matches host bit size)
# -- Set to ON to compile for 64 bit device code, OFF for 32 bit device code.
......@@ -43,19 +41,11 @@
# nvcc in the generated source. If you compile to PTX and then load the
# file yourself, you can mix bit sizes between device and host.
#
#
#
# ::
#
# CUDA_ATTACH_VS_BUILD_RULE_TO_CUDA_FILE (Default ON)
# -- Set to ON if you want the custom build rule to be attached to the source
# file in Visual Studio. Turn OFF if you add the same cuda file to multiple
# targets.
#
#
#
# ::
#
# This allows the user to build the target from the CUDA file; however, bad
# things can happen if the CUDA source file is added to multiple targets.
# When performing parallel builds it is possible for the custom build
......@@ -68,44 +58,24 @@
# this script could detect the reuse of source files across multiple targets
# and turn the option off for the user, but no good solution could be found.
#
#
#
# ::
#
# CUDA_BUILD_CUBIN (Default OFF)
# -- Set to ON to enable and extra compilation pass with the -cubin option in
# Device mode. The output is parsed and register, shared memory usage is
# printed during build.
#
#
#
# ::
#
# CUDA_BUILD_EMULATION (Default OFF for device mode)
# -- Set to ON for Emulation mode. -D_DEVICEEMU is defined for CUDA C files
# when CUDA_BUILD_EMULATION is TRUE.
#
#
#
# ::
#
# CUDA_GENERATED_OUTPUT_DIR (Default CMAKE_CURRENT_BINARY_DIR)
# -- Set to the path you wish to have the generated files placed. If it is
# blank output files will be placed in CMAKE_CURRENT_BINARY_DIR.
# Intermediate files will always be placed in
# CMAKE_CURRENT_BINARY_DIR/CMakeFiles.
#
#
#
# ::
#
# CUDA_HOST_COMPILATION_CPP (Default ON)
# -- Set to OFF for C compilation of host code.
#
#
#
# ::
#
# CUDA_HOST_COMPILER (Default CMAKE_C_COMPILER, $(VCInstallDir)/bin for VS)
# -- Set the host compiler to be used by nvcc. Ignored if -ccbin or
# --compiler-bindir is already present in the CUDA_NVCC_FLAGS or
......@@ -113,19 +83,11 @@
# $(VCInstallDir)/bin is a special value that expands out to the path when
# the command is run from withing VS.
#
#
#
# ::
#
# CUDA_NVCC_FLAGS
# CUDA_NVCC_FLAGS_<CONFIG>
# -- Additional NVCC command line arguments. NOTE: multiple arguments must be
# semi-colon delimited (e.g. --compiler-options;-Wall)
#
#
#
# ::
#
# CUDA_PROPAGATE_HOST_FLAGS (Default ON)
# -- Set to ON to propagate CMAKE_{C,CXX}_FLAGS and their configuration
# dependent counterparts (e.g. CMAKE_C_FLAGS_DEBUG) automatically to the
......@@ -137,10 +99,6 @@
# CUDA_ADD_LIBRARY, CUDA_ADD_EXECUTABLE, or CUDA_WRAP_SRCS. Flags used for
# shared library compilation are not affected by this flag.
#
#
#
# ::
#
# CUDA_SEPARABLE_COMPILATION (Default OFF)
# -- If set this will enable separable compilation for all CUDA runtime object
# files. If used outside of CUDA_ADD_EXECUTABLE and CUDA_ADD_LIBRARY
......@@ -148,38 +106,22 @@
# CUDA_COMPUTE_SEPARABLE_COMPILATION_OBJECT_FILE_NAME and
# CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS should be called.
#
#
#
# ::
#
# CUDA_VERBOSE_BUILD (Default OFF)
# -- Set to ON to see all the commands used when building the CUDA file. When
# using a Makefile generator the value defaults to VERBOSE (run make
# VERBOSE=1 to see output), although setting CUDA_VERBOSE_BUILD to ON will
# always print the output.
#
#
#
# The script creates the following macros (in alphebetical order):
#
# ::
# The script creates the following macros (in alphebetical order)::
#
# CUDA_ADD_CUFFT_TO_TARGET( cuda_target )
# -- Adds the cufft library to the target (can be any target). Handles whether
# you are in emulation mode or not.
#
#
#
# ::
#
# CUDA_ADD_CUBLAS_TO_TARGET( cuda_target )
# -- Adds the cublas library to the target (can be any target). Handles
# whether you are in emulation mode or not.
#
#
#
# ::
#
# CUDA_ADD_EXECUTABLE( cuda_target file0 file1 ...
# [WIN32] [MACOSX_BUNDLE] [EXCLUDE_FROM_ALL] [OPTIONS ...] )
# -- Creates an executable "cuda_target" which is made up of the files
......@@ -193,42 +135,28 @@
# nvcc. Such flags should be modified before calling CUDA_ADD_EXECUTABLE,
# CUDA_ADD_LIBRARY or CUDA_WRAP_SRCS.
#
#
#
# ::
#
# CUDA_ADD_LIBRARY( cuda_target file0 file1 ...
# [STATIC | SHARED | MODULE] [EXCLUDE_FROM_ALL] [OPTIONS ...] )
# -- Same as CUDA_ADD_EXECUTABLE except that a library is created.
#
#
#
# ::
#
# CUDA_BUILD_CLEAN_TARGET()
# -- Creates a convience target that deletes all the dependency files
# generated. You should make clean after running this target to ensure the
# dependency files get regenerated.
#
#
#
# ::
#
# CUDA_COMPILE( generated_files file0 file1 ... [STATIC | SHARED | MODULE]
# [OPTIONS ...] )
# -- Returns a list of generated files from the input source files to be used
# with ADD_LIBRARY or ADD_EXECUTABLE.
#
#
#
# ::
#
# CUDA_COMPILE_PTX( generated_files file0 file1 ... [OPTIONS ...] )
# -- Returns a list of PTX files generated from the input source files.
#
# CUDA_COMPILE_FATBIN( generated_files file0 file1 ... [OPTIONS ...] )
# -- Returns a list of FATBIN files generated from the input source files.
#
#
# ::
# CUDA_COMPILE_CUBIN( generated_files file0 file1 ... [OPTIONS ...] )
# -- Returns a list of CUBIN files generated from the input source files.
#
# CUDA_COMPUTE_SEPARABLE_COMPILATION_OBJECT_FILE_NAME( output_file_var
# cuda_target
......@@ -242,10 +170,6 @@
# automatically for CUDA_ADD_LIBRARY and CUDA_ADD_EXECUTABLE. Note that
# this is a function and not a macro.
#
#
#
# ::
#
# CUDA_INCLUDE_DIRECTORIES( path0 path1 ... )
# -- Sets the directories that should be passed to nvcc
# (e.g. nvcc -Ipath0 -Ipath1 ... ). These paths usually contain other .cu
......@@ -253,17 +177,9 @@
#
#
#
#
#
# ::
#
# CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS( output_file_var cuda_target
# nvcc_flags object_files)
#
#
#
# ::
#
# -- Generates the link object required by separable compilation from the given
# object files. This is called automatically for CUDA_ADD_EXECUTABLE and
# CUDA_ADD_LIBRARY, but can be called manually when using CUDA_WRAP_SRCS
......@@ -273,91 +189,51 @@
# specified by CUDA_64_BIT_DEVICE_CODE. Note that this is a function
# instead of a macro.
#
#
#
# ::
#
# CUDA_WRAP_SRCS ( cuda_target format generated_files file0 file1 ...
# [STATIC | SHARED | MODULE] [OPTIONS ...] )
# -- This is where all the magic happens. CUDA_ADD_EXECUTABLE,
# CUDA_ADD_LIBRARY, CUDA_COMPILE, and CUDA_COMPILE_PTX all call this
# function under the hood.
#
#
#
# ::
#
# Given the list of files (file0 file1 ... fileN) this macro generates
# custom commands that generate either PTX or linkable objects (use "PTX" or
# "OBJ" for the format argument to switch). Files that don't end with .cu
# or have the HEADER_FILE_ONLY property are ignored.
#
#
#
# ::
#
# The arguments passed in after OPTIONS are extra command line options to
# give to nvcc. You can also specify per configuration options by
# specifying the name of the configuration followed by the options. General
# options must preceed configuration specific options. Not all
# configurations need to be specified, only the ones provided will be used.
#
#
#
# ::
#
# OPTIONS -DFLAG=2 "-DFLAG_OTHER=space in flag"
# DEBUG -g
# RELEASE --use_fast_math
# RELWITHDEBINFO --use_fast_math;-g
# MINSIZEREL --use_fast_math
#
#
#
# ::
#
# For certain configurations (namely VS generating object files with
# CUDA_ATTACH_VS_BUILD_RULE_TO_CUDA_FILE set to ON), no generated file will
# be produced for the given cuda file. This is because when you add the
# cuda file to Visual Studio it knows that this file produces an object file
# and will link in the resulting object file automatically.
#
#
#
# ::
#
# This script will also generate a separate cmake script that is used at
# build time to invoke nvcc. This is for several reasons.
#
#
#
# ::
#
# 1. nvcc can return negative numbers as return values which confuses
# Visual Studio into thinking that the command succeeded. The script now
# checks the error codes and produces errors when there was a problem.
#
#
#
# ::
#
# 2. nvcc has been known to not delete incomplete results when it
# encounters problems. This confuses build systems into thinking the
# target was generated when in fact an unusable file exists. The script
# now deletes the output files if there was an error.
#
#
#
# ::
#
# 3. By putting all the options that affect the build into a file and then
# make the build rule dependent on the file, the output files will be
# regenerated when the options change.
#
#
#
# ::
#
# This script also looks at optional arguments STATIC, SHARED, or MODULE to
# determine when to target the object compilation for a shared library.
# BUILD_SHARED_LIBS is ignored in CUDA_WRAP_SRCS, but it is respected in
......@@ -366,27 +242,17 @@
# <target_name>_EXPORTS is defined when a shared library compilation is
# detected.
#
#
#
# ::
#
# Flags passed into add_definitions with -D or /D are passed along to nvcc.
#
#
#
# The script defines the following variables:
#
# ::
# The script defines the following variables::
#
# CUDA_VERSION_MAJOR -- The major version of cuda as reported by nvcc.
# CUDA_VERSION_MINOR -- The minor version.
# CUDA_VERSION
# CUDA_VERSION_STRING -- CUDA_VERSION_MAJOR.CUDA_VERSION_MINOR
#
#
#
# ::
#
# CUDA_TOOLKIT_ROOT_DIR -- Path to the CUDA Toolkit (defined if not set).
# CUDA_SDK_ROOT_DIR -- Path to the CUDA SDK. Use this to find files in the
# SDK. This script will not directly support finding
......@@ -412,13 +278,13 @@
# Only available for CUDA version 3.2+.
# CUDA_cusparse_LIBRARY -- CUDA Sparse Matrix library.
# Only available for CUDA version 3.2+.
# CUDA_npp_LIBRARY -- NVIDIA Performance Primitives library.
# CUDA_npp_LIBRARY -- NVIDIA Performance Primitives lib.
# Only available for CUDA version 4.0+.
# CUDA_nppc_LIBRARY -- NVIDIA Performance Primitives library (core).
# CUDA_nppc_LIBRARY -- NVIDIA Performance Primitives lib (core).
# Only available for CUDA version 5.5+.
# CUDA_nppi_LIBRARY -- NVIDIA Performance Primitives library (image processing).
# CUDA_nppi_LIBRARY -- NVIDIA Performance Primitives lib (image processing).
# Only available for CUDA version 5.5+.
# CUDA_npps_LIBRARY -- NVIDIA Performance Primitives library (signal processing).
# CUDA_npps_LIBRARY -- NVIDIA Performance Primitives lib (signal processing).
# Only available for CUDA version 5.5+.
# CUDA_nvcuvenc_LIBRARY -- CUDA Video Encoder library.
# Only available for CUDA version 3.2+.
......@@ -427,32 +293,15 @@
# Only available for CUDA version 3.2+.
# Windows only.
#
#
#
#
#
# ::
#
# James Bigler, NVIDIA Corp (nvidia.com - jbigler)
# Abe Stephens, SCI Institute -- http://www.sci.utah.edu/~abe/FindCuda.html
#
#
#
# ::
#
# Copyright (c) 2008 - 2009 NVIDIA Corporation. All rights reserved.
#
#
#
# ::
#
# Copyright (c) 2007-2009
# Scientific Computing and Imaging Institute, University of Utah
#
#
#
# ::
#
# This code is licensed under the MIT License. See the FindCUDA.cmake script
# for the text of the license.
......@@ -481,11 +330,6 @@
# FindCUDA.cmake
# We need to have at least this version to support the VERSION_LESS argument to 'if' (2.6.2) and unset (2.6.3)
cmake_policy(PUSH)
cmake_minimum_required(VERSION 2.6.3)
cmake_policy(POP)
# This macro helps us find the location of helper files we will need the full path to
macro(CUDA_FIND_HELPER_FILE _name _extension)
set(_full_name "${_name}.${_extension}")
......@@ -608,7 +452,17 @@ set(CUDA_NVCC_FLAGS "" CACHE STRING "Semi-colon delimit multiple arguments.")
if(CMAKE_GENERATOR MATCHES "Visual Studio")
set(CUDA_HOST_COMPILER "$(VCInstallDir)bin" CACHE FILEPATH "Host side compiler used by NVCC")
else()
set(CUDA_HOST_COMPILER "${CMAKE_C_COMPILER}" CACHE FILEPATH "Host side compiler used by NVCC")
# Using cc which is symlink to clang may let NVCC think it is GCC and issue
# unhandled -dumpspecs option to clang. Also in case neither
# CMAKE_C_COMPILER is defined (project does not use C language) nor
# CUDA_HOST_COMPILER is specified manually we should skip -ccbin and let
# nvcc use its own default C compiler.
if(DEFINED CMAKE_C_COMPILER AND NOT DEFINED CUDA_HOST_COMPILER)
get_filename_component(c_compiler_realpath "${CMAKE_C_COMPILER}" REALPATH)
else()
set(c_compiler_realpath "")
endif()
set(CUDA_HOST_COMPILER "${c_compiler_realpath}" CACHE FILEPATH "Host side compiler used by NVCC")
endif()
# Propagate the host flags to the host compiler via -Xcompiler
......@@ -759,15 +613,11 @@ endif()
set(CUDA_VERSION_STRING "${CUDA_VERSION}")
# Support for arm cross compilation with CUDA 5.5
set(__cuda_toolkit_target_dir_initial "${CUDA_TOOLKIT_ROOT_DIR}")
if(CUDA_VERSION VERSION_GREATER "5.0" AND CMAKE_CROSSCOMPILING AND ${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm")
if(ANDROID AND EXISTS "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-androideabi")
set(__cuda_toolkit_target_dir_initial "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-androideabi")
elseif(EXISTS "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-gnueabihf")
set(__cuda_toolkit_target_dir_initial "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-gnueabihf")
endif()
if(CUDA_VERSION VERSION_GREATER "5.0" AND CMAKE_CROSSCOMPILING AND ${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm" AND EXISTS "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-gnueabihf")
set(CUDA_TOOLKIT_TARGET_DIR "${CUDA_TOOLKIT_ROOT_DIR}/targets/armv7-linux-gnueabihf" CACHE PATH "Toolkit target location.")
else()
set(CUDA_TOOLKIT_TARGET_DIR "${CUDA_TOOLKIT_ROOT_DIR}" CACHE PATH "Toolkit target location.")
endif()
set(CUDA_TOOLKIT_TARGET_DIR "${__cuda_toolkit_target_dir_initial}" CACHE PATH "Toolkit target location.")
mark_as_advanced(CUDA_TOOLKIT_TARGET_DIR)
# Target CPU architecture
......@@ -853,18 +703,6 @@ if(CUDA_BUILD_EMULATION AND CUDA_CUDARTEMU_LIBRARY)
else()
set(CUDA_LIBRARIES ${CUDA_CUDART_LIBRARY})
endif()
if(APPLE)
# We need to add the path to cudart to the linker using rpath, since the
# library name for the cuda libraries is prepended with @rpath.
if(CUDA_BUILD_EMULATION AND CUDA_CUDARTEMU_LIBRARY)
get_filename_component(_cuda_path_to_cudart "${CUDA_CUDARTEMU_LIBRARY}" PATH)
else()
get_filename_component(_cuda_path_to_cudart "${CUDA_CUDART_LIBRARY}" PATH)
endif()
if(_cuda_path_to_cudart)
list(APPEND CUDA_LIBRARIES -Wl,-rpath "-Wl,${_cuda_path_to_cudart}")
endif()
endif()
# 1.1 toolkit on linux doesn't appear to have a separate library on
# some platforms.
......@@ -1044,15 +882,15 @@ macro(CUDA_GET_SOURCES_AND_OPTIONS _sources _cmake_options _options)
set( ${_options} )
set( _found_options FALSE )
foreach(arg ${ARGN})
if(arg STREQUAL "OPTIONS")
if("x${arg}" STREQUAL "xOPTIONS")
set( _found_options TRUE )
elseif(
arg STREQUAL "WIN32" OR
arg STREQUAL "MACOSX_BUNDLE" OR
arg STREQUAL "EXCLUDE_FROM_ALL" OR
arg STREQUAL "STATIC" OR
arg STREQUAL "SHARED" OR
arg STREQUAL "MODULE"
"x${arg}" STREQUAL "xWIN32" OR
"x${arg}" STREQUAL "xMACOSX_BUNDLE" OR
"x${arg}" STREQUAL "xEXCLUDE_FROM_ALL" OR
"x${arg}" STREQUAL "xSTATIC" OR
"x${arg}" STREQUAL "xSHARED" OR
"x${arg}" STREQUAL "xMODULE"
)
list(APPEND ${_cmake_options} ${arg})
else()
......@@ -1148,7 +986,7 @@ function(CUDA_COMPUTE_BUILD_PATH path build_path)
endif()
endif()
# This recipie is from cmLocalGenerator::CreateSafeUniqueObjectFileName in the
# This recipe is from cmLocalGenerator::CreateSafeUniqueObjectFileName in the
# CMake source.
# Remove leading /
......@@ -1177,7 +1015,7 @@ endfunction()
# a .cpp or .ptx file.
# INPUT:
# cuda_target - Target name
# format - PTX or OBJ
# format - PTX, CUBIN, FATBIN or OBJ
# FILE1 .. FILEN - The remaining arguments are the sources to be wrapped.
# OPTIONS - Extra options to NVCC
# OUTPUT:
......@@ -1355,7 +1193,7 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
foreach(file ${ARGN})
# Ignore any file marked as a HEADER_FILE_ONLY
get_source_file_property(_is_header ${file} HEADER_FILE_ONLY)
if(${file} MATCHES ".*\\.cu$" AND NOT _is_header)
if(${file} MATCHES "\\.cu$" AND NOT _is_header)
# Allow per source file overrides of the format.
get_source_file_property(_cuda_source_format ${file} CUDA_SOURCE_PROPERTY_FORMAT)
......@@ -1363,16 +1201,22 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
set(_cuda_source_format ${format})
endif()
if( ${_cuda_source_format} MATCHES "OBJ")
set( cuda_compile_to_external_module OFF )
else()
set( cuda_compile_to_external_module ON )
if( ${_cuda_source_format} MATCHES "PTX" )
set( compile_to_ptx ON )
elseif( ${_cuda_source_format} MATCHES "OBJ")
set( compile_to_ptx OFF )
set( cuda_compile_to_external_module_type "ptx" )
elseif( ${_cuda_source_format} MATCHES "CUBIN")
set( cuda_compile_to_external_module_type "cubin" )
elseif( ${_cuda_source_format} MATCHES "FATBIN")
set( cuda_compile_to_external_module_type "fatbin" )
else()
message( FATAL_ERROR "Invalid format flag passed to CUDA_WRAP_SRCS for file '${file}': '${_cuda_source_format}'. Use OBJ or PTX.")
message( FATAL_ERROR "Invalid format flag passed to CUDA_WRAP_SRCS for file '${file}': '${_cuda_source_format}'. Use OBJ, PTX, CUBIN or FATBIN.")
endif()
endif()
if(compile_to_ptx)
if(cuda_compile_to_external_module)
# Don't use any of the host compilation flags for PTX targets.
set(CUDA_HOST_FLAGS)
set(CUDA_NVCC_FLAGS_CONFIG)
......@@ -1387,7 +1231,7 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
if(CUDA_GENERATED_OUTPUT_DIR)
set(cuda_compile_output_dir "${CUDA_GENERATED_OUTPUT_DIR}")
else()
if ( compile_to_ptx )
if ( cuda_compile_to_external_module )
set(cuda_compile_output_dir "${CMAKE_CURRENT_BINARY_DIR}")
else()
set(cuda_compile_output_dir "${cuda_compile_intermediate_directory}")
......@@ -1397,10 +1241,10 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
# Add a custom target to generate a c or ptx file. ######################
get_filename_component( basename ${file} NAME )
if( compile_to_ptx )
if( cuda_compile_to_external_module )
set(generated_file_path "${cuda_compile_output_dir}")
set(generated_file_basename "${cuda_target}_generated_${basename}.ptx")
set(format_flag "-ptx")
set(generated_file_basename "${cuda_target}_generated_${basename}.${cuda_compile_to_external_module_type}")
set(format_flag "-${cuda_compile_to_external_module_type}")
file(MAKE_DIRECTORY "${cuda_compile_output_dir}")
else()
set(generated_file_path "${cuda_compile_output_dir}/${CMAKE_CFG_INTDIR}")
......@@ -1423,7 +1267,7 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
set(custom_target_script "${cuda_compile_intermediate_directory}/${generated_file_basename}.cmake")
# Setup properties for obj files:
if( NOT compile_to_ptx )
if( NOT cuda_compile_to_external_module )
set_source_files_properties("${generated_file}"
PROPERTIES
EXTERNAL_OBJECT true # This is an object file not to be compiled, but only be linked.
......@@ -1438,7 +1282,7 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
set(source_file "${CMAKE_CURRENT_SOURCE_DIR}/${file}")
endif()
if( NOT compile_to_ptx AND CUDA_SEPARABLE_COMPILATION)
if( NOT cuda_compile_to_external_module AND CUDA_SEPARABLE_COMPILATION)
list(APPEND ${cuda_target}_SEPARABLE_COMPILATION_OBJECTS "${generated_file}")
endif()
......@@ -1455,7 +1299,7 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
# Build the NVCC made dependency file ###################################
set(build_cubin OFF)
if ( NOT CUDA_BUILD_EMULATION AND CUDA_BUILD_CUBIN )
if ( NOT compile_to_ptx )
if ( NOT cuda_compile_to_external_module )
set ( build_cubin ON )
endif()
endif()
......@@ -1482,8 +1326,8 @@ macro(CUDA_WRAP_SRCS cuda_target format generated_files)
# Create up the comment string
file(RELATIVE_PATH generated_file_relative_path "${CMAKE_BINARY_DIR}" "${generated_file}")
if(compile_to_ptx)
set(cuda_build_comment_string "Building NVCC ptx file ${generated_file_relative_path}")
if(cuda_compile_to_external_module)
set(cuda_build_comment_string "Building NVCC ${cuda_compile_to_external_module_type} file ${generated_file_relative_path}")
else()
set(cuda_build_comment_string "Building NVCC (${cuda_build_type}) object ${generated_file_relative_path}")
endif()
......@@ -1576,18 +1420,27 @@ function(CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS output_file cuda_target options
# If -ccbin, --compiler-bindir has been specified, don't do anything. Otherwise add it here.
list( FIND nvcc_flags "-ccbin" ccbin_found0 )
list( FIND nvcc_flags "--compiler-bindir" ccbin_found1 )
if( ccbin_found0 LESS 0 AND ccbin_found1 LESS 0 )
if( ccbin_found0 LESS 0 AND ccbin_found1 LESS 0 AND CUDA_HOST_COMPILER )
list(APPEND nvcc_flags -ccbin "\"${CUDA_HOST_COMPILER}\"")
endif()
# Create a list of flags specified by CUDA_NVCC_FLAGS_${CONFIG}
set(config_specific_flags)
set(flags)
foreach(config ${CUDA_configuration_types})
string(TOUPPER ${config} config_upper)
# Add config specific flags
foreach(f ${CUDA_NVCC_FLAGS_${config_upper}})
list(APPEND config_specific_flags $<$<CONFIG:${config}>:${f}>)
endforeach()
set(important_host_flags)
_cuda_get_important_host_flags(important_host_flags ${CMAKE_${CUDA_C_OR_CXX}_FLAGS_${config_upper}})
foreach(f ${important_host_flags})
list(APPEND flags $<$<CONFIG:${config}>:-Xcompiler> $<$<CONFIG:${config}>:${f}>)
endforeach()
endforeach()
# Add our general CUDA_NVCC_FLAGS with the configuration specifig flags
set(nvcc_flags ${CUDA_NVCC_FLAGS} ${config_specific_flags} ${nvcc_flags})
file(RELATIVE_PATH output_file_relative_path "${CMAKE_BINARY_DIR}" "${output_file}")
# Some generators don't handle the multiple levels of custom command
......@@ -1713,21 +1566,29 @@ endmacro()
###############################################################################
###############################################################################
# CUDA COMPILE
# (Internal) helper for manually added cuda source files with specific targets
###############################################################################
###############################################################################
macro(CUDA_COMPILE generated_files)
macro(cuda_compile_base cuda_target format generated_files)
# Separate the sources from the options
CUDA_GET_SOURCES_AND_OPTIONS(_sources _cmake_options _options ${ARGN})
# Create custom commands and targets for each file.
CUDA_WRAP_SRCS( cuda_compile OBJ _generated_files ${_sources} ${_cmake_options}
CUDA_WRAP_SRCS( ${cuda_target} ${format} _generated_files ${_sources} ${_cmake_options}
OPTIONS ${_options} )
set( ${generated_files} ${_generated_files})
endmacro()
###############################################################################
###############################################################################
# CUDA COMPILE
###############################################################################
###############################################################################
macro(CUDA_COMPILE generated_files)
cuda_compile_base(cuda_compile OBJ ${generated_files} ${ARGN})
endmacro()
###############################################################################
###############################################################################
......@@ -1735,17 +1596,28 @@ endmacro()
###############################################################################
###############################################################################
macro(CUDA_COMPILE_PTX generated_files)
cuda_compile_base(cuda_compile_ptx PTX ${generated_files} ${ARGN})
endmacro()
# Separate the sources from the options
CUDA_GET_SOURCES_AND_OPTIONS(_sources _cmake_options _options ${ARGN})
# Create custom commands and targets for each file.
CUDA_WRAP_SRCS( cuda_compile_ptx PTX _generated_files ${_sources} ${_cmake_options}
OPTIONS ${_options} )
set( ${generated_files} ${_generated_files})
###############################################################################
###############################################################################
# CUDA COMPILE FATBIN
###############################################################################
###############################################################################
macro(CUDA_COMPILE_FATBIN generated_files)
cuda_compile_base(cuda_compile_fatbin FATBIN ${generated_files} ${ARGN})
endmacro()
###############################################################################
###############################################################################
# CUDA COMPILE CUBIN
###############################################################################
###############################################################################
macro(CUDA_COMPILE_CUBIN generated_files)
cuda_compile_base(cuda_compile_cubin CUBIN ${generated_files} ${ARGN})
endmacro()
###############################################################################
###############################################################################
# CUDA ADD CUFFT TO TARGET
......
cmake/FindCUDA/make2cmake.cmake
View file @ c4246bc5
......@@ -37,12 +37,11 @@
file(READ ${input_file} depend_text)
if (${depend_text} MATCHES ".+")
if (NOT "${depend_text}" STREQUAL "")
# message("FOUND DEPENDS")
# Remember, four backslashes is escaped to one backslash in the string.
string(REGEX REPLACE "\\\\ " " " depend_text ${depend_text})
string(REPLACE "\\ " " " depend_text ${depend_text})
# This works for the nvcc -M generated dependency files.
string(REGEX REPLACE "^.* : " "" depend_text ${depend_text})
......
cmake/FindCUDA/parse_cubin.cmake
View file @ c4246bc5
......@@ -37,11 +37,10 @@
file(READ ${input_file} file_text)
if (${file_text} MATCHES ".+")
if (NOT "${file_text}" STREQUAL "")
# Remember, four backslashes is escaped to one backslash in the string.
string(REGEX REPLACE ";" "\\\\;" file_text ${file_text})
string(REGEX REPLACE "\ncode" ";code" file_text ${file_text})
string(REPLACE ";" "\\;" file_text ${file_text})
string(REPLACE "\ncode" ";code" file_text ${file_text})
list(LENGTH file_text len)
......@@ -57,7 +56,7 @@ if (${file_text} MATCHES ".+")
# Extract kernel names.
if (${entry} MATCHES "[^g]name = ([^ ]+)")
string(REGEX REPLACE ".* = ([^ ]+)" "\\1" entry ${entry})
set(entry "${CMAKE_MATCH_1}")
# Check to see if the kernel name starts with "_"
set(skip FALSE)
......@@ -76,19 +75,19 @@ if (${file_text} MATCHES ".+")
# Registers
if (${entry} MATCHES "reg([ ]+)=([ ]+)([^ ]+)")
string(REGEX REPLACE ".*([ ]+)=([ ]+)([^ ]+)" "\\3" entry ${entry})
set(entry "${CMAKE_MATCH_3}")
message("Registers: ${entry}")
endif()
# Local memory
if (${entry} MATCHES "lmem([ ]+)=([ ]+)([^ ]+)")
string(REGEX REPLACE ".*([ ]+)=([ ]+)([^ ]+)" "\\3" entry ${entry})
set(entry "${CMAKE_MATCH_3}")
message("Local: ${entry}")
endif()
# Shared memory
if (${entry} MATCHES "smem([ ]+)=([ ]+)([^ ]+)")
string(REGEX REPLACE ".*([ ]+)=([ ]+)([^ ]+)" "\\3" entry ${entry})
set(entry "${CMAKE_MATCH_3}")
message("Shared: ${entry}")
endif()
......
cmake/FindCUDA/run_nvcc.cmake
View file @ c4246bc5
......@@ -62,7 +62,7 @@ set(cmake_dependency_file "@cmake_dependency_file@") # path
set(CUDA_make2cmake "@CUDA_make2cmake@") # path
set(CUDA_parse_cubin "@CUDA_parse_cubin@") # path
set(build_cubin @build_cubin@) # bool
set(CUDA_HOST_COMPILER "@CUDA_HOST_COMPILER@") # bool
set(CUDA_HOST_COMPILER "@CUDA_HOST_COMPILER@") # path
# We won't actually use these variables for now, but we need to set this, in
# order to force this file to be run again if it changes.
set(generated_file_path "@generated_file_path@") # path
......@@ -106,7 +106,7 @@ list(APPEND CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS_${build_configuration}})
# Any -ccbin existing in CUDA_NVCC_FLAGS gets highest priority
list( FIND CUDA_NVCC_FLAGS "-ccbin" ccbin_found0 )
list( FIND CUDA_NVCC_FLAGS "--compiler-bindir" ccbin_found1 )
if( ccbin_found0 LESS 0 AND ccbin_found1 LESS 0 )
if( ccbin_found0 LESS 0 AND ccbin_found1 LESS 0 AND CUDA_HOST_COMPILER )
if (CUDA_HOST_COMPILER STREQUAL "$(VCInstallDir)bin" AND DEFINED CCBIN)
set(CCBIN -ccbin "${CCBIN}")
else()
......@@ -126,7 +126,7 @@ endif()
# and other return variables are present after executing the process.
macro(cuda_execute_process status command)
set(_command ${command})
if(NOT _command STREQUAL "COMMAND")
if(NOT "x${_command}" STREQUAL "xCOMMAND")
message(FATAL_ERROR "Malformed call to cuda_execute_process. Missing COMMAND as second argument. (command = ${command})")
endif()
if(verbose)
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment