Get the latest tech news

B200 Attention Kernel from Scratch to Near-SOTA in 60 Diagrams

B200 attention from scratch in 60 diagrams: 14 CUDA/PTX kernels from naive to near-SOTA, plus a video-generation capstone.

None

Get the Android app

Or read this on Hacker News

Related news:

I were 17, I'd learn how to build LLMs from scratch

'There's no realistic chance all of that content will get remastered from scratch': Dolby Vision 2 will come to more titles than you might expect, thanks to an automated process — but there will be two key creative-led features that aren't automated

AI-Powered Browser Just Generates Every Website From Scratch

« Nvidia's controversial, face-changing, gen-AI DLSS 5 tech releases tomorrow for one game, and seems to tank frame-rates

ICE Is Paying a Controversial AI Firm to Hide the Identities of Agents | In a leaked memo, an ICE official tells employees the new AI tech will protect them from doxing. Some worry it could be used to root out whistleblowers »