We have some code built with CUDA 11.0 and want to ship to Fedora.
However we can only find 10.2 (only for Fedora 29) and 11.1 (only for Fedora 32) from https://developer.download.nvidia.com/compute/cuda/repos/
Fedora 29 (10.1/10.2):
Fedora 32 (11.1):
Since the ecosystem (cudnn/tensorrt/other 3rd-party libs) is not ready for the newly released 11.1 yet, we have to stay with 11.0 for a while.
Is this gap intentional or by mistake?
Can someone from Nvidia helps bridge the gap?
Thanks