[A.M. Kuchling]
[CC'ing to python-dev] Confirmed:
[amk@mira Python-2.0]$ ./python Lib/test/pystone.py Pystone(1.1) time for 10000 passes = 3.14 This machine benchmarks at 3184.71 pystones/second [amk@mira Python-2.0]$ python2.1 Lib/test/pystone.py Pystone(1.1) time for 10000 passes = 3.81 This machine benchmarks at 2624.67 pystones/second
The ceval.c changes seem a likely candidate to have caused this. Anyone want to run Marc-Andre's microbenchmarks and see how the numbers have changed?
Want to, yes, but it looks hopeless on my box: **** 2.0 C:\Python20>python lib/test/pystone.py Pystone(1.1) time for 10000 passes = 0.851013 This machine benchmarks at 11750.7 pystones/second C:\Python20>python lib/test/pystone.py Pystone(1.1) time for 10000 passes = 1.24279 This machine benchmarks at 8046.41 pystones/second **** 2.1a1 C:\Python21a1>python lib/test/pystone.py Pystone(1.1) time for 10000 passes = 0.823313 This machine benchmarks at 12146 pystones/second C:\Python21a1>python lib/test/pystone.py Pystone(1.1) time for 10000 passes = 1.27046 This machine benchmarks at 7871.15 pystones/second **** CVS C:\Code\python\dist\src\PCbuild>python ..\lib\test\pystone.py Pystone(1.1) time for 10000 passes = 0.836391 This machine benchmarks at 11956.1 pystones/second C:\Code\python\dist\src\PCbuild>python ..\lib\test\pystone.py Pystone(1.1) time for 10000 passes = 1.3055 This machine benchmarks at 7659.9 pystones/second That's after a reboot: no matter which Python I use, it gets about 12000 on the first run with a given python.exe, and about 8000 on the second. Not shown is that it *stays* at about 8000 until the next reboot. So there's a Windows (W98SE) Mystery, but also no evidence that timings have changed worth spit under the MS compiler. The eval loop is very touchy, and I suspect you won't track this down on your box until staring at the code gcc (I presume you're using gcc) generates. May be sensitive to which release of gcc you're using too. switch-to-windows-and-you'll-have-easier-things-to-worry-about<wink>-ly y'rs - tim