895409 - Pref on SkiaGL canvas

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Toolbar, defect)

Description

[James Willcox (:snorp) ([e-mail]) (he/him)]

Boom.

Comment 1

[James Willcox (:snorp) ([e-mail]) (he/him)]

Attachment: patch

Comment 2

[James Willcox (:snorp) ([e-mail]) (he/him)]

This of course depends on a whole bunch of bugs, but mainly 858237.

Comment 3

[James Willcox (:snorp) ([e-mail]) (he/him)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/4303c1af09f6

Comment 4

[Ed Morley [:emorley]]

Backed out in the hope that this will fix the Android 2.2 reftest failures of form:
https://tbpl.mozilla.org/php/getParsedLog.php?id=25471979&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=25472104&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=25481187&tree=Mozilla-Inbound

...and that I can avoid having to back out the rest of the push.

Out of curiosity, did this patch set get sent to try? I've looked in several of the mentioned bugs, but couldn't find a link.

Backout:
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/6b3289859dd1

Comment 5

[James Willcox (:snorp) ([e-mail]) (he/him)]

Ugh. Thanks for not backing out all of them. The pref one should do the trick.

We've been developing on the graphics branch, and had a couple green runs there.

Comment 7

[James Willcox (:snorp) ([e-mail]) (he/him)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/bdd1a266176f

Comment 8

[James Willcox (:snorp) ([e-mail]) (he/him)]

You can see the bazillion R2 runs here: https://tbpl.mozilla.org/?tree=Graphics&rev=a88543734633

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/bdd1a266176f

Comment 10

[Alex Keybl [:akeybl]]

Hey Erin, should we be communicating this as a perf win, or something different?

Comment 11

[Erin Lancaster [:elan]]

Hi Alex, sorry to just be getting back to you. I don't have quantitative data but I can say:
*This is the first iteration of SkiaGL support for devices with NVIDIA GPUs (only)
*We expect canvas perf improvements in general, one immediate example could be increased FPS with sites such as:
**http://www.craftymind.com/factory/guimark2/HTML5GamingTest.html
**http://ie.microsoft.com/testdrive/Performance/FishBowl/Default.html
*Snorp working to turn SkiaGL on for more devices and therefore, our scope of testing will increase
*Not to pass the buck, but I am need-infoing blassey and snorp to weigh in on how much we should message today. This is certainly interesting from a development point of view and will be more so for users as we increase support/test/stabilize

Comment 12

[Alex Keybl [:akeybl]]

Don't worry too much about the timeline all, we can change the notes throughout the Aurora cycle.

Comment 13

[Brad Lassey [:blassey] (use needinfo?)]

I think relnote-wise we should talk about the perf win for nvidia devices. We should probably wait until 26 to any PR noise about it since it is a bit nuanced (perf win on certain canvas ops on certain devices)