That's a nice tutorial but just to be clear: that is not a deep dive in any sens... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		almostgotcaught on Nov 12, 2024 \| parent \| context \| favorite \| on: Optimizing a WebGPU Matmul Kernel for 1 TFLOP That's a nice tutorial but just to be clear: that is not a deep dive in any sense. It's just the bog standard tricks. It doesn't cover MMA and WMMA, which today is table stakes for fast matmul. Also doesn't cover software pipelining. It's basically a good summary of the basics.

saagarjha on Nov 12, 2024 [–]

It’s a deep dive as of like 2015 probably. I don’t know if anyone has done something similar for modern GEMMs. Maybe the CUTLASS or Colfax people?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact