Python · MIT
RightNow-AI/AutoMegaKernel
An agent harness that compiles a model into one provably-correct, self-retargeting CUDA megakernel and self-tunes it past cuBLAS at batch-1 LLM decode, paper: https://arxiv.org/abs/2606.09682
★ 161
Stars
16 forks · 4 open issues
+2.0/wk
Growth · measured
Stars counted between two of our own snapshots roughly a week apart. The repository page gives the exact span.
19.5
Momentum / 100
see the breakdown below
5mo ago
Created
last push 51d ago
Star history
How this score was built
| Component | Weight | Earned | Why |
|---|---|---|---|
| Velocity | 55 | 2.1 | +2.0/wk, log-scaled against a 750/day ceiling |
| Acceleration | 15 | 0.0 | last week ran 0.74× the monthly pace |
| Freshness | 15 | 11.0 | last push 51d ago; decays to zero at 180 days |
| Scale | 15 | 6.4 | 161 stars, log-scaled against 150k — stops a 3-star repo topping the board |
| Momentum | 100 | 19.5 | sum of the above |
Detail
| Topics | agent-harnesscudagpugpu-programmingkernel-fusionllm-inferencemachine-learningmegakernelmlsys |
| Matched on | LLM & Inference (llm-inference) |
| Short window | +2 stars over 7 days = 0.3/day |
| Long window | +10 stars over 26 days = 0.4/day |
| Lifetime average | 1.07 stars/day over 150 days |
| First seen here | 2026-10-11 |
| Last read | 2d ago |