896627 - [8.1] Browser has extremely poor performance on recent m-c

Status: RESOLVED DUPLICATE of bug 893412

(Firefox for Metro Graveyard :: General, defect, P2)

Description

[Jonathan Wilde [:jwilde]]

On -metrodesktop, things are okay performance-wise. Like, the animation for bringing in the tabs is a little jankier than I remember it last being, but that's about it. Bubblemark running in a tab is reasonably fast (~150 - 200 fps).

On metro immersive, nearly everything--Bubblemark (~20 fps), page loading times, scrolling speed--is exceedingly janky to the point of being unusable for me.

Attaching video demos of each shortly.

---

This seems to be a Windows 8.1 issue. Current performance and behavior on 8.0 for me is pretty similar to -metrodesktop on 8.1.

---

Application Basics
------------------

Name: MetroFirefox
Version: 25.0a1
User Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:25.0) Gecko/20130722 Firefox/25.0

Extensions
----------

Important Modified Preferences
------------------------------

browser.cache.disk.capacity: 358400
browser.cache.disk.smart_size.first_run: false
browser.cache.disk.smart_size.use_old_max: false
browser.cache.disk.smart_size_cached_value: 358400
browser.startup.homepage_override.mstone:
extensions.lastAppVersion: 25.0a1
gfx.direct3d.last_used_feature_level_idx: 0
network.cookie.prefsMigrated: true
places.history.expiration.transient_current_max_pages: 56348
plugin.disable_full_page_plugin_for_types: application/pdf

Graphics
--------

Adapter Description: AMD Radeon HD 6770M
Adapter Drivers: aticfx64 aticfx64 aticfx64 aticfx32 aticfx32 aticfx32 atiumd64 atidxx64 atidxx64 atiumdag atidxx32 atidxx32 atiumdva atiumd6a atitmm64
Adapter RAM: 1024
Device ID: 0x6740
Direct2D Enabled: true
DirectWrite Enabled: true (6.3.9431.0)
Driver Date: 9-27-2012
Driver Version: 9.2.0.0
GPU #2 Active: false
GPU Accelerated Windows: 1/1 Direct3D 11 (OMTC)
Vendor ID: 0x1002
WebGL Renderer: Google Inc. -- ANGLE (AMD Radeon HD 6770M Direct3D9Ex vs_3_0 ps_3_0)
windowLayerManagerRemote: true
AzureCanvasBackend: direct2d
AzureContentBackend: direct2d
AzureFallbackCanvasBackend: cairo
AzureSkiaAccelerated: 0

JavaScript
----------

Incremental GC: true

Accessibility
-------------

Activated: true
Prevent Accessibility: 0

Library Versions
----------------

NSPR
Expected minimum version: 4.10
Version in use: 4.10

NSS
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSSMIME
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSSSL
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSUTIL
Expected minimum version: 3.15.1 Beta
Version in use: 3.15.1 Beta

Comment 1

[Yuan Wang(:Yuan) – Mobile Firefox Design Lead]

Application Basics
------------------

Name: MetroFirefox
Version: 25.0a1
User Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:25.0) Gecko/20130719 Firefox/25.0

Extensions
----------

Important Modified Preferences
------------------------------

browser.cache.disk.capacity: 358400
browser.cache.disk.smart_size.first_run: false
browser.cache.disk.smart_size.use_old_max: false
browser.cache.disk.smart_size_cached_value: 358400
browser.startup.homepage_override.mstone:
extensions.lastAppVersion: 25.0a1
network.cookie.prefsMigrated: true
places.database.lastMaintenance: 1374244627
places.history.expiration.transient_current_max_pages: 53674
plugin.disable_full_page_plugin_for_types: application/pdf
privacy.donottrackheader.enabled: true
storage.vacuum.last.index: 1
storage.vacuum.last.places.sqlite: 1373315233

Graphics
--------

Adapter Description: Microsoft Basic Display Adapter
Adapter Drivers: Unknown
Adapter RAM: 0
Device ID: 0x0405
Direct2D Enabled: Blocked for your graphics card because of unresolved driver issues.
DirectWrite Enabled: false (6.3.9431.0)
Driver Date: 6-21-2006
Driver Version: 6.3.9431.0
GPU #2 Active: false
GPU Accelerated Windows: 1/1 Direct3D 11 (OMTC)
Vendor ID: 0x15ad
WebGL Renderer: Blocked for your graphics card because of unresolved driver issues.
windowLayerManagerRemote: true
AzureCanvasBackend: skia
AzureContentBackend: none
AzureFallbackCanvasBackend: cairo
AzureSkiaAccelerated: 0

JavaScript
----------

Incremental GC: true

Accessibility
-------------

Activated: true
Prevent Accessibility: 0

Library Versions
----------------

NSPR
Expected minimum version: 4.10
Version in use: 4.10

NSS
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSSMIME
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSSSL
Expected minimum version: 3.15.1 Basic ECC Beta
Version in use: 3.15.1 Basic ECC Beta

NSSUTIL
Expected minimum version: 3.15.1 Beta
Version in use: 3.15.1 Beta

Comment 2

[Jim Mathies [:jimm]]

Juan is there anyone in QA with an 8.1 tablet? We'd like to track down a regression range on this. Apparently it started fairly recently on mc.

Comment 3

[juan becerra [:juanb]]

Samvedna is using a Samsung machine with 8.1. I'll work with her to try to find a regression range.

Comment 4

[Samvedana (:Samvedana)]

Yes, I have Samsung machine with 8.1. I am trying to find regression range.

Comment 5

[Samvedana (:Samvedana)]

@Jim Mathies
Please assign priority to this bug. Due to mid air collision it is removed from this bug. Sorry about that.

Comment 6

[Samvedana (:Samvedana)]

I tried to find out the regression window. The details are as below:

7/12/2013 nightly build is good.
User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:25.0) Gecko/20130712 Firefox/25.0
Build ID: 20130712030203

7/13/2013 nightly build is bad. 
User Agent: Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:25.0) Gecko/20130713 Firefox/25.0
Build ID: 20130713030205

I am seeing issues from 7/13 to 7/23 nightly builds. Metro Firefox has become slow on each nightly. 7/23 nightly is slower compare to 7/13 nightly.

I am trying to find change-sets.

Comment 7

[Samvedana (:Samvedana)]

Pushlog URL: http://hg.mozilla.org/mozilla-central/pushloghtml?startdate=2013-07-12&enddate=2013-07-13


Let me know if more detail is required.

Comment 8

[Masatoshi Kimura [:emk]]

Could you narrow down the regression range on m-i?

Comment 9

[Samvedana (:Samvedana)]

(In reply to Masatoshi Kimura [:emk] from comment #8)
> Could you narrow down the regression range on m-i?

Hi Masatoshi,
How can I narrow-down it? Can you please provide me some hint?

Comment 10

[Masatoshi Kimura [:emk]]

mozregression --repo=mozilla-inbound

Comment 11

[Samvedana (:Samvedana)]

(In reply to Masatoshi Kimura [:emk] from comment #10)
> mozregression --repo=mozilla-inbound

I am getting following output:

Sam@WINDOWS8_1 ~
$ mozregression --repo=mozilla-inbound --good=2013-07-12 --bad=2013-07-13


Last good nightly: 2013-07-12
First bad nightly: 2013-07-13

Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?startdate=2013-07-12&enddate=2
013-07-13

do you want to bisect further by fetching the repository and building? (y or n)

Am I doing right?

Comment 13

[Marco Mucci [:MarcoM]]

Hey Brian, would you like Bug 896627 to block Bug 898795?

Comment 14

[Brian R. Bondy [:bbondy]]

yep, done.

Comment 15

[Tim Abraldes [:TimAbraldes] [:tabraldes]]

Consolidating a couple bugs that seem to cover the same issue:
  Bug 896627 and bug 893412 are both about the poor performance of metroFx on Windows 8.1. Marking the newer bug as a duplicate of the older bug.

It is likely that bug 907410 is causing a large portion of our performance issues. Marking that as blocking bug 893412.