Memory pressure causes 40FPS graphics regression

RESOLVED WORKSFORME

Status

()

Core
Graphics
RESOLVED WORKSFORME
4 years ago
3 years ago

People

(Reporter: ladamski, Unassigned)

Tracking

Firefox Tracking Flags

(Not tracked)

Details

(Reporter)

Description

4 years ago
STR:

Latest 2.0 nightly (20140725000201)

1) Get game "X Type" from marketplace

2) Configure device to 512MB, game loads and you get 55-60 FPS at main menu and during gameplay

3) Configure device to 319MB, you get 15-20 FPS

Everything else is equal.  I even swapped devices and saw the same effect.  Seems like we're thrashing somewhere and wasting a lot of FPS and likely power too
I'm not sure if the regression in the title meant this is a regression from a previous version.
Keywords: regressionwindow-wanted
I guess there are multiple questions here at this point:
1) Does it still reproduce on newer B2G versions?
2) Is this a memory config we're committed to supporting at this point?
Flags: needinfo?(nhirata.bugzilla)
We're supposedly moving to 512 MB; it's not official though.  I'm not sure if it will be official before the end of 2.5; we need to get new baseline measurements.

QAWanted to check to make sure that 512 MB maintains 55-60 FPS at main menu and game play on flame device.
I'm not sure if it's worth checking 319 MB...  though please do a quick check.  It will be good to know where we are at with the current version.
Flags: needinfo?(nhirata.bugzilla)
Keywords: qawanted
I checked both 512 and 319MB on the latest Nightly Central Flame and both saw a ~30 FPS drop during some portions of gameplay (enemies shoot out multiple projectiles at once, sometimes in multiple waves) but would otherwise run between 50-60 fps.  This was using the FPS counter in the Developer HUD though which as I understand is not always reflective of actual performance.

I don't believe that the regression in the title meant that this was a regression between versions considering the content of the description so I'm removing that tag for now.

Environmental Variables:
Device: Flame 2.5
BuildID: 20151013030230
Gaia: d400cda6bf0f8b30dcf7d7d71bfa61f29a3f1588
Gecko: 607a236c229994df99766c005f9ec729532d7747
Gonk: c4779d6da0f85894b1f78f0351b43f2949e8decd
Version: 44.0a1 (2.5) 
Firmware Version: v18D
User Agent: Mozilla/5.0 (Mobile; rv:44.0) Gecko/44.0 Firefox/44.0
QA Whiteboard: [QAnalyst-Triage?]
Flags: needinfo?(nhirata.bugzilla)
Flags: needinfo?(ktucker)
Keywords: qawanted, regressionwindow-wanted
QA Whiteboard: [QAnalyst-Triage?] → [QAnalyst-Triage+]
Flags: needinfo?(ktucker)
There's also a possibility that it's SWC versus HWC.
In any cases, it seems like it doesn't reproduce anymore?

RyanVM, passing this back to you.
Flags: needinfo?(nhirata.bugzilla) → needinfo?(ryanvm)
I have no horse in this race. If B2G QA is satisfied, then so am I.
Flags: needinfo?(ryanvm)
Resolving this issue as WFM.  If it happens again please reopen.
Status: NEW → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.