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.

Sign up for GitHub

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

Jump to bottom

half2 Intrinsic #2

Open
Zhiwei35 opened this issue Nov 13, 2023 · 1 comment
Open

half2 Intrinsic #2

Zhiwei35 opened this issue Nov 13, 2023 · 1 comment

Comments

@Zhiwei35
Copy link

hello, thanks for the code of GEMV, I see the gemv_fp16 did FMA with fp32 by cast fp16 to fp32, I think maybe its better here to use half2 intrinsic, did you have a try? Thanks~
@TylunasLi
Copy link

I tried different types of half / half2 Intrinsics, on CUDA > 11.0 and spercified compute compatibility, there's no significant difference. cause rhe GEMV kernel is memory bandwidth bound.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@TylunasLi @Zhiwei35