Beginner · Systems

GPU

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. Graphics Processing Unit — a massively parallel processor that powers most modern AI workloads.

Technical Definition

Graphics Processing Unit — a massively parallel processor that powers most modern AI workloads.

How it works

GPUs perform thousands of simple operations in parallel, ideal for the matrix math at the core of neural networks. NVIDIA dominates the market with CUDA-based H100, A100, and B200 chips, though AMD and custom silicon (TPUs, Trainium, MI300) are competing.

Related Concepts

• TPU — Tensor Processing Unit — Google's custom-designed AI accelerator chip.
• CPU — Central Processing Unit — the general-purpose chip that runs most software.
• Accelerator — Any specialized hardware designed to run AI workloads faster or more efficiently than a general CPU.