Comparing Memory Usage of Firefox 2 vs 3

DaMan writes "ZDNet picks up on yesterday's Firefox 3 beta 1 review by comparing the memory usage of Firefox 2 against the latest beta. The results from one of the tests is quite interesting, after loading 12 pages and waiting 5 minutes, 2 used 103,180KB and 3 used 62,312KB. IE used 89,756KB.""

And Opera

[Anonymous Coward]

is using 34mb (winXP)

How are they measuring?

[MyLongNickName]

Are they using the handy dandy Task Manager? If so, this is not even remotely accurate. In the age of managed memory, this is an estimate at best. Don't believe me. Open up internet explorer, run it a while and look at the memory usage. Now minimize IE. Watch the number drop like a lead balloon.

Re:not to point out the obvious

[ceejayoz]

From the article:

Both Firefox 2.0.0.9 and Firefox 3.0 b 1 were installed fresh using a standard install.

This explains a part of it (memory fragmenting)

[Anonymous Coward]

http://blog.pavlov.net/2007/11/10/memory-fragmentation/ [pavlov.net]

An interesting read on how memory fragmentation adversely affects FireFox... & why/how.

APK

P.S.=> I also recommend Opera for these reasons (less security holes period, & the 1 it had yesterday? Patched yesterday too... fast!)

SECUNIA DATA ON BROWSER SECURITY (dated 11/20/2007):

Opera 9.24 security advisories @ SECUNIA (0% unpatched):

http://secunia.com/product/10615/?task=advisories [secunia.com]

----

Netscape 9.0.0.3 (0% unpatched)

http://secunia.com/product/14690/ [secunia.com]

----

FireFox 2.0.0.9 security advisories @ SECUNIA (29% unpatched):

http://secunia.com/product/12434/ [secunia.com]

----

IE 7 (latest cumulative update from MS) security advisories @ SECUNIA (37% unpatched):

http://secunia.com/product/12366/ [secunia.com]

----

Those %'s are the latest for FireFox 2.0.0.9, Netscape 9.0.0.3, IE7 after last "patch Tuesday" from MS with the "CUMULATIVE IE UPDATES" they have (see the security downloads URL I post in the 12 steps above to secure yourself), & Opera 9.24... all latest/greatest models.

So, as you can see?

Well, NOT ONLY IS OPERA MORE SECURE/BEARING LESS SECURITY VULNERABILITIES?

It's faster too, on just about ANYTHING a browser does, & is probably the MOST standards compliant browser under the sun (not counting HTML dev tools). This is borne out in these tests:

http://www.howtocreate.co.uk/browserSpeed.html [howtocreate.co.uk]

AND, yes others (most recently in Javascript parsing speeds, oddly enough, lol... given the topic of my post here that is), right here:

http://nontroppo.org/timer/kestrel_tests/ [nontroppo.org]

Opera's just more std.'s compliant, faster, & more secure than the others... so, "where do you want to go today?"...

apk

Exact opposite on my machine

[Darth_brooks]

Either I got a bad build, or I've got a weird system setup. FF3b1 was using 180 megs (yes, 180 megs) of memory to load my intranet page, and would try and scream upwards from there before my poor IBM laptop (P3 800, 320 megs of ram) ground to a halt. FF 2.0.9 was using 30 megs.

I wish I could have submitted a bug report, but my machine would freeze before firefox actually crashed.

(and no, it does also take me 15 minutes to move a 20 meg file on my mac.....)

Re:Strange, 1p/10 mins more than 12pp/5 mins?

[stony3k]

Form the tests that the developers have been running, most of the memory leaks in Firefox itself seem to be fixed (there are probably still some left). However, memory usage still remains a problem. I think this blog post [pavlov.net] summarizes their findings. They've been using dtrace and other tools to find out exactly what is going on.

Unfortunately, I think the damage to Firefox's reputation is already done. There are many people who have had negative experiences with Firefox who keep on harping about the "memory leaks" and I don't see how Mozilla devs can change this public perception.

Yet flash....

[webmaster404]

On Ubuntu 7.04 and 7.10 if you install the flash plugin nonfree package from apt-get flash works fine but whenever you try installing it from Adobe's site or the auto plugin installer, FF grinds to a halt on it using around 100 CPU on anything Flash related like Youtube or Slashdot's ads, disabling flash solves it, however on my other computer that is not much more powerful (slower clock speed of CPU but higher bus speed) when I installed it from the auto plugin installer it works fine getting only around 50% of CPU Max. Firefox or Adobe needs to fix this so Linux people can test the binary that requires you to install the auto-plugin and doesn't work with flash-plugin-nonfree. However, Firefox 3 is my preferred browser on my other computer and it was on Windows even more. My question is, why can't Firefox produce either a sane way to compile it (its a pain to compile it already...) or supplying .deb and .rpm for the builds to make it easier to install? Linux seems to be neglected by Firefox lately, with more strategy of stealing IE's market share then making a better browser on Linux. And Konqueror is painfully slow when on XFCE or GNOME (or just about anything thats not KDE) but perhaps KDE 4 will fix that....

Re:How are they measuring?

[volpe]

When you minimize, the working set size is reduced. This causes pages to be swapped out to the pagefile. When you maximize (or restore), the working set size is increased, meaning that the application is *allowed* to use more physical memory, but that doesn't mean it's going to immediately start loading back the same pages it swapped out. It's going to wait for page faults to compel it to do so. That is why #6 is lower than #2.

Won't be going back

[scubamage]

So I started using the beta yesterday, and I can say that I won't be going back to IE or FF2. It runs extremely fast, stable, and is nice and polished. It seriously reminds me of the early releases of FF, but much, much faster. I've got about 14 tabs open right now, and its still running screaming fast. The earlier /. article is no lie, it installs in a heartbeat, opens fast, closes fast, even browses fast (as would be assumed given that it uses a smaller memory footprint, though I could be wrong about that). I reccomend.

Memory still an issue for me...

[Anonymous Coward]

Installed and fired up firefox 3 beta 1. Went to visit www.speakeasy.net/speedtest, couldn't even hit enter. The default page wasn't even loading. My system slowed to a crawl. I checked the availible RAM, and of the 1GB I have in this system I had 2 megs free. Here Firefox was using 707.13 Megs of RAM... don't think the memory leak has been complete fixed (yes this was a windows machine...)

Re:And Opera

[CastrTroy]

Not raw like your camera raw, but raw uncompressed format. All those JPGs and GIFs and PNGs must be converted to raw or bitmap files to display them on the screen.

Re:Yes, but... looking in the wrong spot! :)

[damaki]

Here is what you are searching for : http://www.microsoft.com/technet/sysinternals/utilities/processexplorer.mspx [microsoft.com]

Re:Memory still an issue for me...

[magicsquid]

I wish I had mod points, because this needs to be brought to people's attention. Everyone seems to be claiming victory over the memory bugs, but for me (and you and many others) there are still random problems.

My system exhibits the exact same problem you describe. My Firefox will spike from around 66 MB of RAM usage to 700 then 800 then 900 and will just sit there chewing up more RAM until I kill it. I'd love to know the cause and even better, the solution to this problem.

It is happening in FF2 and in the 3 Beta. It doesn't happen on the same site every time. It happens most frequently when using JavaScript, but not always. I can't seem to narrow it down unfortunately.

Opera 9.5

[Nicolay77]

Your complains are real.

Because of that, Opera has two features you could find useful:

• Mask as: as most sites don't work because of improperly and outdated sniffing code, you can make Opera pretend to be IE or Fx, this setting can be global, or can be per domain.
  
• Opera Sync: The upcoming Opera 9.5 has syncing between accounts too.
  

As you see, Opera deserves its good reputation because they are updating the browser all the time adressing all kind of issues.

(And I'm glad you posted real issues, not the same old 'extensions, extensions, extensions!')

Just did this test in linux

[Adult film producer]

FC 6 .. kernel 2.6.22.. Firefox 1.5.0.12 vs 3.0b1

I created a new user, logged in and loaded up FF 1.5.. opened up 12 tabs and logged into these sites

www.bbc.co.uk
www.slashdot.org
www.dailykos.com
www.news.com
www.abc.com
www.foxnews.com
www.freep.com
www.youtube.com
www.youporn.com
www.liveleak.com
www.rawstory.com
www.drudge.com

Here are the numbers for ff 1.5. The first line is when it loaded up with 12 empty tabs. The second line is the 12 websites loaded initially.. and the third line is 12 minutes afterwards

  3876 perfume 20 0 175m 54m 38m S 0.0 14.5 0:18.19 firefox-bin
  3876 perfume 20 0 348m 124m 49m R 72.0 33.2 1:47.83 firefox-bin
  3876 perfume 20 0 338m 135m 49m R 46.8 36.0 7:30.93 firefox-bin

I logged out, rm -rf ./.mozilla then logged back in and fired up FF 3.0b1.. same procedure, same 12 websites and 12 minutes of idling on them

  4231 perfume 20 0 202m 58m 38m S 3.6 15.6 0:11.79 firefox-bin
  4231 perfume 20 0 273m 106m 40m S 59.7 28.4 1:31.37 firefox-bin
  4231 perfume 20 0 254m 107m 40m S 1.3 28.5 2:27.26 firefox-bin

CPU usage seemed to be much better with FF 3B1 as well.. not sure why the difference but everything was clean...

Regrettably It Also Locked Up!

[Toad-san]

After about 2 minutes of use, 2 or three different pages online .. the new 3.0 slowed down my entire system to a crawl, and finally to a lockup. Had to pull the plug.

Rebooted (Win2K, 2.8 MHZ Pentium 4, 1GB RAM), manually fired up ye olde Firefox, went to same pages, ran fine.

Closed, re-ran 3.0 .. same problem.

Sorry boys, not ready for Prime Time IMHO.

Still giving issues

[kryliss]

I just downloaded and installed FF3beta, opened up slashdot and BAM....

http://home.windstream.net/slashdot/pics/firefox3beta.jpg [windstream.net]

Re:And Opera

[LizardKing]

The "w3" web browser extension for Emacs can display images.

Re:Yeah but it's still beta

[GoodbyeBlueSky1]

Try creating a new firefox profile and see if you get the same thing.

I've been running nightly FF3 pre-beta builds for a few months now, and even on the occasional day where a new patch causes regular crashes I've not seen this happen.

Re:And Opera

[s4m7]

All those JPGs and GIFs and PNGs must be converted to raw or bitmap files to display them on the screen.

Well, yes and no. They do have to be converted when placed into display memory (VRAM) but the decompression *can* happen on the fly. I believe that it usually doesn't for performance reasons though, since each scroll of the page would require another decompression pass.

Dillo is back

[Kelson]

The real shame of it is that the Dillo project is on hold now, even though with the tiniest fraction of the resources of the Mozilla project, it could very quickly become an absolutely amazing web browser. It's really the same thing that happened with Links-GUI... Two amazingly promising browsers, going nowhere.

They finally managed to get the code released for the half-finished port to FLTK last month, and there's been a massive flurry of activity on the developers mailing list [wearlab.de] and in CVS. I guess no one's updated the project web page yet.

Re:And Opera

[BZ]

> Point me to the "borked pages" code, and I'll be damn happy to remove it
> if it will give a huge performance boost

http://lxr.mozilla.org/seamonkey/source/parser/htmlparser/src/CNavDTD.cpp [mozilla.org]
http://lxr.mozilla.org/seamonkey/source/content/html/document/src/nsHTMLContentSink.cpp [mozilla.org]

Re:Yes, but... looking in the wrong spot! :)

[X0563511]

An excellent tool. Careful using it though, as it attaches to the system through debugging hooks and hence certain copy protection systems scream at you and make you reboot, and not run it... I'm staring at you SecureROM!

Re:How are they measuring?

[Lemmingue]

When you minimize, the working set size is reduced. This causes pages to be swapped out to the pagefile

Nope. This way your hd would trash like hell every time you minimize a big program. Just imagine a program using 1GB ram being minimized in a notebook with a 4200 RPM hard drive...

Windows Memory Manager just set the flag V (for Valid) in the Page Table Entry (PTE) as zero, but the virtual/physical address association is still held. If you restore the window the Memory Manager will just restore the PTE flags and you're good to go. You can check the "Windows Internals" book or the Intel Manuals.

Re:And Opera

[dvice_null]

> On some of my machines, I just never installed it. Others, I have removed it. But, it is so damned pervasive that there are some web sites that simply don't work without flash

I repeat what you were already told. Install Flashblock add-on:

https://addons.mozilla.org/en-US/firefox/addon/433 [mozilla.org]

What does it do you ask? Well, it does everything you want. It disables all flash by default. You can whitelist a site (like youtube) to always show flash. Or you can simply single click the space where flash would normally be to play it. I seriously suggest that _everyone_ should try it. It is one of those add-ons that make like a lot easier.

Re:Yeah but it's still beta

[jnaujok]

It's already been reported. It's caused by the anti-phishing stuff built into Firefox. Apparently, despite the fact they could simply copy the 16MB Sqlite file down and use it, they choose to send down the data and then reload it into the local database. That's what burns the time, CPU, Disk, etc. for nearly 2 and a half hours.

Re:And Opera

[daniel_newton]

plugins (dlls) generally run in-process

Re:And Opera

[rtaylor]

Unfortunately that is an operating system problem.

We need a way to tell the operating system that some memory is important and other segments can be dropped at any time (cached or precalculated data) provided the application is told so it can rebuild it when necessary.

The OS scheduler would choose which applications are idle to the user and dump some of the applications data.

Experiencing RAM/CPU Problems? Try Again.

[asa]

I think we've got to the root of the problem that you and some other Firefox 3 Beta 1 testers are seeing.

Starting yesterday, we began receiving reports, like yours, of a new memory/cpu usage issue that happens shortly after a normal startup and can spike the CPU and chew up hundreds of MB of RAM. This is apparently happening to people with new profiles or in profiles that have a very outdated list of bad sites for the Phishing Protection feature.

What's going on is that soon after Firefox is started, Firefox tries to fetch updates to the site forgery list -- the lists of bad sites that allows Firefox to warn users about suspected Phishing attacks. If the profile has very outdated or no local list, as is the case for a new Firefox profile, Firefox is trying to bring down a complete, rather large, list in one big chunk rather than slowly in small chunks. This causes Firefox to consume large amounts of CPU and memory and can slow the users machine to a crawl.

This problem is due to the change in the "SafeBrowsing Protocol" which only affects Firefox 3 Beta 1 and nightly build users. If you're on Firefox 2, this isn't going to affect you.

The work-around for this problem was for us to throttle it on the server side. We've done that and if you try Firefox 3 Beta 1 again, it should be fine.

- A

Re:Just did this test in linux

[dkf]

Can someone please tell me what the columns are in English? While it's great to know how much "NI" Firefox has, I'd rather see the memory usage.

That's an ignorable column. Instead, pay attention to "VIRT" (virtual memory used) and "RES" (approximately the physical memory used). In particular, note that both of those figures are relatively lower once a few sites are visited, meaning that FF3.0b1 is both less memory hungry and less inclined to touch pages that it is using (i.e. should be better performance in practice, especially if you're doing other things with the machine too, such as running an IDE.)

Re:How are they measuring?

[Foolhardy]

When you minimize, Win32 sets the working set of the process is set to its minimum (a few MB). Each of the pages removed is indeed marked as invalid in the page table entries of the process. Physical pages have a reference count: removing the page from the process working set reduces the count by 1. When the reference count reaches 0, the page is moved to the standby list. In the standby list, the page cannot be modified (as it's no longer mapped anywhere). A copy is written to disk lazily, but it still exists in memory as long as it's in standby. That way, if the page needs to go back into a working set, it doesn't have to be read from disk, and if the memory is needed for something else, a copy has already been written to disk. When looking for new memory to allocate, the memory manager uses up the free memory first, and then standby memory. When that runs out, it aggressively trims process working sets.

In short, you're both right. In the 2000/XP task manager, standby memory is counted as both "available" and "system cache" (I guess because it's both available for re-use without disk access, and a type of cache).