Catalog · Chips & Density

AI Chips and Compute Density Foundations

Name: AI Chips and Compute Density Foundations
Author: Mach Eigen

101Mach Eigen27 min readNarratedFree during beta

The constraint is simple: buildings take years to deliver, and the chips inside them now change every year. This course explains why artificial intelligence moved onto accelerators, what sits inside one, how the rack became the computer, and how to read a chip roadmap without being carried along by it. Memory, links, power per rack and the building that has to follow. No background assumed. Glad you're here. — Myke Founder, Faraday Intelligence

Start reading

What’s inside

Module 1
Why Artificial Intelligence Runs on Accelerators
- Why the Work Moved to Accelerators · 3 min · narrated
- What the CPU Still Does, and What It Stopped Doing · 3 min · narrated
- Module review
Module 2
Inside the Accelerator
- Inside the Accelerator: The Arithmetic Engine · 3 min · narrated
- The Memory Wall and the Links Between Chips · 3 min · narrated
- Module review
Module 3
From Chip to Rack to Building
- From Chip to Rack: When the Rack Becomes the Computer · 4 min · narrated
- Power per Rack, and What It Does to a Building · 3 min · narrated
- Module review
Module 4
Reading the Roadmap
- The Annual Cadence: How Chip Roadmaps Are Built · 3 min · narrated
- Competing Silicon, and How a Non-Engineer Should Read a Chip Roadmap · 3 min · narrated
- Module review

Take the quiz Glossary

AI Chips and Compute Density Foundations

What’s inside

Why Artificial Intelligence Runs on Accelerators

Inside the Accelerator

From Chip to Rack to Building

Reading the Roadmap