Support building (or rather exporting) targets which will statically link against CUDA libraries #699
Assignees
Labels
Comments
eyalroz
pushed a commit
that referenced
this issue
Dec 5, 2024
* Now exporting `_static` versions of our targets - which indicate the eventual executable should link ingainst the `_static` versions of `CUDA::` targets. * Added an configuration option for building the example programs statically
eyalroz
pushed a commit
that referenced
this issue
Dec 5, 2024
* Now exporting `_static` versions of our targets - which indicate the eventual executable should link ingainst the `_static` versions of `CUDA::` targets. * Added an configuration option for building the example programs statically
eyalroz
pushed a commit
that referenced
this issue
Dec 5, 2024
* Now exporting `_static` versions of our targets - which indicate the eventual executable should link ingainst the `_static` versions of `CUDA::` targets. * Added an configuration option for building the example programs statically
eyalroz
pushed a commit
that referenced
this issue
Dec 5, 2024
* Now exporting `_static` versions of our targets - which indicate the eventual executable should link ingainst the `_static` versions of `CUDA::` targets. * Added an configuration option for building the example programs statically
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
NVIDIA offers kinda-static versions of CUDA libraries, especially the CUDA Runtime library and NVRTC. However - our wrappers opt for linking against the dynamic/shared versions of these libraries. We should allow both a configuration in which the wrappers depend on shared libraries, and one where they depend on the static ones.
The text was updated successfully, but these errors were encountered: