mirror of
https://github.com/Kitware/CMake.git
synced 2026-01-05 21:31:08 -06:00
c4ae9384ff
NVCC's default architecture may be newer than the one supported by the machine's GPU. In such cases it's useful to have an environment variable for initializing CMAKE_CUDA_ARCHITECTURES to avoid specifying it for every invocation.
21 lines
716 B
ReStructuredText
21 lines
716 B
ReStructuredText
CMAKE_CUDA_ARCHITECTURES
|
|
------------------------
|
|
|
|
.. versionadded:: 3.18
|
|
|
|
Default value for :prop_tgt:`CUDA_ARCHITECTURES` property of targets.
|
|
|
|
Initialized by the :envvar:`CUDAARCHS` environment variable if set.
|
|
Otherwise as follows depending on :variable:`CMAKE_CUDA_COMPILER_ID <CMAKE_<LANG>_COMPILER_ID>`:
|
|
|
|
- For ``Clang``: the oldest architecture that works.
|
|
|
|
- For ``NVIDIA``: the default architecture chosen by the compiler.
|
|
See policy :policy:`CMP0104`.
|
|
|
|
Users are encouraged to override this, as the default varies across compilers
|
|
and compiler versions.
|
|
|
|
This variable is used to initialize the :prop_tgt:`CUDA_ARCHITECTURES` property
|
|
on all targets. See the target property for additional information.
|