CUDA 11.6 added the `nvcc -arch=native` flag to automatically compile
for the host GPUs' architectures. Add support for specifying this
special `native` value in `CMAKE_CUDA_ARCHITECTURES` and
`CUDA_ARCHITECTURES`. During the compiler ABI detection step,
detect the native architectures so we can pass them explicitly
when using Clang or older versions of nvcc.
Fixes: #22375
Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
2021-08-17) added all and all-major options to CUDA_ARCHITECTURES. These are
fairly generic and likely to see real-world use by distributors. Thus it's
desirable to support these also for Clang and older NVCC versions.
The supported architectures are dependent on the toolkit version. We determine
the toolkit version prior to compiler detection. For NVCC we get the version
from the vendor identification output, but for Clang we need to invoke NVCC
separately.
The architecture information is mostly based on the Wikipedia list with the
earliest supported version being CUDA 7.0. This could be documented and
expanded in the future to allow projects to query CUDA toolkit version and
architecture information.
For Clang we additionally constrain based on its support.
Additionally the architecture mismatch detection logic is fixed, improved and
updated for generic support:
* Commit 01428c55 (CUDA: Fail fast if CMAKE_CUDA_ARCHITECTURES doesn't work
during detection, 2020-08-29) enabled CMAKE_CUDA_COMPILER_ID_REQUIRE_SUCCESS
if CMAKE_CUDA_ARCHITECTURES is specified. This results in
CMakeDetermineCompilerID.cmake printing the compiler error and our code for
presenting the mismatch in a user-friendly way being useless. The custom
logic seems preferable so go back to not enabling it.
* Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
2021-08-17) tried to support CMP0054 but forgot to add x to the interpolated
result. Thus the conditions would always evaluate to false. This is fixed as
a byproduct of removing NVIDIA specific checks, improving the error message
and replacing architectures_mode with a simpler architectures_explicit.
Visual Studio support omits testing the flags during detection due to
complexities in determining the toolkit version when using it.
A long-term proper implementation would be #23161.
Implements #22860.
Since commit 0d0145138f (CUDA: Add abstraction for cuda runtime
selection, 2019-11-29, v3.17.0-rc1~83^2) we add CUDA runtime library
selection flags by default.
To maintain backwards compatibility the default CUDA runtime
library needs to be computed based on what libraries are found
on the initial compiler invocation. For example a toolchain
could establish initial flags that have all CUDA compilations
using the runtime version, and if we don't detect this we will
try to link to both the static and shared runtime.
Co-Author: Brad King <brad.king@kitware.com>
Fixes: #20708
`link.exe /lib` is an undocumented flag and it just calls `lib.exe`.
Also `link.exe` doesn't parse the `/lib` option correctly when in a
response file.
The CUDA language failed to persist CMAKE_SIZEOF_VOID_P
and CMAKE_INTERNAL_PLATFORM_ABI, causing find_ calls to fail when
the only enabled language was CUDA. This specifically occurred
when having to locate libraries inside `lib64` directories.
The CUDA language failed to persist CMAKE_CUDA_LIBRARY_ARCHITECTURE
and CMAKE_LIBRARY_ARCHITECTURE, causing find_ calls to fail when
the only enabled language was CUDA.
- CMakeParseImplicitIncludeInfo.cmake: new parser that extracts the
compiler's include path from verbose output. If the parser cannot
parse the output, we fall back to the old behavior. On osx we skip
over framework directories (handled elsewhere).
- CMakeDetermineCompilerABI.cmake:
- use verbose flag in try_compile for ${src}
- use new cmake_parse_implicit_include_info() to attempt extract
implicit include directory path and if successful set
CMAKE_${LANG}_IMPLICIT_INCLUDE_DIRECTORIES
- CMakeCCompiler.cmake.in and CMakeCXXCompiler.cmake.in - preserve
CMAKE_${LANG}_IMPLICIT_INCLUDE_DIRECTORIES value between runs in
the same way CMAKE_${LANG}_IMPLICIT_LINK_DIRECTORIES is preserved
- Tests/RunCMake/ParseImplicitIncludeInfo: tests for parse
based on the older Tests/CMakeTests/ImplicitLinkInfoTest.cmake.in.
The test runs a set of verbose compiler outputs collected from
various machines through the parser and checks the results. New
compiler files can be added by dropping input/output files in the
ParseImplicitIncludeInfo/data subdirectory and then adding the new set
of files to the ${targets} list in ParseImplicitIncludeInfo.cmake.
There is a helper CMakeLists.txt in ParseImplicitIncludeInfo/data
that can help with the generation of test input files.
NOTE: the standard cmake pre-commit hook rejects verbose compiler
output with trailing spaces... you have to manually edit them out.
This shouldn't impact the test.
Note that both the parser and the test code can use CMAKE_${LANG}_COMPILER_*
variables such as ${CMAKE_CXX_COMPILER_ID} to decide how to parse
verbose compiler output. For the test code, this requires us to
save the variables values in the test input files.
Fixes: #16291
CMake commands vs_link_dll and vs_link_exe, performing linking on MSVC,
are responsible for calling resource compiler and manifest tool.
Before this commit, both of these tools were called directly, with the
expectation that they are available in the `PATH`. This has been fixed
by respecting CMake variables `CMAKE_RC_COMPILER` and `CMAKE_MT`
defining paths to these tools.
Fixes: #17804
Previously we had a two issues when building cuda executables
that required separable compilation. The first was that we didn't
propagate FLAGS causing any -arch / -gencode flags to be dropped, and
secondly generators such as ninja would use the CXX language flags
instead of CUDA when the executable was mixed language.
Brad King