Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add documentation FAQ entry about .nv cache directory #3458

Open
seibert opened this issue Oct 31, 2018 · 0 comments
Open

Add documentation FAQ entry about .nv cache directory #3458

seibert opened this issue Oct 31, 2018 · 0 comments
Labels
CUDA CUDA related issue/PR doc Task

Comments

@seibert
Copy link
Contributor

seibert commented Oct 31, 2018

Based on this post:

https://groups.google.com/a/continuum.io/d/msg/numba-users/DcWYd7IR5HE/MtzPSIzcAQAJ

It sounds like there are situations where PTX that has been JIT compiled by the CUDA driver has been cached to $HOME/.nv (on Linux anyway) and can cause linker problems with Numba. We should add an entry about this to the FAQ, and possibly even detect this error and linux to the FAQ.

For more information about the CUDA JIT cache (not to be confused with the caching Numba does on the CPU), see this NVIDIA blog post:

https://devblogs.nvidia.com/cuda-pro-tip-understand-fat-binaries-jit-caching/

@stuartarchibald stuartarchibald added Task doc CUDA CUDA related issue/PR labels Oct 31, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CUDA CUDA related issue/PR doc Task
Projects
Usability / Docs
Awaiting triage
Development

No branches or pull requests

2 participants