PyPy in production
summary
PyPy claims large speedups over CPython — benchmarks show many cases 10x–100x and microbenchmarks here run about 5x faster than CPython and within roughly 2–5x of optimized C for some loops. It uses a tracing JIT that records hot loops, optimizes a trace and emits machine code for x86/ARM. Tooling includes PyPy log + JIT viewer to inspect traces and IR. Known pitfall: some Python programs break on PyPy because they rely on reference counting semantics rather than PyPy’s GC.
AI-written summary · about
chapters (7)
- 0:00Welcome
- 1:11What PyPy is
- 3:20Running Python on PyPy
- 6:15PyPy JIT compiler
- 9:49Live demo: PyPy speed
- 13:09Inspecting JIT-compiled code
- 20:05Preparation
related talks
57mPython and PyPy performance (not) for dummies
22mPyPy and the future of the Python ecosystem
42mThe state of PyPy
39mCPython, Grumpy, PyPy - When, How, Why
33mWhy is Python slow?
58mPython performance Past, Present, Future
01h 03mA tour of CPython's runtime
48mPython Performance: Past, Present and Future