New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 774678 link

Starred by 4 users
Status: Assigned
Owner:
█ marc...@chromium.org
Last visit > 30 days ago
Cc:
pucchakayala@chromium.org
█ marc...@chromium.org
█ bccheng@chromium.org
vsu...@chromium.org
█ bhthompson@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug-Regression



Sign in to add a comment

FishBowl FPS is low on Spring device

Project Member Reported by songsuk@chromium.org, Oct 13 2017 
Chrome Version       : 62.0.3202.55
Platform             : 9901.46.0 Daisy_Spring
URLs (if applicable) : https://testdrive-archive.azurewebsites.net/Performance/FishBowl/Default.html


What steps will reproduce the problem?
(1)  navigate to https://testdrive-archive.azurewebsites.net/Performance/FishBowl/Default.html 

(2) set the fish number to 10, and check the fps meter 

What is the expected result?
fps should be 60.

What happens instead?
Maximum fps is around 15. Fish is slowly swimming. Please see attached m62.webm file. 

Please provide any additional information below. Attach a screenshot if
possible.
Unable to reproduce the issue on 61.0.3163.120/9765.81.0- Spring,(m61.webm).
Unable to reproduce the issue on 62.0.3202.55/9901.46.0 - Reks/Peppy
m62.webm
376 KB View Download
m61.webm
158 KB View Download

Comment 1 by songsuk@chromium.org, Oct 13 2017 

Attached the Spring device gpu information.
gpu.zip
9.4 KB Download

Comment 2 by bhthompson@google.com, Oct 13 2017

Cc: bccheng@chromium.org marc...@chromium.org
Any ideas what we may have done that would impact the old Exynos 5250?

Comment 3 by marc...@chromium.org, Oct 13 2017 

This test uses both canvas (the fishes are) and video playback (the water animation is actually a video). There were probably changes on the Chrome side to both of these, but it's too far behind to act upon it at this point.

In the future, if we care about this benchmark, we should track the results as part of chromeperf so we can act right away.

Comment 4 by songsuk@chromium.org, Oct 13 2017 

Good build : Chrome 62.0.3202.50 / CrOS 9901.42.0
Bad build  : Chrome 62.0.3202.52 / CrOS 9901.44.0

CL : https://chromium.googlesource.com/chromium/src/+log/62.0.3202.50..62.0.3202.52?pretty=fuller&n=10000

Comment 5 by marc...@chromium.org, Oct 16 2017

Owner: songsuk@chromium.org
Status: Assigned (was: Untriaged)
I just tested on daisy (no more working daisy_spring here, sorry) and the perf looks fine on 63. Can you retest on 63? If that's fine we can wontfix this bug.

Comment 6 by songsuk@chromium.org, Oct 17 2017

Owner: marc...@chromium.org
Able to reproduce the issue on Chrome 63.0.3239.7/CrOS 10032.4.0 - daisy_spring

Comment 7 by bhthompson@google.com, Oct 18 2017 

Daisy and daisy_spring should be functionally identical, so it is strange this only repros on one.

Does this happen on multiple daisy_spring units or just one?

Comment 8 by songsuk@chromium.org, Oct 18 2017 

Since browser qa team has 1 Spring device, I checked the issue on 1 device. 
>> Does this happen on multiple daisy_spring units or just one?

Comment 9 by songsuk@chromium.org, Oct 18 2017 

Able to reproduce the issue on 62.0.3202.63/9901.49.0 - Mickey, Daisy

Comment 10 by songsuk@chromium.org, Oct 18 2017

Labels: Arch-ARM
The issue is reproducible on Chrome 63.0.3239.8/CrOS 10032.5.0 - daisy

Comment 11 by bhthompson@google.com, Oct 23 2017

Labels: -Pri-1 Pri-0
Upping priority as we build the R62 stable RC tomorrow, we need a fix in the next 24 hours.

Comment 12 by marc...@chromium.org, Oct 23 2017

Labels: -Pri-0 -ReleaseBlock-Stable Pri-1
This one isn't really a blocker or a P0...

Comment 13 by afakhry@chromium.org, Sep 28

Components: Blink>Canvas OS>Kernel>Graphics

Comment 14 by schenney@chromium.org, Oct 22

Components: -Blink>Canvas
Labels: -Pri-1 Pri-2
►

Sign in to add a comment