Skip to content

Conversation

@iomaganaris
Copy link
Contributor

@iomaganaris iomaganaris commented Feb 8, 2024

  • Added benchmark scripts for P100 (based on the configuration of A100) and H100 GPUs
  • Fixed issue with regex when using --label-regex with sbench-analyze plot
  • Write each benchmark source code in benchmarks_source_code folder for easier inspection
  • Rename GPU kernel functions to corresponding implementation name
  • Removed deprecated gpu_timers parameter
  • Added --generate-line-info compilation option for nvcc to improve source code investigation in NSight Compute. (To be used with --import-source yes in ncu)

@fthaler
Copy link
Contributor

fthaler commented Apr 2, 2024

Could you please split these changes into multiple PRs for easier review? That is, one for the benchmark scripts, one for the fixes, one for the changed kernel names, and one for the updated naming of source files.

@iomaganaris iomaganaris marked this pull request as draft April 2, 2024 12:23
@iomaganaris
Copy link
Contributor Author

Could you please split these changes into multiple PRs for easier review? That is, one for the benchmark scripts, one for the fixes, one for the changed kernel names, and one for the updated naming of source files.

Good point 👍 I split this PR to the following ones:

@fthaler
Copy link
Contributor

fthaler commented Jun 24, 2024

This PR was split into #17, #18, #19, and #20, thus closing now.

@fthaler fthaler closed this Jun 24, 2024
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