Skip to content

The Chip Letter personal blog

Demystifying GPU compute architectures

by Babbage thechipletter.substack.com published 2024-02-04

Getting 'low level' with Nvidia and AMD GPUs

Read on thechipletter.substack.com → Discuss on HN (59) Archived copy

Points: 210
HN comments: 59
Submissions: 1
Cited by: 0 commenters

Hardware & Electronics AI & Machine Learning 2024

On Hacker News

Submitted	Title on HN	Points	Comments
Feb 6, 2024	Demystifying GPU compute architectures by rbanffy	210	59

More from The Chip Letter

All 18 posts →

Google's First Tensor Processing Unit: Architecture

2024354 points187 comments Hardware & Electronics details

Ericsson to WhatsApp: The Story of Erlang

2023270 points79 comments Programming Languages details

Satya Nadella uses an IBM AS/400 in 1993

2024254 points147 comments History & Retro Computing details

Intel's Immiseration

2024185 points218 comments Tech Industry & Business details

A History of C Compilers – Part 1: Performance, Portability and Freedom

2024180 points109 comments Programming Languages details

The Nvidia Way

2024164 points71 comments Tech Industry & Business details

Related posts from the same era

Apple Vision Pro: Apple’s first spatial computer

Apple Newsroom20232,608 points2,843 commentscited by 4 details

I designed a cube that balances itself on a corner

Willem Pennings20242,460 points377 comments details

MomBoard: E-ink display for a parent with amnesia

Jan Miksovsky20242,106 points251 comments details

We should have the ability to run any code we want on hardware we own

Hugo Tunius20252,071 points1,202 comments details

Air Con: $1697 for an on/off switch

Hopefully Useful20242,001 points478 comments details

What's that touchscreen in my room?

Nikita Lapkov20241,890 points560 comments details