Closed Bug 1592169 Opened 5 years ago Closed 2 years ago

High heap unclassified memory usage with OpenSpeedTest

Categories

(Core :: DOM: File, defect, P2)

70 Branch
defect

Tracking

()

RESOLVED WORKSFORME
Tracking Status
firefox-esr68 --- wontfix
firefox70 --- wontfix
firefox71 --- wontfix

People

(Reporter: enigma1052666, Unassigned)

References

()

Details

(Whiteboard: [MemShrink:P2])

Attachments

(1 file)

User Agent: Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:70.0) Gecko/20100101 Firefox/70.0

Steps to reproduce:

Go to http://openspeedtest.com/ and do the speed test

Actual results:

While the test was successful, RAM usage of it's tab went to 2GB+ while the test was being made (returns to normal or almost normal after it ends), on lower spec machines, that can cause a lockup.

Expected results:

Not using 2GB of memory (even temporary) for the test.

Tried on Chromium Edge browser and Chrome, doesn't happen there, it stays at ~160MB

Status: UNCONFIRMED → NEW
Ever confirmed: true
Whiteboard: [MemShrink]

I took a memory report while this was running, and I ended up with a content process with 2.1GB of heap-unclassified. There was also around 100MB of memory-file-data/large/file. I'm going to guess this is DOM storage related, but somebody should put together a DMD build to be sure. I wonder if it is related to the string blob issue I was looking at earlier.

Component: Untriaged → DOM: Core & HTML
Product: Firefox → Core
Summary: OpenSpeedTest causes a memory leak → High heap unclassified memory usage with OpenSpeedTest
Whiteboard: [MemShrink] → [MemShrink:P2]

I also noticed that my heap-unclassified was really high in the parent process, and it didn't go away when I closed the speed test page, but I had that browser session open for a long time, so it might have been unrelated.

Component: DOM: Core & HTML → DOM: File
See Also: → 1591722
Priority: -- → P2
Severity: normal → S3

I made a test with Nightly 96 and took a profile. Memory peak usage of the openspeedtest tab is 425MB. CPU usage is very high and most of it seems to be graphics related for painting the updates. But it all turns normal once the test finishes.

I think it is (was) unrelated to the Deezer issue, too. I'll close this WORKSFORME but feel free to re-open if you still see this.

Status: NEW → RESOLVED
Closed: 2 years ago
Resolution: --- → WORKSFORME
See Also: 1591722

I don't think users care about FF70, FF71, and FF68ESR at this point?

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: