Skip to content

feat: add analyze_speedup_v6 with integrated token analysis - #49

Closed
factnn wants to merge 1 commit into
flagos-ai:mainfrom
factnn:feature/add-analyze-speedup-v6-token
Closed

feat: add analyze_speedup_v6 with integrated token analysis#49
factnn wants to merge 1 commit into
flagos-ai:mainfrom
factnn:feature/add-analyze-speedup-v6-token

Conversation

@factnn

@factnn factnn commented Jun 26, 2026

Copy link
Copy Markdown
Collaborator

Summary

  • Add analyze_speedup_v6.py: combines speedup analysis and token usage statistics in a single script
  • Add analyze_tokens.py: token counting utility that scans workspaces for CC/OC output files
  • Merged from feature/agent-bench-multi-chips branch

Key Features

  • Speedup analysis with robust statistics (geometric mean, median, IQM)
  • Per-type breakdown (aten / cublas / vllm13)
  • Integrated token usage per operator and per category
  • Anti-hack filtering support
  • Markdown report generation

Test plan

  • Run on existing result directories to verify speedup + token output
  • Verify compatibility with KernelGenBench dataset (210 ops)

🤖 Generated with Claude Code

Merge analyze_speedup_v6.py and analyze_tokens.py from feature/agent-bench-multi-chips.
v6 now supports both speedup analysis and token usage statistics in a single script.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@factnn factnn closed this Jun 26, 2026
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.

1 participant