News for hackers. Stuff that matters.

A comment on Writing Speed-of-Light Flash Attention for 5090 in CUDA C++. Parent comment

Re:Writing Speed-of-Light Flash Attention for 5090 in CUDA C++ by saagarjha on Saturday August 23, 2025 @10:32PM (#45000807)
No, GPUs are a lot simpler. You can mostly just take the clock rate and scale it directly for the instruction being compared.
[ Parent | on HN ]

There are no comments on this comment yet.