Bad cuda error when using torch.matmul with CUDA 10.0

My computer refuses to perform the matmul operation. I have torch1.0, python 2.7, CUDA 10.0 and GPUs are Nvidia 2080 RTX.

Is your code working on the CPU?
If so, could you share a (small) executable code snippet so that we could have a look?

Hey thanks for help. I actually figured out how to fix it. I had to compile pytorch for Cuda 10.0. it was currently configured for Cuda 9.0