Bug 77952 – REGRESSION(r106819): ~28% or so performance regression on the ManInBlue HTML benchmark

• Status: RESOLVED
• Resolution: FIXED
• Component: DOM

[Ryosuke Niwa]

From https://bugs.webkit.org/show_bug.cgi?id=77879:

Comment #4 From Tim Horton 2012-02-06 13:32:31 PST (-) [reply] 
This change seems to significantly (~28% or so) regress performance of the ManInBlue HTML (http://themaninblue.com/experiment/AnimationBenchmark/html/) benchmark, among others. Looking at the change, though, I'm not sure how that's possible. Could you potentially take a look at this?

[Ryosuke Niwa]

Created attachment 125791 [details]
Fix attempt

This is a fix attempt per analysis on the bug 77879. As much as I'd like to finish writing this patch, my brain isn't functioning anymore at 2am in the morning. I'll try to finish up the patch tomorrow if cmarcelo hadn't done so by the time I can get back to it.

[Ryosuke Niwa]

Okay, not adding a new test isn't as bad as I thought. DOM/Accessors caught this regression:

It also caused a regression on DOM/Accessors:
http://webkit-perf.appspot.com/graph.html?#tests=[[3113,2001,32196]]&sel=1328551109987.2344,1328555918820.6528&displayrange=7&datatype=running

It seems like we imported Dromaeo tests just in time :D

[Andreas Kling]

(In reply to comment #1)
> Created an attachment (id=125791) [details]
> Fix attempt
> 
> This is a fix attempt per analysis on the bug 77879. As much as I'd like to finish writing this patch, my brain isn't functioning anymore at 2am in the morning. I'll try to finish up the patch tomorrow if cmarcelo hadn't done so by the time I can get back to it.

This patch looks fine to me, so I'm taking the liberty of saying r+ and landing it for you to get rid of the regression.

[Andreas Kling]

Committed r106925: <http://trac.webkit.org/changeset/106925>

[Ryosuke Niwa]

(In reply to comment #3)
> (In reply to comment #1)
> > Created an attachment (id=125791) [details] [details]
> > Fix attempt
> > 
> > This is a fix attempt per analysis on the bug 77879. As much as I'd like to finish writing this patch, my brain isn't functioning anymore at 2am in the morning. I'll try to finish up the patch tomorrow if cmarcelo hadn't done so by the time I can get back to it.
> 
> This patch looks fine to me, so I'm taking the liberty of saying r+ and landing it for you to get rid of the regression.

Thanks!