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

Beignet Worksize trial-and-error bug #29

Closed
sjkelly opened this issue Apr 16, 2019 · 2 comments
Closed

Beignet Worksize trial-and-error bug #29

sjkelly opened this issue Apr 16, 2019 · 2 comments

Comments

@sjkelly
Copy link
Member

sjkelly commented Apr 16, 2019

On beignet with HD620 gpu:

Beignet: "unable to find good values for local_work_size[i], please provide\n" " local_work_size[] explicitly, you can find good values with\n" " trial-and-error method."

test/example001.jl will trigger this.

This is not an issue on Beignet on Sandy Bridge, presumably because it uses OpenCL -> SIMD.

@sjkelly
Copy link
Member Author

sjkelly commented Apr 20, 2019

It seems like the "everything inline" approach is pressuring registers. Separating the primitive distance fields into different kernels might fix this: https://stackoverflow.com/questions/9504828/write-multiple-kernels-or-a-single-kernel

@sjkelly
Copy link
Member Author

sjkelly commented Jun 28, 2019

no longer using opencl

@sjkelly sjkelly closed this as completed Jun 28, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant