Question about JIT performance improvement

Python Help
1

As a long-term Python user, I closely follow the updates of Python performance improvement. According to benchmarking-public/RESULTS.md at main · faster-cpython/benchmarking-public · GitHub, the JIT performance has greatly improved recently.

Under the linux x86_64 (linux) section, the benchmarks show a great improvement between 2fd09b0 (JIT) and 9fbd66a (JIT), jumping from 1.310x to 1.913x faster, as shown below,

screenshot
screenshot2071×939 151 KB

Out of my curiosity,

  • which commit(s)/change(s) bring us the great improvement?
  • will this improvement been released with Python 3.14?
1 Like
2

Looks a whole lot like that is from the branch listed right below it from @brandtbucher…

3

This one, which accidentally made it so that certain JIT-compiled for loops (like those in the pprint benchmarks) exit immediately instead of iterating.

No, because this bug has since been fixed.

:slight_smile:

9 Likes