README.md

---
title: The Memory Wall
emoji: 📉
colorFrom: gray
colorTo: indigo
sdk: static
app_file: index.html
pinned: false
license: apache-2.0
short_description: Why your GPU's TFLOPs do not decide your tokens per second
---

# The Memory Wall

An interactive calculator for the one equation that governs single-stream
generation speed:

```
tokens/second  ≤  memory bandwidth  ÷  model bytes
```

No model download, no GPU, runs instantly. Useful on its own, and it is also the
argument for everything else Exeaon publishes.