# PyPy

> Three times the pure-Python CPU of Gunicorn, after a JIT warmup. Poor C-extension support means no ML ecosystem bonus. 200 MB per worker.

## Trade-offs

| Axis | Score |
| --- | --- |
| Throughput | |
| Tail latency | |
| Memory | |
| Velocity | |
| Resilience | |
| Operability | |

## Frameworks

- [Django](https://wiki.playquorum.dev/frameworks/django.md)
- [Flask](https://wiki.playquorum.dev/frameworks/flask.md)
- [FastAPI](https://wiki.playquorum.dev/frameworks/fastapi.md)

## Engineering terms

- [Process per request](https://wiki.playquorum.dev/glossary/process-per-request.md)
- [JIT warmup](https://wiki.playquorum.dev/glossary/jit-warmup.md)

## Facts

PyPy
- **Language:** Python
- **Concurrency:** One process per request
- **Speed:** ×0.6
- **Memory:** 200 MB
- **Startup:** 5 s
- **Warm-up:** 15 s
- **GC pause:** None
- **Unlocked by:** [PyPy](https://wiki.playquorum.dev/research/pypy.md)
- **Velocity:** ×1.15
- **Salary:** ×1.05
- **Build:** 1 min

## See also

- [Runtimes and frameworks](https://wiki.playquorum.dev/runtimes.md)
- [Django](https://wiki.playquorum.dev/frameworks/django.md)
- [Flask](https://wiki.playquorum.dev/frameworks/flask.md)
- [FastAPI](https://wiki.playquorum.dev/frameworks/fastapi.md)

Source: https://wiki.playquorum.dev/runtimes/pypy