These are current tasks at the sprint. Also various people are evaluating patches from the ListOfPerformanceRelatedPatches, and initials record the participant(s) taking those resposibilities.
Continue function call speed enhancements. Neal Norwitz patch review. Zombie frames? TimPeters will continue.
Faster parsing of strings and bytes into int, long, and back. TimPeters: patch ready - needs clean-up
Buffer for use with network I/O MartinBlais working in blair-bytebuf branch --> sandbox? [Performance comparisons first]
Patch list review DONE (see ListOfPerformanceRelatedPatches)
Implement portions of the decimal module in C: GeorgBrandl and JackDeiderich
Create a PyInt64 type (for 32 bit machines) Evaluate source impact ...GvR: Not in 2.5! [/Deferred?]
- Build-out struct module to support fast, high-volume binary conversions [perhaps with regexp analogs such struct.compile() and struct.finditer()] Offset arg? DONE
Optimize methods in stringobject.c FredrikLundh + AndrewDalke
Allow selective removal of unused features such as profiling support DONE [/DeferredSuccesses?]
Expose/develop pre-allocation arguments/methods for dict/list? API? [/DeferredSucceses]
- String benchmark DONE
Improve interpreter startup time, like in patch [SF]921466? GeorgBrandl
Patch that Skip wrote for a possible implementation of str.partition() which returns a string subclass with pointers to the slices instead of a full copy. Any subsequent string ops lazily evaluate the slice into a full string. [--> RaymondHettinger]
- Unicode improvements per CCP requirements: DONE
Build-out the collections module for optimized data structures: ordered dictionary: API? Use case?! red/black? pivot tbl? skip list: contribution? radix/crit-bit trees? ChristianTismer + RichardEmslie
Struct benchmark DONE - needs to be incorporated into pybench BobIppolito
- Improve the API for timeit.py [SH to search python-dev and call Raymond Hettinger for details if necessary]
Update PythonSpeed/PerformanceTips Everyone can make suggestions SeanReifschneider + JohnBenediktsson
- Evaluate the PEPs for optimizing global and attribute lookups
- PEP 266 PEP 267 PEP 280
Psyco improvements: JohnBenediktsson
- Support for generator expressions Support for nested scopes Support for more dictionary operations Speedup float arithmetic Support for more built-ins (e.g. int(), long(), float(), etc.) Upgrade for python 2.5 Better tools for profiling psyco-ness of application Investigate usefulness of IVM (with aim to producing a more streamlined dispatch loop) LLVM backend Virtualized longs (for long longs) Virtualized slots (Ability to cache getattribute() values)
* Py3000: Make a wishlist for possible performance gains in Py3.0