arekpaterek/Faster_SGEMM_CUDA

4 best points · 4 comments · 1 Hacker News post · View on GitHub

Hacker News discussions

• Show HN: FP32 matmul of large matrices up to 24% faster than cuBLAS on a 4090 (4 points, 4 comments, by ap4, 2024-07-31)