| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
| |
The OLD behaviors of all policies are deprecated, but only by
documentation. Add an explicit deprecation diagnostic for policies
introduced in CMake 3.18 and below to encourage projects to port
away from setting policies to OLD.
|
| |
|
|
|
|
|
| |
With CUDAARCHS there's the possibility of tests being run with multiple
architectures or having CMAKE_CUDA_ARCHITECTURES set with NVCC despite CMP0104
being OLD.
Fix a few tests to work properly in such cases.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the following scenario (with 3.18 policies):
1. A CXX target is created.
2. CUDA language is enabled.
CMake 3.18 introduced CMP0104, which requires CUDA_ARCHITECTURES to be
set. Because the CXX target was created before CUDA was enabled it
wouldn't have it set. The Visual Studio generator would however end up
computing CUDA compile options for the CXX target, which would result in
a fatal error due to the policy violation.
There doesn't seem to be a reason to do this for targets that don't
actually use the CUDA language, so we can skip and generate the CXX
target just fine.
Fixes: #21341
|
| |
|
|
|
|
|
|
|
|
| |
The ability to disable adding architectures completely for packaging purposes
and cases requiring passing the architectures flags explicitly has been
requested.
Support a false value for CUDA_ARCHITECTURES and CMAKE_CUDA_ARCHITECTURES
for this purpose.
Implements #20821.
|
| | |
|
|
|
Simplifies CUDA target architecture handling.
Required for Clang support as Clang doesn't automatically select a supported architecture.
We detect a supported architecture during compiler identification and set CMAKE_CUDA_ARCHITECTURES to it.
Introduces CMP0104 for backwards compatibility with manually setting code generation flags with NVCC.
Implements #17963.
|
