1553575 - MotionMark 1.1 - test "Multiply": scores with Firefox are inconsistent

Status: RESOLVED FIXED

(Core :: Graphics: WebRender, defect, P3)

Description

[Rival]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:67.0) Gecko/20100101 Firefox/67.0

Steps to reproduce:

With the landing of Webrender on Firefox stable, in my opinion, the most used benchmark to highlight the performance improvements obtained thanks to Webrender, will be MotionMark.
Of course it is desirable that Firefox works in the best way with this benchmark.

With my configuration, with tests run with Webrender enabled and not, I noticed some problems.

MotionMark ver1.1 and ver1.0
Firefox 67.0 64bit
Windows 10 build 1803 (display: 2560x1440 60hz)
GPU GTX1600 - 6GB (driver: 430.64)
CPU i5 6600K - 16GB

• Test "Multiply"
  https://browserbench.org/MotionMark1.1/developer.html?test-interval=30&display=minimal&tiles=big&controller=ramp&frame-rate=50&kalman-process-error=1&kalman-measurement-error=4&time-measurement=performance&suite-name=MotionMark&test-name=Multiply

Actual results:

Scores are extremely variable and ridiculously scarce; in addition, with an activated webrender, they fall further.
Perhaps this test in version 1.1, has heavy fps measurement problems with Firefox, and Webrender makes it worse.

With other browsers (for example, the "old" Microsoft Edge) instead the scores are fairly constant.
With the version 1.0 of MotionMark, these inconsistencies in the scores in the Multiply test are not there, in fact, even with Firefox the scores are good and are not so fluctuating, moreover, by activating webrender, it gets (as expected) higher scores.

Comment 1

[Jeff Muizelaar [:jrmuizel]]

Running MotionMark 1.2 on macOS I see something similar.

Comment 2

[Jeff Muizelaar [:jrmuizel]]

Attachment: Chrome frame graph

Comment 3

[Jeff Muizelaar [:jrmuizel]]

Attachment: FF frame graph

The two graphs make the reason for the very different scores pretty obvious. Chrome's frame rate converges and ours is very bimodal

Comment 4

[Jeff Muizelaar [:jrmuizel]]

This should be fixed now.

Comment 5

[Jeff Muizelaar [:jrmuizel]]

This was fixed by bug 1778575