Tavis,
If you want to take the next step, then you should test it to see if it
makes a noticeable difference in the performance of WebKit. If so, then we
can consider including it in WebKit. I'd leave it up to Chuck to decide if
the improved performance is worth the extra hassle and confusion of having
our own souped-up version of a standard module.
These kinds of optimizations don't interest me personally very much, but it
does seem that a lot of people are interested in squeezing as much speed
out of WebKit as possible.
At 04:26 PM 6/4/01 -0700, Tavis Rudd wrote:
>Hi,
>I've gone through the Queue class and done a number
>of optimizations that resulted in a very significant speed
>increase. The modified verion and a test script are
>attached. I've only done a very simple test so far.
>
>prior to optimizations using the existing Queue class
>======================================
>tavis@lucy: ~ > time python test.py
>
>real 0m6.367s
>user 0m6.260s
>sys 0m0.020s
>
>after to optimizations
>======================================
>tavis@lucy: ~ > time python test.py
>
>real 0m4.361s
>user 0m4.320s
>sys 0m0.020s
>
>Difference
>======================================
>It shaves off 2.006 seconds (31%) from the original time
>
>This could be worthwhile considering how central
>queues are in WebKit.
>Tavis
>
>p.s. sourceforge's mail server is taking hours
>to deliver things today. anyone know why?
>That's why I cc'd all of you...
>
>
--
- Geoff Talvola
[EMAIL PROTECTED]
_______________________________________________
Webware-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/webware-devel
