Stackless 1.01 Homepage

[Christian Tismer]

Robin Becker wrote:
> 
> In article <388BE68E.B26985CA at tismer.com>, Christian Tismer <tismer at tismer.com> writes
> 
> I'm very impressed. here are the pybench timing differences (absolute and expressed as % of
> original python) for my 333Mhz win32 machine. here negative is bad for stackless, positive is
> good. My only disappointment is that Zope wouldn't start up.

Whow. I guess there is a bug in exec. In any case, it cannot be a
conceptual but but just a failure of mine. Idle, PyWin, pystone,
Metakit, they all work fine.

Will look into it after the conference.
Thanks for giving the measures.

These optimizations were done last January partially, before
I started to think of SLP. They didn't make it into the core,
like SLP didn't as well. As a consequence I bundled them and pushed
optimization even further, to outperform Standard Python.
If these patches would be applied to the original, I expect about
15 % speed gain, because SLP actually looses about 8%.

But please don't apply them :-))

cheers - chris

-- 
Christian Tismer             :^)   <mailto:tismer at appliedbiometrics.com>
Applied Biometrics GmbH      :     Have a break! Take a ride on Python's
Düppelstr. 31                :    *Starship* http://starship.python.net
12163 Berlin                 :     PGP key -> http://wwwkeys.pgp.net
PGP Fingerprint       E182 71C7 1A9D 66E9 9D15  D3CC D4D7 93E2 1FAE F6DF
     we're tired of banana software - shipped green, ripens at home