[Python-Dev] tiny optimization in ceval mainloop
Fri, 30 Aug 2002 11:24:20 -0400
> The difference I saw with only the ticker check in ceval was only
> about 1% for pystone. Python was always faster with the change, but
> never by much.
> A bunch of 0.5% improvements add up. If there's not much cost in
> complexity, why not go for it?
There isn't, and we should <wink>. I'd do it even if it slowed things by
1%: reducing the test+branch count on the critical path will *eventually*
pay off. The SET_LINENO removal worked in the other direction, and that
proved a timing mini-disaster under MSVC6. Doing even random things in the
right direction may very well nudge MSVC6 back into the local minimum it got
knocked out of.