Recommended Posts

Alright, I've been holding off posting this, but I just can't ignore it anymore. For the past week I've been trying to fix this.

I previously talked about Google Maps being stuttery and actually running better in the in-game Steam browser.

Well I recently realized that ALL Youtube videos stutter as well, dropping around 5 to 10 frames every 10 to 20 seconds. Always.

Now, back then I was still running a 4890, so I thought maybe the issue with Google Maps was due to AMD's Legacy drivers.

But no, I've since upgraded to a GTX 660 Twin Frozr III and the issue is still there, plus, the flash video issue.

The kicker? If I use IETab in Firefox on Youtube, there's zero frames dropped. Both Chrome and IE run flash (and maps) flawlessly, so the only conclusion I can come up with is that the issue is Firefox, not my computer.

I tried both the stable (17.0.1) and latest Nightly and the issue was there. I tried to do alot of things short of reinstalling Windows. Safe mode, new profile, disabling protected mode, HW acceleration off in Fx, off in Flash, off in both, on in both. Nothing could fix it.

I knew Firefox had performance issues but I didn't realize it was this bad. Makes me sad considering I've been using it since the Firebird days...

Now, a few questions, has anyone noticed this issue? Or was able to fix it? And when is Youtube gonna implement HTML5 on all videos!?

Servo is actually going to do that :)

what i dont get, if they are gonna ditch Gecko at some stage, why not just Ditch the whole Browser ( firefox ) an make a New Browser with the New rendering engine in it. , surely that would be easier

full HTML5 implementation wont come for years.IMO. so i wouldnt get ya hopes up to high

anyone know if there are any New builds with inContent ?

Not yet but might be soon since Download Panel is on its polishing phase so it requires Library window for Downloads handling (in future) that's why.

Servo is actually going to do that :)

So it is where they are doing Rust based future Firefox bug thanks for sharing it but I still consider as a business student, time to market is too late. Ehsan and Areyh did some excellent work to make Gecko little C++11 compliant and still working on it. It will help somewhat but not much. The reason why we see too much progress and intention on JS side is the good example that it code changed quite a lot and well documented except GC case where BillM and Terrence are doing heroic work.

what i dont get, if they are gonna ditch Gecko at some stage, why not just Ditch the whole Browser ( firefox ) an make a New Browser with the New rendering engine in it. , surely that would be easier

I'm not sure you understand quite how complicated a web browser is. There's a reason there are only, what, 4 codebases out there. Gecko, Trident, Presto and WebKit.

  • Like 2

I'm not sure you understand quite how complicated a web browser is. There's a reason there are only, what, 4 codebases out there. Gecko, Trident, Presto and WebKit.

And WebKit started as a fork of KHTML, Apple thought it was easier than writing their own engine (And Google decided the same)

Ion Monkey now do compilation off the main thread, enabling bug landed on Inbound. Several amazing bugs related to JS engine landing which will probably improve performance quite a lot. Few Snappy P1 bugs are near to land out of which one even landed it was related to Download API async.

Download attribute bug fixes and some other performance improvement coming. I am loving FF20 without Australis. Amazing performance.

Interesting bugs:

Cache font mechanism and sharing between pages (something like this) - https://bugzilla.mozilla.org/show_bug.cgi?id=816483

Ion Monkey - https://bugzilla.mozilla.org/show_bug.cgi?id=727857

Constant Invalidation fix on Tomshardware - https://bugzilla.mozilla.org/show_bug.cgi?id=805343

Mac only I think DLBI issue fix - https://bugzilla.mozilla.org/show_bug.cgi?id=819837

Regression regarding background - https://bugzilla.mozilla.org/show_bug.cgi?id=818643

Regression of CSS3D Transforms - https://bugzilla.mozilla.org/show_bug.cgi?id=815666

  • Like 3

I agree with The_Decryptor that current DLBI implementation is better than previous since new implementation improved rendering, scrolling and painting a lot, site with DHTML improved a great deal.

And WebKit started as a fork of KHTML, Apple thought it was easier than writing their own engine (And Google decided the same)

Indeed. Although I think (and correct me if I'm wrong) that now KHTML and WebKit share patches and remain broadly compatible in terms of source code, so they are, for all intents and purposes, one codebase.

Yeah, main development has shifted over to WebKit, but KHTML is still kicking around for a few reasons (Although it's not as refined as WebKit or up to date, but some of the Linux guys don't like Apple and Google controlling their web browsers, quite rightly)

Zlip792, once more thank you for your Interesting/Important bugs list, always appreciated, I always come to this topic to find your posts :)

But let's be honest Firefox performance sucks... And yes, I am a Firefox user, been using it for the last 8 years.

I don't care about browser benchmarks, most of them are pointless. """" woohoo I'm 1ms faster than you!!!! """"

What I care is that I have PC's where I install Firefox and not only have I to turn off HWA but sometimes uninstall it completely.

It's GUI get's stuck on "Not responding" loops many times. It can't even use more than a core. And most netbooks I came across just run it horribly.

My Asus 1201N, with a dual core Atom, overclocked to 2.0GHz, with 4GB of RAM, ION GPU with 512 Ram, runs it like crap, HWA is off always.It's new tab, close tab animation is like watching a turtle go from left to right on my 24" screen!

Seriously ever since Electrolysis was canceled (Snappy and Super Snappy don't have the same goals, or have any ambition like it had), I lost hope that Firefox will be a better product from a performance standpoint in the near (or even longer) future.

Quite honestly, In the beginning of 2013 I'm going to give Chrome a shot. If the extensions do work like the ones I use in Firefox I'll probably just jump ship.

Zlip792, once more thank you for your Interesting/Important bugs list, always appreciated, I always come to this topic to find your posts :)

But let's be honest Firefox performance sucks... And yes, I am a Firefox user, been using it for the last 8 years.

I don't care about browser benchmarks, most of them are pointless. """" woohoo I'm 1ms faster than you!!!! """"

What I care is that I have PC's where I install Firefox and not only have I to turn off HWA but sometimes uninstall it completely.

It's GUI get's stuck on "Not responding" loops many times. It can't even use more than a core. And most netbooks I came across just run it horribly.

My Asus 1201N, with a dual core Atom, overclocked to 2.0GHz, with 4GB of RAM, ION GPU with 512 Ram, runs it like crap, HWA is off always.It's new tab, close tab animation is like watching a turtle go from left to right on my 24" screen!

Seriously ever since Electrolysis was canceled (Snappy and Super Snappy don't have the same goals, or have any ambition like it had), I lost hope that Firefox will be a better product from a performance standpoint in the near (or even longer) future.

Quite honestly, In the beginning of 2013 I'm going to give Chrome a shot. If the extensions do work like the ones I use in Firefox I'll probably just jump ship.

Really appreciated that bro.

Also I think Patrick changes regarding async proxy checking etc caused this Not Responding loop regressions and Honza is behind fixing it. Lets hope I am not confusing two issues although they might fix bugs. Firefox sucks when Windows Power Plan is not High Performance, there was a bug about it as well.

For all others, latest hourly support "per window" private browsing... Yuppeeeeeeeee!!!!!!!!!

Is this why the video on this page doesn't play in Nightly, but does in IE?

It might be. This will add Firefox support for H.264 codec as well and MP3 and other audio support. So overall win win situation for us and kind of Flash Plugin more quickly hurry to death.

I might try Nightly again soon. I haven't really used it much for the last few months after experiencing a period of instability.

I've been getting a lot of bug spam about the private browsing window feature since then and i am hearing good things about performance improvents so I might as well give it a look.

Nightly since last week feels a lot faster to me , the context menus open up faster , scrolling is quick. Feels great!

It is reverse for me, few days back Nightly was smacking fast but now it regressed little bit.

Off the Main Decoding is soon to come in Nightly so watch out for another performance improvement. Some session restore cleanup work in also work in progress which will make it faster. So overall going in good direction except Australis work.

*snip*

I have to agree here. It is like Mozilla is lying about doing anything to address performance issues.

DLBI? No performance benefit - scrolling and UI is laggy.

Some JS fix here or there? No performance benefit - scrolling and UI is laggy.

Azure something or other? No performance benefit - scrolling and UI is laggy.

Invalidation something or other? No performance benefit - scrolling and UI is laggy

And threading / multi-process simple moves bad computation to another core (if you have one).

What you will see is simple more of cores maxed out and same performance.

I think the low-end systems they test FF on can run Skyrim on high settings.

  • Like 1

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Posts

    • Microsoft: Windows 11 KB5094126, KB5093998 finally stops trusting a critical system threat by Sayan Sen This week Microsoft released the Patch Tuesday updates for June 2026 with KB5094126 on Windows 11 25H2, 24H2, and KB5093998 on Windows 11 23H2. On Windows 10 22H2 it's under KB5094127. Alongside the announced release notes for the new builds, Microsoft has revealed another change that is coming to Windows with these new releases. It has been confirmed that custom folders are getting a significant change with the June 2026 updates as such folders or folder names defined by desktop.ini will no longer appear after this update is successfully installed. While you may inititally think this is a bug with the new release, Microsoft has stated that this is in fact "expected behaviour" in its new support article regarding this which Neowin spotted today while browsing. Essentially it's a security hardening measure such that custom folder presentations are treated as potentially unsafe whenever Windows is not sure about their origin and whether that desktop.ini folder can be trusted or not. Here is list of such untrusted files and folders: Files downloaded from the internet that carry Mark-of-the-Web (MOTW). Files copied from certain remote locations, such as some WebDAV or HTTP-based locations. Files on network paths that are not classified as intranet or trusted by zone policy. For those who may not be familiar, Desktop.ini is a special configuration file used by Windows to customize the appearance and behavior of individual folders. Basically Windows can read specific instructions stored in Desktop.ini instead of displaying every folder with the same default settings. This can be used to apply custom icons, thumbnail images, localized folder names, and such informational tooltips (infotip). The file can also influence certain folder-specific behaviors and properties. It is typically stored as a hidden system file within a folder that has been designated to support Desktop.ini customization. However, because Windows Shell automatically reads and applies these attributes whenever a customized folder is opened, they have historically (since the Windows XP days) presented an attack surface as a result of an unchecked buffer in the Shell component responsible for extracting custom attributes from Desktop.ini files. As such an attacker could create a specially crafted Desktop.ini containing a malicious or corrupted attributes and place it on a network share. So if a user were to browse that folder, Windows would automatically process the file, potentially triggering a buffer overflow. This could allow arbitrary code to run with the same permissions as the logged-in user. Hence a seemingly harmless folder could become a security risk when their contents are not properly validated. For admins and users alike looking to manage this behaviour, Microsoft has shared a few ways. One of them is to assign a trusted mark on the folder in case you are sure of its source. Secondly a policy can be used to revert back to the previous state. Finally, the MOTW can be removed too to indicate to Windows that this is a safe file. The company explains: Option 1: Add the source to Trusted Sites (Recommended) If the affected content is stored on a known internal or managed source, add that source to the Trusted Sites list. Once the source is treated as trusted, Windows processes desktop.ini from that source normally. This keeps the protection in place for other locations and is the lower-risk option. Option 2: Use policy to restore previous behavior Organizations that need broader compatibility can enable the policy Allow the use of remote paths in file shortcut icons.Enabling this policy restores the pre-June 2026 behavior for affected remote or untrusted scenarios. Option 3: Check for and remove the Mark of the Web (MotW) If the desktop.ini file has a Mark of the Web (MotW), Windows may treat it as coming from an untrusted source and block customization. Verify whether MotW is present and, if appropriate, remove it from the desktop.ini file. This can restore expected behavior, but should only be done for trusted content, as it removes the associated security protection. To remove the MotW tag, open PowerShell and run one of the following commands: For a single desktop.ini file: Unblock-File "C:\Your\Folder\Path\desktop.ini" For all desktop.ini files in a folder: Get-ChildItem "C:\Your\Folder\Path" -Recurse -Filter desktop.ini -Force | Unblock-File Microsoft has warned though against using a broad opt-out using the provided policy as it reduces protection against potentially malicious remote folder-customization content. As such the tech giant recommends trusting only controlled internal sources and keeping trust settings as strict as possible. You can check out the official support article here on Microsoft's website.
    • LAV Filters 0.82.0 by Razvan Serea LAVFSplitter is a multi-format media splitter that uses libavformat (the demuxing library from ffmpeg) to demux all sorts of media files. LAV Splitter is a Souce Filter/Splitter required to demux the files into their separate elementary streams. LAV Audio and Video Decoder are powerful decoders with a focus on quality and performance, without any compromises. Supported Formats: MKV/WebM, AVI, MP4/MOV, MPEG-TS/PS (including basic EVO support), FLV, OGG, and many more that are supported by ffmpeg! LAV Filters are based on ffmpeg and libbluray and is aimed to offer a all-around solution to perfect playback of file-based Media as well as Blu-rays. LAV Filters 0.82.0 changelog: LAV Splitter NEW: Support for demuxing Dolby Vision Enhancement Layer streams NEW: Support for Animated WebP images Changed: When demuxing Blu-ray discs, Dolby Vision metadata is available on the primary video stream LAV Video NEW: Support for Animated WebP images Changed: Hardware decoding support for DVDs has been removed Download: LAV Filters 0.82.0 | 15.5 MB (Open Source) View: LAV Filters Website | Screenshot Get alerted to all of our Software updates on Twitter at @NeowinSoftware
    • For some reason when EU forced Microsoft to allow users to change the default browser and search provider in Windows (also no ads for Office and the likes) - it was good. But when it comes to Apple - then it's bad. BTW, Apple would have gone out of business if Microsoft wasn't pressed by US government several decades ago. 😉
  • Recent Achievements

    • One Month Later
      Sopa flores earned a badge
      One Month Later
    • First Post
      StaticMatrix earned a badge
      First Post
    • Week One Done
      StaticMatrix earned a badge
      Week One Done
    • Rookie
      lamborghiniv10 went up a rank
      Rookie
    • One Month Later
      pinnclepd earned a badge
      One Month Later
  • Popular Contributors

    1. 1
      +primortal
      506
    2. 2
      PsYcHoKiLLa
      207
    3. 3
      +Edouard
      156
    4. 4
      Steven P.
      88
    5. 5
      ATLien_0
      79
  • Tell a friend

    Love Neowin? Tell a friend!