aaa370

3 ms·

1. 1.▲
   How to Write a Fast CUDA Matrix Multiplication with Nvidia Tensor Cores(alexarmbr.github.io)

   2 points by aaa370 2y ago|0 comments
2. 2.▲
   by aaa370 2y ago

   I gotta see it to believe it ;)
3. 3.▲
   by aaa370 2y ago

   Another point to consider here is that this project of writing a cuBLAS level GEMM kernel becomes much more challenging if you are doing it with fp16, and are thus competing with the cuBLAS kernels that use tensor cores. The (theoretical) a