thron7 wrote:
> You're right, there's something inherently weird about this patch. My 
> own measurements do not match with my initial figures (using a different 
> platform, though). I undid it in trunk (r20591). BTW, your figures look 
> strange too (e.g. build times are *bigger* with cache than without -  if 
> this were true the whole cache would be worthless).

That may be a little too late since it's undone, but here are our results for a 
~400 classes application when using the file TreeLoader.py of r20582 and 
sticking with the 0.8.3 release for the other files.

./generate.py build (no cache) - 0.8.3 release:

real    1m47.422s
user    1m44.002s
sys     0m1.792s

Memory usage: ~50 MB


./generate.py build (no cache) - r20582:

real    2m19.662s
user    2m17.909s
sys     0m1.612s

Memory usage: ~550 MB

====================================================

./generate.py source (no cache) - 0.8.3 release:

real    0m38.235s
user    0m37.466s
sys     0m0.680s

Memory usage: ~46 MB


./generate.py source (no cache) - r20582:

real    0m40.432s
user    0m39.554s
sys     0m0.720s

Memory usage: up to 280 MB


The new version was always slower and used more memory.


> Thanks to all for testing anyway.
Thanks for trying to improve the build speed :)


-- 
Loïc Bresson


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to