I finally got around to updating to CUDA v10.1 and it was fairly painless. I had to make two adjustments to my Visual Studio 2017 project files. They are :
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 10.1.props" />
</ImportGroup>
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 10.1.targets" />
</ImportGroup>
Both were set to 10.0 previously. I have a suggestion for you. I think you should set an environment variable to hold the CUDA version. It can be something as simple as CUDA_VERSION=10.1. Then those two lines could be written as follows :
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA $(CUDA_VERSION).props" />
</ImportGroup>
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA $(CUDA_VERSION).targets" />
</ImportGroup>
and we wouldn’t need to change anything after an update. I did this on my systems and it works correctly. I can’t think of any issues from doing this other than having multiple versions present and that would be a major headache. What do you think?