I am new to this forum, please tell me if there is a more appropriate place to post my issue.
I’m trying to get my 3060 on a Windows laptop to work with WSL, following the instructions in CUDA on WSL :: CUDA Toolkit Documentation to install the cuda toolkit 11.7 without the linux drivers attached. I have downloaded and installed the latest windows drivers.
When I try to confirm installation in bash, the command “nvcc --version” returns Command ‘nvcc’ not found.
I then tried to install the toolkit via the runfile here CUDA Toolkit 11.7 Update 1 Downloads | NVIDIA Developer but it did not seem to help.
When I navigate to /usr/local/cuda-11.7/, I notice there is no “bin” folder.
I also tried to install cudnn. I suspect the cudnn errors I have are related to the cuda toolkit errors. I first followed Installation Guide :: NVIDIA Deep Learning cuDNN Documentation for the tar file installation, replacing /usr/local/cuda/ with /usr/local/cuda-11.7/ since …/cuda/ didn’t exist. If I remember correctly, I had to create the “lib64” and “include” directories myself.
It says to verify installation by make-ing something in the /usr/src/cudnn_samples_v8/ folder, but the only file in there is some generic license readme.txt from nvidia.
I then tried the package manager installation from the same tutorial with wsl-ubuntu as the OS, but nothing appeared to change.
in powershell:
wsl cat /proc/version
Linux version 5.10.102.1-microsoft-standard-WSL2 (oe-user@oe-host) (x86_64-msft-linux-gcc (GCC) 9.3.0, GNU ld (GNU Binutils) 2.34.0.20200220) #1 SMP Wed Mar 2 00:30:59 UTC 2022
in wsl:
which nvcc
(returns no result)
$ ls /usr/local/cuda-11.7/
doc extras gds include lib64 targets
I am new to solving these kinds of problems, so if I’ve left out any critical information please tell me. If anyone can help me, I’d really appreciate it.