KernelSeek beats FlashAttention-3 by asking questions, not adding tools
A tool-rich agent stalled at 0.50x after 133 steps; KernelSeek reached 1.13x, then 1.16x and 1.46x on later kernels.
Stories per day · 30 days
2 RSS
5 stories, newest first · one card per story
Stories tagged #kernel-optimization
DuoDiscover searches optimizations with cheap proxy evaluators
Authors say it doubles expert speedup on SWE-fficiency and reports 1.92 times vLLM and 1.38 times SQLite.
MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code
MetaKernelBench measures whether GPU kernel skills transfer across CuTe and TIRx on 74 problems.
Approximating Softmax in Pretrained LLMs: Model Sensitivity and Kernel Acceleration
Approximating softmax in ten frozen LLMs yields a FlashAttention-4 kernel up to 25.8% faster on NVIDIA B200.
Show HN: Agentic CUDA Kernel Optimizer
Hacker News highlights a Show HN project for an agentic CUDA kernel optimizer.