Skip to content
Sign in

Course

GPUs & Hardware

Why AI runs on parallel accelerators, and what limits them.

L3 · AdvancedEvolvingKnown~2 h

What you’ll learn

  • Explain why GPUs and TPUs suit deep learning and how they differ from CPUs
  • Reason about memory capacity and bandwidth as the dominant performance limit
  • Estimate whether a model fits in VRAM and how it is split across devices

Module 1. Accelerators

Why AI runs on GPUs and TPUs, and the memory wall.

Module 2. Fitting & Scaling

Estimating VRAM and splitting a model across devices.