HipKittens: Fast and furious AMD kernels(hazyresearch.stanford.edu)

223 points by dataminer 11 months ago | 74 comments

GPUs Go Brrr(hazyresearch.stanford.edu)

1056 points by nmstoker 2 years ago | 243 comments

Stanford Hazy Research blog post on what GPU hardware wants for efficient AI kernels, introducing the ThunderKittens CUDA kernel library.

We bought the whole GPU, so we're damn well going to use the whole GPU(hazyresearch.stanford.edu)

472 points by sydriax 1 year ago | 94 comments

Stanford Hazy Research blog post introducing a throughput-optimized megakernel for tensor-parallel Llama-70B inference on H100 GPUs that overlaps compute, memory, and communication.

AMD Releases ROCm Version 6.3(insidehpc.com)

11 points by ankitg12 1 year ago | 5 comments

What happens when a GPU reads memory(blog.doubleword.ai)

100 points by ibobev 2 months ago | 18 comments

Nvidia's $20B antitrust loophole(ossa-ma.github.io)

509 points by ossa-ma 10 months ago | 159 comments

Blog post arguing Nvidia's $20B deal for Groq's assets and hiring of CEO Jonathan Ross exploits an antitrust loophole to avoid merger scrutiny.