Does anyone have profiling numbers that Azure on top of Skia has a measurable performance overhead?
Thanks, Andreas > On Nov 11, 2014, at 2:53 PM, Cameron Zwarich <zwar...@mozilla.com> wrote: > > On Nov 11, 2014, at 1:19 PM, Patrick Walton <pcwal...@mozilla.com> wrote: >> >> On 11/11/14 1:02 PM, Robert O'Callahan wrote: >>> https://github.com/servo/servo/wiki/Workweek-rasterization >>> >>> We talked about writing our own rasterizer, which makes sense because web >>>> pages only render solid colored rectangles, image, and display text. >>>> >>> >>> No. SVG, border-radius (rounded borders plus clipping to rounded shapes), >>> complex border styles, box-shadows and text-shadows are all commonly used, >>> and their performance matters. Plus you want to be able to rasterize these >>> precisely under arbitrary 2D affine transforms. >> >> I think that this was probably mistranscribed and/or missing context. I >> wasn't in this meeting, but I would assume that this was talking about Servo >> at an earlier stage of development, when people were thinking about writing >> a rasterizer just to see what the performance would be. We have a box-shadow >> PR and a work-in-progress border-radius implementation, after all. > > Yes, it seems that there was something missing in the transcription. I was > making the point that writing a rasterizer is so tempting because the > features used by web content that Servo currently supports wouldn’t be too > difficult to implement. Other CSS features, Canvas, SVG, etc. mean that you > quickly need to support the full gamut. > >> Incidentally, I see no reason to move away from Azure, given how nice >> Direct2D is on Windows. Nor do I think that it realistically makes sense to >> write our own rasterizer, given that we have no reason to believe that we'll >> do better than Direct2D/Skia-GL and the security track record of Azure/Moz2D >> is, based on my analysis, much better than that of (for example) layout or >> the DOM. As I said earlier, I wasn't in this meeting and I do not think that >> we should make any architectural decisions as a result of it. > > The point about moving away from Azure is that we pay a cost for using Skia > via Azure, but don’t get anything for it; none of the alternative backends > actually work. We have to import a lot of code outside of Azure proper in > order to make Azure work, and we had to add Skia-specific surface handling > support. As Martin commented, this adds a decent amount of administrative > overhead. As far as I can tell, the only reason to keep doing this would be > the hope that one of the non-Skia backends is useful to us in the future. > > Cameron > _______________________________________________ > dev-servo mailing list > dev-servo@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-servo _______________________________________________ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
