cuda: set library pathes on non-gpu systems

This commit is contained in:
2024-11-22 10:44:34 +01:00
parent ef76df3a07
commit c9bda4b5ee

View File

@@ -13,15 +13,12 @@ harnessing the power of the graphics processing unit (GPU).
prepend-path CUDA_PATH "$PREFIX"
# libnvidia-ml.so in GDK package is a stub library that is attached only for build purposes
prepend-path LIBRARY_PATH "$PREFIX/lib64/stubs"
prepend-path LIBRARY_PATH "$PREFIX/targets/x86_64-linux/lib/stubs"
if {[catch {exec /usr/sbin/modinfo -d nvidia} result]} {
# add if the module is *not* loaded => we are on a system without GPU
prepend-path LIBRARY_PATH "$PREFIX/lib64/stubs"
prepend-path LD_LIBRARY_PATH "$PREFIX/lib/stubs"
}
prepend-path C_INCLUDE_PATH "$PREFIX/targets/x86_64-linux/include"
prepend-path CPLUS_INCLUDE_PATH "$PREFIX/targets/x86_64-linux/include"
# We should always run with libnvidia-ml.so that is installed with your NVIDIA Display Driver.
# By default it's installed in /usr/lib and /usr/lib64.
# Therefore, LD_LIBRARY_PATH must not contain any libnvidia-ml.so reference in the GDK package
# prepend-path LD_LIBRARY_PATH "$PREFIX/lib/stubs"
# prepend-path LD_LIBRARY_PATH "$PREFIX/targets/x86_64-linux/lib/stubs"