[Webkit-unassigned] [Bug 64262] Small speed up, which switches some virtual functions to inline ones.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=64262





--- Comment #26 from Gabor Loki <loki at webkit.org>  2011-07-29 09:25:53 PST ---
> I, like Dave, don't understand the testing methodology here.  Is this a speed up on page load time?

Methanol measures only the page loads, nothing else.
Lets see how you can reproduce this numbers. ;)

1.- Make a local mirror of a webpage (for example: http://gitorious.org/methanol/methanol/trees/master/sample/openbsd.org )
2.- Add an event listener - which can call back the Methanol's engine - to that local webpage (see here: http://gitorious.org/methanol/methanol/blobs/master/sample/openbsd.org/index.html#line1 )
3.- Register the relative path of that website in the test container (lets say: sample.js).
4.- Open Methanol's URL from the browser which loads the local website several times.
5.- Each page loads are saved separately and the final page (results.html) shows the average of page loads and the standard deviation as well.

So, at the end we have the average of page loads (in millisecond). No gprof, valgrind or such thing is needed.


Tamas's percentages are computed with the following formula:
  ((old value) / (new value) -1).
For example: (100 ms)/(89 ms)-1 ~= 12.35%

I hope that I have been a help.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.