[issue37774] Micro-optimize vectorcall using PY_LIKELY

[STINNER Victor]

STINNER Victor <vstinner at redhat.com> added the comment:

Few more links about likely/__builtin_expect:

* GCC documentation says: "In general, you should prefer to use actual profile feedback for this (-fprofile-arcs), as programmers are notoriously bad at predicting how their programs actually perform."

  https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html

* https://softwareengineering.stackexchange.com/questions/188853/how-much-usage-of-likely-and-unlikely-macros-is-too-much

  "If you're writing for x86/x64 (and are not using 20-year-old CPUs), the performance gain from using __builtin_expect() will be negligible if any."

* http://blog.man7.org/2012/10/how-much-do-builtinexpect-likely-and.html

  "This optimized version [PGO] runs significantly faster (...) than our version that used __builtin_expect()."

  [When __builtin_expect() is misused] "In this case, unsurprisingly, we made each check run slower (...)"

* https://stackoverflow.com/questions/109710/how-do-the-likely-unlikely-macros-in-the-linux-kernel-work-and-what-is-their-ben

  "Like all such performance optimisations you should only do it after extensive profiling to ensure the code really is in a bottleneck (...)"

* https://stackoverflow.com/questions/10922607/gcc-likely-unlikely-macro-usage

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue37774>
_______________________________________