ListOfPerformanceRelatedPatches

Legacy Wiki Page

This page was migrated from the old MoinMoin-based wiki. Information may be outdated or no longer applicable. For current documentation, see python.org.

Pending:

• [ 1479611 ] speed up function calls TP

• [ 876193 ] reorganize, extend function call optimizations TP

• [ 1107887 ] Speed up function calls/can add more introspection info TP

• [ 1087418 ] long int bitwise ops speedup TP

• [ 1366311 ] SRE engine do not release the GIL

• [ 1346214 ] Better dead code elimination for the AST compiler GB

  • New patch posted, ready for review

• [ 1243654 ] Faster output if message already has a boundary SH

Addressed:

• [ 1145039 ] Remove some invariant conditions and assert in ceval TP

• [ 1442927 ] PyLong_FromString{.backtick} optimization

• [ 1335972 ] combo speedup and bugfix for string->int conversion

• [ 876206 ] scary frame speed hacks [SUCCESS]

• [ 1281707 ] Speed up gzip.readline (~40%) [SUCCESS]

• [ 1337051 ] remove 4 ints from PyFrameObject

• [ 1243730 ] Big speedup in email message parsing SH no speedup observed

• [ 921466 ] Reduce number of open calls on startup GB

• [ 1359618 ] Speed charmap encoder [patch approved, MAL to apply]

Deferred:

• [ 813436 ] Scalable zipfile extension

  • Addressed by SoC student

• [ 738094 ] for i in range(N) optimization

  • Related to [ 1472639 ] make range be xrange

  • Topic was discussed endlessly for some times.

• [ 936813 ] fast modular exponentiation

  • Going to need a lot of attention, not suitable for the sprint.

• [ 923643 ] long <-> byte-string conversion

  • This is a feature request.

• [ 1353872 ] a faster Modulefinder (Related to py2exe runtime performance)

  • No one interested in it.

• [ 1492828 ] Improvements to ceval.c –> RaymondHettinger

• [ 1346238 ] A constant folding optimization pass for the AST