# Prerequisites

<!-- https://learn-kernels.com/chapters/prerequisites -->

The rest of this book optimizes one computation on one kind of machine, in numbers narrow enough to be worth arguing about. This chapter covers all three, so that nothing later has to stop and explain them. If you already write kernels, skip it.

## 5 sections

1.  [1.1Why a GPU looks like that](https://learn-kernels.com/chapters/prerequisites/why-a-gpu-looks-like-that)
2.  [1.2The languages used here](https://learn-kernels.com/chapters/prerequisites/the-languages-here)
3.  [1.3How numbers are stored](https://learn-kernels.com/chapters/prerequisites/how-numbers-are-stored)
4.  [1.4What a transformer computes](https://learn-kernels.com/chapters/prerequisites/what-a-transformer-computes)
5.  [1.5Generating one token at a time](https://learn-kernels.com/chapters/prerequisites/generating-one-token)