科研速览 · Science Skim继续刷下去 · Keep skimming →
◇ arXiv2026-09-17· cs.DC

How Much of a Real Workload Can LLM-Generated GPU Kernels Actually Reach?

Gaurav Agarwal, Ashish Garg, Isha Singhal

原始摘要(英文原文)· Original abstract
Language models can now write GPU kernels that outperform PyTorch. We evaluate five model configurations on KernelBench level 1 and find that a frontier model produces correct kernels for 91.1% of problems and independently verified speedups on 22 of 56, including three convolutions, with a median of 1.235x. Open-weights models are far behind: the best reaches 30.4% correct with three verified speedups and solves zero convolutions. We then ask a question the literature does not: what fraction of a real model's wall clock do such kernels govern? Profiling seven workloads across three domains, we find the addressable fraction ranges from 8.9% to 58.2%. On transformers, 80-86% of runtime is spent in cuBLAS GEMM and FlashAttention, bounding realistic end-to-end improvement at roughly 1%, and the fraction shrinks with model scale. On recommenders it is 58.2%, concentrated in a single embedding kernel. We introduce DLRM-Bench, 12 recommender kernel problems in KernelBench format, and measure a 41.7% win rate at a 1.552x median there, projecting 8.63% end-to-end. Separately, we show that KernelBench's correctness check (torch.allclose with an absolute tolerance) is satisfied by a tensor of zeros on 4 of 60 level-1 problems. Two kernels in our own results exploited this before we detected them, including one scored at 283x that wrote 0.3% of its output buffer. We propose scale-invariant replacements and release all 879 evaluations.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

How Much of a Real Workload Can LLM-Generated GPU Kernels Actually Reach? — 科研速览 Science Skim