Closed Bug 768752 Opened 12 years ago Closed 1 year ago

Firefox uses 600MB, freezes and becomes completely unusable

Categories

(Core :: Layout, defect)

13 Branch
x86
Windows Vista
defect

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: invented0, Unassigned)

References

(Blocks 1 open bug)

Details

(Whiteboard: [MemShrink:P2][Snappy:p3])

User Agent: Mozilla/5.0 (Windows NT 6.0; rv:13.0) Gecko/20100101 Firefox/13.0.1
Build ID: 20120614114901

Steps to reproduce:

Opened the following link.
http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&range=129376:135598&mode=html




Actual results:

Firefox becomes completely unusable for more than 2 minutes while using more than 600MB. Google Chrome has NO problem with the page and only consumes 180MB memory.
Both the browsers are at their latest version (FF13.0.1 and Chrome 20) and have the same set of plugins (Java, Flash)


Expected results:

Firefox should have been usable/available. Should be able to open other tabs/pages.
Google Chrome works just fine.
Whiteboard: [MemShrink]
My FF hangs for 30s or so while rendering the page, but that's no different than the HTML5 spec.

In general, we're bad at loading giant walls of text.  And since Firefox is single-threaded, doing a long operation like rendering this giant wall of text freezes the browser.

I'll mark this as blocking our text-suck metabug, but I don't think anyone is going to devote energy to fixing this particular bug anytime soon.
Component: Untriaged → Layout
Product: Firefox → Core
QA Contact: untriaged → layout
FWIW, about:memory on a current Linux/x86-64 nightly says:

│    │  ├──221.73 MB (19.07%) -- window(http://build.chromium.org/cgi-bin/svn-log?url=http://src.chromium.org/svn//trunk/src&range=129376:135598)
│    │  │  ├──192.73 MB (16.57%) -- layout
│    │  │  │  ├──150.24 MB (12.92%) -- frames
│    │  │  │  │  ├───69.75 MB (06.00%) ── nsTextFrame
│    │  │  │  │  ├───33.31 MB (02.87%) ── nsInlineFrame
│    │  │  │  │  ├───30.01 MB (02.58%) ── nsContinuingTextFrame
│    │  │  │  │  ├───17.17 MB (01.48%) ── nsBlockFrame
│    │  │  │  │  └────0.00 MB (00.00%) ── sundries
│    │  │  │  ├───35.70 MB (03.07%) ── line-boxes
│    │  │  │  └────6.79 MB (00.58%) ++ (5 tiny)
│    │  │  ├───28.98 MB (02.49%) -- dom
│    │  │  │   ├──16.60 MB (01.43%) ── text-nodes
│    │  │  │   ├──12.37 MB (01.06%) ── element-nodes
│    │  │  │   └───0.01 MB (00.00%) ── other [2]
│    │  │  └────0.03 MB (00.00%) ── style-sheets
│    │  └───83.10 MB (07.15%) -- window(http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&range=129376:135598&mode=html)
│    │      ├──51.97 MB (04.47%) -- dom
│    │      │  ├──38.46 MB (03.31%) ── element-nodes
│    │      │  ├──12.50 MB (01.08%) ── text-nodes
│    │      │  └───1.01 MB (00.09%) ++ (2 tiny)
│    │      ├──31.12 MB (02.68%) -- layout
│    │      │  ├──24.19 MB (02.08%) ++ frames
│    │      │  └───6.93 MB (00.60%) ++ (7 tiny)
│    │      └───0.02 MB (00.00%) ++ (2 tiny)
I also got ~30% heap-unclassified with that page, so we're probably missing something.
Oh wow.  Yes, I still see 25% heap-unclassified with that page even on a recent nightly.  I'll try DMD'ing that.
Filed bug 768903 for the high heap-unclassified numbers seen here.
Whiteboard: [MemShrink] → [MemShrink:P2]
Whiteboard: [MemShrink:P2] → [MemShrink:P2][Snappy]
Status: UNCONFIRMED → NEW
Ever confirmed: true
Whiteboard: [MemShrink:P2][Snappy] → [MemShrink:P2][Snappy:p3]
Severity: normal → S3

Unable to reproduce.

Status: NEW → RESOLVED
Closed: 1 year ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.