[Webkit-unassigned] [Bug 90019] [chromium] Add time spent painting to GPU benchmarking renderingStats() API.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=90019





--- Comment #13 from Dave Tu <dtu at chromium.org>  2012-07-19 16:50:07 PST ---
(From update of attachment 152358)
View in context: https://bugs.webkit.org/attachment.cgi?id=152358&action=review

>> Source/WebCore/platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.cpp:-98
>> -    SkPictureCanvasLayerTextureUpdater::prepareToUpdate(contentRect, tileSize, contentsWidthScale, contentsHeightScale, resultingOpaqueRect);
> 
> I think this class' prepareRect time spent in updater->paintContentsRect should increment timePainting. While you're at it, add a timeSpentRasterizing property to the stats that is separate that we bump when this class does drawPicture, lower in the file.

Done. When I tested this, it doesn't go through BitmapSkPictureCanvasLayerTextureUpdater::prepareToUpdate at all. Are there chrome://flags that I should enable? And what should timeSpentRasterizing do when it's not enabled? Right now it just says 0.

>>> Source/WebCore/platform/graphics/chromium/cc/CCLayerTreeHost.cpp:218
>>> +    ++m_renderingStats.numAnimationFrames;
>> 
>> why ++x and not x++? Does webkit style guide advise on this?
> 
> The style guide doesn't, but Chromium compositor code pretty consistently uses post-increment except in for loops.

Changed to post-increment.

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