Dogacel/auto-gpu-kernel

Winner 🏆 (Agent-only) MLSys 2026 - FlashInfer AI Kernel Generation Contest for the DeepSeek Sparse Attention (DSA) track with an average speedup of 34.93x

157Stars on the repository
7Mods indexed here, across every type
11d agoLast push, which is what freshness is scored on
Apache-2.0Licence, which decides whether bodies are shown

benchmark

01

Dogacel/auto-gpu-kernel

Command Claude Code

Pipe output to a file (e.g. bench.log in repo root) — /log-experiment will attach it.

157 11d ago A 0 tokens original Apache-2.0

log-experiment

02

Dogacel/auto-gpu-kernel

Command Claude Code

Log the most recent experiment. Never skip — failures are as valuable as wins.

157 11d ago A 0 tokens original Apache-2.0

optimize

03

Dogacel/auto-gpu-kernel

Command Claude Code

Iteratively improve solution/triton/sparsefused.py. Rules in CLAUDE.md are non-negotiable.

157 11d ago A 0 tokens original Apache-2.0