[dev-servo] Meeting notes 9/28 (Meeting time change; additional TWiS reporting; script crate rebuilds; webrender)
https://github.com/servo/servo/wiki/Meeting-2015-09-28 ___ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
[dev-servo] WebRender
This sounds very promising! A question: one of the graphs shows Servo taking 35ms per frame to paint the slide-in animation. What exactly is in this animation that makes it take such a relatively long time to paint? Rob -- lbir ye,ea yer.tnietoehr rdn rdsme,anea lurpr edna e hnysnenh hhe uresyf toD selthor stor edna siewaoeodm or v sstvr esBa kbvted,t rdsme,aoreseoouoto o l euetiuruewFa kbn e hnystoivateweh uresyf tulsa rehr rdm or rnea lurpr .a war hsrer holsa rodvted,t nenh hneireseoouot.tniesiewaoeivatewt sstvr esn ___ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
Re: [dev-servo] WebRender
It's a fairly simple off canvas menu - it is full height (1024px in this test), and expands from off-screen to 600px wide. It has a linear gradient for the background, and a small amount of text on it. I was also surprised the existing Servo paint code is so slow with this - it's on my todo list to profile the existing renderer and see if there's something silly going on. I'm planning to tidy up the test cases I'm using and push them to git, so that we have a set of references for benchmarking purposes. On 29/09/15 11:52, Robert O'Callahan wrote: This sounds very promising! A question: one of the graphs shows Servo taking 35ms per frame to paint the slide-in animation. What exactly is in this animation that makes it take such a relatively long time to paint? Rob ___ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
Re: [dev-servo] WebRender
On Tue, Sep 29, 2015 at 3:55 PM, Glenn Watson wrote: > It's a fairly simple off canvas menu - it is full height (1024px in this > test), and expands from off-screen to 600px wide. It has a linear gradient > for the background, and a small amount of text on it. > > I was also surprised the existing Servo paint code is so slow with this - > it's on my todo list to profile the existing renderer and see if there's > something silly going on. > That sounds like the kind of thing most browsers would pre-render and animate in the compositor. Rob -- lbir ye,ea yer.tnietoehr rdn rdsme,anea lurpr edna e hnysnenh hhe uresyf toD selthor stor edna siewaoeodm or v sstvr esBa kbvted,t rdsme,aoreseoouoto o l euetiuruewFa kbn e hnystoivateweh uresyf tulsa rehr rdm or rnea lurpr .a war hsrer holsa rodvted,t nenh hneireseoouot.tniesiewaoeivatewt sstvr esn ___ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
Re: [dev-servo] WebRender
Well, the idea here is to test the case in which the object isn't layerized for some reason--say, the object was moving by animating "margin-left" in JavaScript and it had complex Z-ordering constraints that caused whatever layerization heuristics the browser had to fail, causing a full repaint of the affected area each frame. This is, of course, a pessimizing assumption :) But we do have reason to believe that this happens not infrequently in practice due to the multitude of ways to animate on the Web (see Paul Irish's writeups on this, for example). Much of the goal here is to make the baseline performance fast, so that even when our layerization code fails to produce the optimal result the user experience remains nice. Patrick On Sep 28, 2015 8:15 PM, "Robert O'Callahan" wrote: > On Tue, Sep 29, 2015 at 3:55 PM, Glenn Watson > wrote: > > > It's a fairly simple off canvas menu - it is full height (1024px in this > > test), and expands from off-screen to 600px wide. It has a linear > gradient > > for the background, and a small amount of text on it. > > > > I was also surprised the existing Servo paint code is so slow with this - > > it's on my todo list to profile the existing renderer and see if there's > > something silly going on. > > > > That sounds like the kind of thing most browsers would pre-render and > animate in the compositor. > > Rob > -- > lbir ye,ea yer.tnietoehr rdn rdsme,anea lurpr edna e hnysnenh hhe uresyf > toD > selthor stor edna siewaoeodm or v sstvr esBa kbvted,t > rdsme,aoreseoouoto > o l euetiuruewFa kbn e hnystoivateweh uresyf tulsa rehr rdm or rnea > lurpr > .a war hsrer holsa rodvted,t nenh hneireseoouot.tniesiewaoeivatewt sstvr > esn > ___ > 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
Re: [dev-servo] WebRender
On Tue, Sep 29, 2015 at 4:36 PM, Patrick Walton wrote: > Well, the idea here is to test the case in which the object isn't > layerized for some reason--say, the object was moving by animating > "margin-left" in JavaScript and it had complex Z-ordering constraints that > caused whatever layerization heuristics the browser had to fail, causing a > full repaint of the affected area each frame. This is, of course, a > pessimizing assumption :) But we do have reason to believe that this > happens not infrequently in practice due to the multitude of ways to > animate on the Web (see Paul Irish's writeups on this, for example). Much > of the goal here is to make the baseline performance fast, so that even > when our layerization code fails to produce the optimal result the user > experience remains nice. > Sure, that makes sense. Rob -- lbir ye,ea yer.tnietoehr rdn rdsme,anea lurpr edna e hnysnenh hhe uresyf toD selthor stor edna siewaoeodm or v sstvr esBa kbvted,t rdsme,aoreseoouoto o l euetiuruewFa kbn e hnystoivateweh uresyf tulsa rehr rdm or rnea lurpr .a war hsrer holsa rodvted,t nenh hneireseoouot.tniesiewaoeivatewt sstvr esn ___ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
