REGRESSION(r157354-r157355): 2.6% regression on Dromaeo/cssquery-jquery

• Component: CSS
• Status: RESOLVED INVALID

[Ryosuke Niwa]

https://perf.webkit.org/#mode=charts&chartList=%5B%5B%22mac-mountainlion%22%2C%22Dromaeo%2Fcssquery-dojo%3ARuns%22%5D%2C%5B%22mac-mountainlion%22%2C%22Dromaeo%2Fcssquery-jquery%3ARuns%22%5D%2C%5B%22mac-lion%22%2C%22Dromaeo%2Fcssquery-jquery%3ARuns%22%5D%5D&days=136&zoom=%5B1381314242746.7024%2C1382083568577.5894%5D

http://trac.webkit.org/log/?rev=157355&stop_rev=157354&verbose=on

[Darin Adler]

r157355 was supposed to just be a refactoring and should not have had any performance impact.

[Darin Adler]

I re-read http://trac.webkit.org/changeset/157355 and I can’t spot any change that should have had performance impact.

[Andreas Kling]

I think <http://trac.webkit.org/changeset/157354> is the likely culprit here.

[Darin Adler]

I guess I misunderstood the title of the bug. I assumed it meant that r157354 was known to fast and r157355 was known to be slow.

[Darin Adler]

r157354 was supposed to just be a refactoring and should not have had any performance impact.

I re-read http://trac.webkit.org/changeset/157354 and I can’t spot any change that should have had performance impact.

[Ryosuke Niwa]

This regression is hardly noticeable on my local machine so it could be something to do with memory alignment and the order file.

[Ryosuke Niwa]

CSS Query 

r157352: http://dromaeo.com/?id=212262

8429.08runs/s (Total)
DOM Query (Dojo): 7327.92runs/s
DOM Query (ExtJS): 5174.90runs/s
DOM Query (Mootools): 11909.97runs/s
DOM Query (Prototype): 9965.06runs/s
DOM Query (Yahoo UI): 5969.95runs/s
DOM Query (jQuery): 13193.79runs/s

r157355: http://dromaeo.com/?id=212266

8484.89runs/s (Total)
DOM Query (Dojo): 7488.50runs/s
DOM Query (ExtJS): 5400.73runs/s
DOM Query (Mootools): 11823.50runs/s
DOM Query (Prototype): 9950.53runs/s
DOM Query (Yahoo UI): 5940.33runs/s
DOM Query (jQuery): 13064.93runs/s

I guess the change is with in the margin of errors here :(