I’ve installed CUDA Toolkit through several ways now and it keeps giving me problems on Fedora 41. If anyone has any troubleshooting tips that would be greatly appreciated.
Fedora 41 is not a supported OS on Cuda 12.6, the current version.
I think the solution is to use the rpm-fusion drivers and not the official NVIDIA drivers. Will update here later.
EDIT: Yes, I now have Cuda 12.6 working on Fedora 41 with Driver Version: 565.77.
Step-by-step guide:
- Install Nvidia drivers through RPM-Fusion: Howto/NVIDIA - RPM Fusion
- Install CUDA through the official NVIDIA repo: CUDA Toolkit 12.6 Update 3 Downloads | NVIDIA Developer
- Add /usr/local/cuda-12.6/bin to $PATH
- Add /usr/local/cuda-12.6/lib64 to /etc/ld.so.conf and run ldconfig as root
1 Like
If you want to compile stuff you will have to tell it to ignore the compiler version (which then leaves you even more unsupported) or install a specific supported version (maybe clang as it will probably be less disruptive?).
From a previous post:
…alternatively upgrade to F41 but stay with current F41 gcc and you will get probably this error message:
"error – unsupported GNU version! gcc versions later than 13 are not supported!
The nvcc flag ‘-allow-unsupported-compiler’ can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk. "
That will make that error go way but as it says: “use at your own risk.”…