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

Gpu optimize and fix precision loss #30

Merged
merged 5 commits into from
Jul 6, 2024

Conversation

nhatdongdang
Copy link
Collaborator

@nhatdongdang nhatdongdang commented Jul 6, 2024

  • Change parameter type to float*
  • Add compiler flag
  • Fix argmax and softmax precision
  • Use shared memory and local memory

@nhatdongdang nhatdongdang changed the title Gpu optimize Gpu optimize and fix precision loss Jul 6, 2024
@nhatdongdang nhatdongdang force-pushed the feat/optimize-param branch 2 times, most recently from 24f02ec to 6e3ba74 Compare July 6, 2024 03:52
@nhatdongdang nhatdongdang force-pushed the feat/optimize-param branch from 6e3ba74 to 9d479e7 Compare July 6, 2024 03:54
@nhatdongdang nhatdongdang force-pushed the feat/optimize-param branch from 8ada910 to d4b590e Compare July 6, 2024 05:49
@rozukke rozukke self-requested a review July 6, 2024 08:27
Copy link
Contributor

@rozukke rozukke left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good besides a few small changes 👍

CMakeLists.txt Outdated Show resolved Hide resolved
src/main.cu Show resolved Hide resolved
@nhatdongdang nhatdongdang requested a review from rozukke July 6, 2024 08:37
Copy link
Contributor

@rozukke rozukke left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good, cheers

@rozukke rozukke merged commit 02dc834 into kachi-group:cuda-staging Jul 6, 2024
1 check passed
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

Successfully merging this pull request may close these issues.

2 participants