Comment on Writing high-performance matrix multiplication kernels for BlackwellparentComments−flakiness11moSo OpenAI has Triton and Google has Pallas. What's the NVIDIA counterpart?−saagarjha11moTilus/CUTLASS I assume−flakiness11moInteresting: https://github.com/NVIDIA/tilus Thanks for the pointer!
Comments
So OpenAI has Triton and Google has Pallas. What's the NVIDIA counterpart?
Tilus/CUTLASS I assume
Interesting: https://github.com/NVIDIA/tilus Thanks for the pointer!