▲ 3 pointsParallelKernelBench: Can LLMs write fast multi-GPU kernels?github.comby matt_d·1mo ago·0 comments·view on hn ↗