1369966 - Firefox 53.0.3 very slow, uses excessive CPU

Status: RESOLVED DUPLICATE of bug 1359550

(Firefox :: Untriaged, defect)

Description

[[e-mail]]

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:53.0) Gecko/20100101 Firefox/53.0
Build ID: 20170518000419

Steps to reproduce:

Browse using Firefox 53.0.3 x86 on Windows 7


Actual results:

After some browsing, it becomes very slow, as in unusable and unresponsive. CPU usage rises to continuous 25%, even doing nothing. Closing tabs does not help. Trying to close the program brings the "unresponsive program" dialog, but if you wait it will eventually close.


Expected results:

Program should be responsive and not consume so much CPU.

Comment 1

[[e-mail]]

Browsing eBay.com seems to trigger the problem.

Comment 2

[Loic]

Could you type about:support in the location bar and paste the section "graphics".

Comment 3

[[e-mail]]

Sure. Snapshot below is from Firefox 52.0.3 64 bits - I ditched the 32 bit version that behaved badly. It looks like the 64-bit also eventually gets into a situation where it eats up 10% CPU doing nothing, and afterwards CPU usage never goes back down. But it does not stop to a complete crawl to the point of becoming unresponsive like the x86 version did on my machine. And may I add, apart from that, your browser is awesome :-).

Graphics
Features
Compositing	Direct3D 11
Asynchronous Pan/Zoom	none
WebGL Renderer	Google Inc. -- ANGLE (NVIDIA GeForce GT 240 Direct3D11 vs_4_1 ps_4_1)
WebGL2 Renderer	Google Inc. -- ANGLE (NVIDIA GeForce GT 240 Direct3D11 vs_4_1 ps_4_1)
Audio Backend	wasapi
Direct2D	true
DirectWrite	true (6.2.9200.22082)
GPU #1
Active	Yes
Description	NVIDIA GeForce GT 240
Vendor ID	0x10de
Device ID	0x0ca3
Driver Version	21.21.13.4200
Driver Date	10-18-2016
Drivers	nvd3dumx,nvwgf2umx,nvwgf2umx nvd3dum,nvwgf2um,nvwgf2um
Subsys ID	075b196e
RAM	512
Diagnostics
AzureCanvasAccelerated	0
AzureCanvasBackend	direct2d 1.1
AzureContentBackend	direct2d 1.1
AzureFallbackCanvasBackend	skia
Decision Log
D3D9_COMPOSITING	
disabled by default: Disabled by default
GPU_PROCESS	
unavailable by runtime: Multi-process mode is not enabled

Comment 4

[[e-mail]]

Sorry, I meant Firefox 53.0.3 64-bit in the first line.

Comment 5

[Mr Alexander]

Using Firefox 53.0.3 on Mac OSX 10.10.5.  Plenty of CPU/RAM.  Java 8 Update 131 installed. Flash NPAPI Plug-in version 25.0.0.171 is installed.

Since Firefox V 53, it is dire!

VERY SLOW, unless I type into the address bar immediately after launch: "about:memory" and click on "Minimise memory usage", and only then start browsing.
On launch, nonetheless, Firefox uses a massive 2 GB of available RAM which is entirely unacceptable.

Some websites no longer function properly.

Firefox has disabled most of the plugins and seems to be set on moving away from its simple, sleek and easy design. 

Problems with current Firefox:
-SLOW
-SLUGGISH (not snappy at all but often lags)
-FAILS to display all of site information (e.g. fails to display mandatory tick / check boxes on some sites)

Larger issues:
The internet design standards (relentless push for html5).  This has killed all fun, beauty, depth of icons, malleability and many plugins.

SOLUTION: Install a different browser and ditch Firefox (sorry).

Comment 6

[Emil Pasca [:emilpasca]]

Since the issue specified here seems to be the same problem as in the following bug 1359550 (high CPU usage, PC setup, ebay webpages and other similar sites), I will mark this as a duplicate of it.

Marc if you may, please add future info related to this issue to that bug, so that we can keep track of this and debug it further.

Thank you.

Comment 7

[(Away)]

I'm not sure if others will disagree with me, but it may be better to wait until we have a clear diagnosis before resolving as duplicate. "High CPU usage" is quite a broad symptom and can have a huge number of causes. If we have two bug reporters commenting on bug 1359550 and their underlying causes turn out to be different, it can result in confusion for the bug investigators.

Comment 8

[[e-mail]]

I'm fine either way. I have to say the bug this report was folded under looks awfully similar indeed.