[Stackless] Stackless and Psyco
bsdbeast at yahoo.com
Mon Mar 1 21:19:07 CET 2004
Indeed, thanks for that great graph library btw, I use
it extensively for other things than GadflyB5 as well.
The problem lies in the between the request and the
respons, I guess I could solve it with some funky
caching techniques as there are about a 200:1 factor
between selects and inserts.
Any ideas if that is a good option?
Oh, and the reason I wanted a stackless psyc(h)o was
the budget and time frame, which are, well, not the
best at the moment.
--- Aaron Watters <aaron at reportlab.com> wrote:
> Just curious: are you using kjbuckets with gadfly?
> That could make around a 10x speed difference for
> database intensive code (stackless or not) depending
> on the kinds of queries you do.
> -- Aaron Watters
> Mellon Gibson <bsdbeast at yahoo.com> wrote:
> I can't get psyco to work with stackless, which is
> pretty critical for my application, as it's a tad
> bit too slow right now (about to 11% speed
> improvement would suffice to serve all clients
> instead of buying new hardware). Both the client and
> server classes are designed around GadflyB5, which
> needs some of the 30% speed improvements I had with
> just vanilla python and psyco, but then the server
> tasks become too slow because of the C stack
> fumbling that vanilla python does. Are there any
> plans to support psyco, is it too much work? I'm not
> very fond of the idea to buy brand new SMP hardware
> to support just 300 more clients. /beast
> Do you Yahoo!?
> Get better spam protection with Yahoo!
> Stackless mailing list
> Stackless at stackless.com
Do you Yahoo!?
Get better spam protection with Yahoo! Mail.
Stackless mailing list
Stackless at stackless.com
More information about the Stackless