CPython's JIT compiler in Python 3.15 alpha has achieved its performance goals ahead of schedule, running 11-12% faster than the tail calling interpreter on macOS AArch64 and 5-6% faster on x86_64 Linux. The post covers the human side of this turnaround: how the project survived the loss of its main sponsor (Microsoft's Faster CPython team), transitioned to community-led development, and made two key technical bets — a tracing-based JIT frontend with a 'dual dispatch' mechanism and reference count elimination — that together boosted JIT code coverage by 50% and delivered meaningful speedups. The author credits luck, a small dedicated team, and deliberate efforts to lower the bus factor and onboard new contributors through well-scoped tasks.
Table of contents
Part 1: A community-led JITPart 2: Lucky betsPart 3: A great teamPart 4: Talking to peopleSort: