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.""

Comments

Yes, but...

[InvisblePinkUnicorn]

How much does it use on Linux... err... does it run... damn!

Sorry, I'm new at this....

And Opera

[Anonymous Coward]

is using 34mb (winXP)

Re:And Opera

[gEvil (beta)]

I'm assuming you loaded the exact same 5 pages with the same ads that the ZDnet testers did?

Re:And Opera

[morgan_greywolf]

Is it just me or does it seem like 60MB or even 34MB is a LOT of memory for something that browses Web pages?

I mean, people used to make fun of GNU Emacs, saying things like it stands for eight megabytes and constantly swapping or eventually malloc()'s all computer storage. Emacs takes somewhere around 10MB or so on a RHEL4 box, and that thing is practically an operating system. It reads mail! Firefox doesn't even read mail, and it takes 60MB. Opera reads mail, but still 34MB seems just too big, too.

Maybe I'm just getting to be a cranky old man. Now you kids get offa my lawn!

Re:And Opera

[gEvil (beta)]

Sorry, the same 12 pages with a 5 minute wait...

Re:And Opera

[beau_west]

I use FireFox and Opera very heavily at work. FireFox for any normal browsing/development. And Opera primarily for . What constantly surprises me is the amount of memory that Opera uses with one tab. It starts at about 74,000K and grows to obscene amounts. (It grows with every new song) Greater than FireFox with several instances running, which is currently using 308,000K on my machine. This could be due to poor Flash on Pandora's part, or, could there be problems with Opera?

Re:And Opera

[bmartin]

Opera's no saint. After running Opera and Firefox 2 with no plugins for quite some time, they used the same amount of memory on my computer (Linux). I've been running FF w/ no plugins for a very long time on my machine due to its limited memory (512 MB).

While the rendering engine has an obvious need for memory, it's nice to see that they're cutting down on memory usage; it has been one of the biggest drawbacks of using Firefox. There's really no need for a web browser to use more than 100MB of memory. For people with limited RAM, it'd make more sense at that point to cache the contents to disk re-render them when they're needed.

That said, Firefox does need a rewrite of the plugin system (which would be a BIG overhaul), but it's more important that they iron out the bugs for this release first. It'd be a huge step up for the Mozilla team, as long as they don't change the existing plugin API.

Re:And Opera

[Jarjarthejedi]

Not for modern webpages. A single flash ad requires a couple of megs, tack on the capability to have multiple pages open in a single browser (that adds to the memory usage a little), a bunch of these ads, and the actual page content and it's pretty small.

Actually something of interest I've noticed is that since I got NoScript my FF ram usage has dropped considerably. I rarely get about 83MB with FF2 now, because it doesn't have to load the plugins and such.

Re:And Opera

[Applekid]

IANAWBD (Web Browser Developer), but... there's just so much data for web pages now. You've got plugged-in interpreted flash code, graphics that need to be kept in RAW formats in memory because of speed, the full length and width of the page on an in-memory surface to pan through on a window.

Even then it still needs a dynamic layout for CSS and scripting on the fly. And even then some scripting is safe, some is not, so there are rules that the code has to implement like pop-up blockers, password managers, warnings on insecure pages, warnings on cross-site scripting, etc. All that and the browsers STILL need to be able to sensibly parse and display completely borked pages with invalid HTML.

Nevermind maintaining history, cache, cookies, referring pages, bookmarks.

Re:And Opera

[CastrTroy]

But Emacs doesn't display images (somebody will probably correct me on this). Just the cached copies of all the images can take up quite a bit of memory. And from what I remember, it has to basically uncompress them to bitmaps, and keeps those in memory, so that can eat up a lot of memory. Also, all the CSS, DOM, and other information that a text editor doesn't have to keep loaded all the time probably uses up a large amount of memory also. Not to mention plugins like flash and other things that probably use quite a bit of memory. A web browser isn't just something that displays text. If you want that, try Lynx, and you'll probably have a browser that uses very little memory. However, for those of us who want a graphical and eye-pleasing web experience, the browser is probably going to take quite a bit of memory for a long time to come.

Re:And Opera

[muyuubyou]

It's the "interwebs" what's really bloated. My Firefox executable here is taking just 7.3MB, but then you open several web pages that take MBs each, and some more in uncompressed, parsed form. Then some browsers cache other stuff like rendered pages in memory, and you get those figures we're talking about.

Re:And Opera

[Anonymous Coward]

I'm running Opera 9.5 right now, and it's a slow, fat pig. A fat pig with lots of useful features, but still a slow, fat pig.

Re:And Opera

[morgan_greywolf]

See what I mean? :-D We have to develop an extremely complicated application just to look at pretty pictures and text on the Intarweb tubes. I think the resident size on lynx is like a couple of meg. Yeah. Just checked.

Re:And Opera

[trosenbl]

Just to clear up one thing

...graphics that must be kept in RAW format...

There are no browsers I'm aware of that support the RAW image format, if that's what you meant. The memory overhead would be insane, not to mention file transfer times.

Don't get me wrong, there's a lot of kids out there that are playing web developer, who don't know that compressing photographs as GIF is bad, or who still use spacer images, or render paragraphs of text using JPGs, so there's room for improving the code that the browser has to handle, but it's not always quite as bad as your post suggests.

That being said, the nesting of tables, which still continues even on mainstream sites, requires a lot of memory to parse. Browser makers do a great job at dealing with the mess of code some people throw at them.

-- Tim

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:And Opera

[nicolastheadept]

There's no point RAM sitting there, not being used. It might as well be used to speed up page loading etc. rather than doing nothing.

Re:And Opera

[Anonymous Coward]

it does

Re:And Opera

[Anonymous Coward]

Let's see, an 800x600 image in 24bit "true color" is about 1.5Mbytes. Sure the jpeg is compressed but a bitmap is what is on screen. That's just a picture in a web page. I don't know about you, but I run my browsers at full screen, the smallest screen I run on is 1440x900, that's nearly 4MB of framebuffer. That's just the visual part, that doesn't do any processing or actually making sense of the data or running any javascript. The frame is a shared component it's not directly in the memory space of the browser but that gives you some idea of sizes of data. What's a browser do? It parses data and produces a picture of it, rendering fonts and what have you. I have not looked closely at the code to any of them but to make them fast, I'd render the whole page in to a buffer and then display the portions of the buffer that the user has selected and then, hopefully you'd be able to scroll fast and search and do all that stuff. So if you're web page is big enough to be like 10 "screens" (say your scroll bar is 1/10th the size) and the whole thing is rendered in memory and it happens to just be a raw picture during that time, you're looking at like 40MB of a page. Now this isn't the most efficient way of doing it but it's really easy to consume that kind of memory with an app like this.

I think I usually have about 10 tabs open on firefox too.
 

Re:And Opera

[fatphil]

Lynx isn't graphical, you want to compare it to w3m, which whilst text-based also renders images if running under X.

29490 phil 15 0 179m 87m 16m S 0.0 40.1 9:46.17 firefox-bin
  8146 root 5 -10 151m 20m 2636 S 0.3 9.1 77:22.35 Xorg
29527 phil 15 0 16616 12m 4532 S 0.0 5.8 0:06.72 emacs21
29682 phil 15 0 10952 8264 2472 T 0.0 3.7 0:00.51 w3m

I kill firefox every day due to its incessant leaking. (I know they've designed it it allocate memory which doesn't get freed, but however they word it, it's still a leak.)

Re:And Opera

[LizardKing]

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

Re:And Opera

[tepples]

Firefox doesn't even read mail citation needed

Re:And Opera

[AmaDaden]

As I understand it a lot of it is for speed. FF loves to forward and back cache pages. So it caches most of the pages you've been to and pages you might go to from where you are. Then you have the raw caches of all the images as has already been talked about. You also have the JavaScript engine. I'm sure I'm missing a lot of things too. It does have a lot to remember so I'm not surprised about the ram usage. By the way for those of you who have not kept up one of the big problems with FF2 was that the memory would start to get fragmented over time. So you ended up with large blocks of memory that contained only a few bits of information. FF2 did not seem to have a good way to handle this. But FF3 does. If you open up Task Manager you can watch it's ram in use drop after a heavy browsing session. A huge memory footprint is bad but a slow browser is just not worth it.

Re:And Opera

[calebt3]

There is one thing that has seriously inhibited me from looking seriously at Opera:
I can't figure out how to make a cookie whitelist (the "Exceptions" list in Firefox). The best I can find is Allowing/Denying cookies on a per-site basis.

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.

Re:And Opera

[forkazoo]

Is it just me or does it seem like 60MB or even 34MB is a LOT of memory for something that browses Web pages?

I mean, people used to make fun of GNU Emacs, saying things like it stands for eight megabytes and constantly swapping or eventually malloc()'s all computer storage. Emacs takes somewhere around 10MB or so on a RHEL4 box, and that thing is practically an operating system. It reads mail! Firefox doesn't even read mail, and it takes 60MB. Opera reads mail, but still 34MB seems just too big, too.

Maybe I'm just getting to be a cranky old man. Now you kids get offa my lawn!

I used to browse the web on a machine with 8 MB of RAM. Total, including the OS. At the time, real time decoding of a JPEG was extremely difficult, but my current CPU has 100 times the clock speed and is 64 bit and has vector processing features. Yet, browsers still seem to make the same class of CPU-memory tradeoffs that made sense on a 68030. For example, I may have ten tabs open in a window. I can only see one of them at any given moment, but the fully decoded images are all sitting in memory for all ten web pages, despite the fact that the page could be re-rendered almost instantly on a modern system.

Since browsing a few web pages is seldom the only thing I do with my computer, I go and do other stuff in Lightwave, Blender, Photoshop, whatever, then I come back to my web browser, and I wait while the whole working set gets swapped back in. Then, I click on the tab I want, and I wait while the working set for that tab gets swapped back in. If it just rerendered the page from the original bits, rather than using cached decoded images sucking up RAM and whatnot, it'd have almost nothing to reload and worst case performance would be orders of magnitude better. Hooray for "optimisation!"

Oh, and can we get some ninjas to fucking kill Flash. Seriously, I shouldn't need a bunch of script blocking and flash blocking extensions just to be able to browse the fucking intarwebs without having a seizure.

Re:And Opera

[evilviper]

You've got plugged-in interpreted flash code,

No, I don't have Flash at all, yet Moz is just as much of a resource hog.

And even then some scripting is safe, some is not, so there are rules that the code has to implement

I use NoScript, so no Javascript in-use here 99% of the time, either...

like pop-up blockers, password managers, warnings on insecure pages,

A pop-up blocker is a passive device, simply refusing to execute certain code, it saves CPU time, not the other way around. Similar for pointless warnings about page contents.

And again, I have the password manager disabled, so it should not be using any resources.

All that and the browsers STILL need to be able to sensibly parse and display completely borked pages with invalid HTML.

Point me to the "borked pages" code, and I'll be damn happy to remove it, if it will give a huge performance boost. No question.

Nevermind maintaining history, cache, cookies, referring pages, bookmarks.

Although it doesn't handle javascript (which I disable with Moz anyhow), Dillo does basically everything you've described, using a fraction as much memory.

Right now, with the same 3 tabs open, Dillo is using 1/5th the memory of FF2.0.0.9, and that's just for starters. The performance disparity is much larger... Much of the time, the next page will be loaded just about the instant I finish clicking on the link. It's like the difference between night and day.

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.

Re:And Opera

[truthsearch]

FF loves to forward and back cache pages. So it caches most of the pages you've been to and pages you might go to from where you are.

I don't mind if some back pages take a little longer to load, so I set the browser.cache.memory.capacity config setting to 8000. Memory usage dropped and I see no difference in performance.

Re:And Opera

[b4k3d b34nz]

Yeah, immediate history caching can take a lot of memory, but Opera does the same thing (better) for less memory usage, generally. My guess is that extensions and the beast that it takes to support that in FF is what sucks down most of the memory. It's interpreted, rather than compiled in, so of course it's going to use up both more processor time and memory.

I'm not a big fan of FF, but I'm glad (for its users' sakes) that they're working to fix these problems, rather than just saying "oh, it's not really a problem" like they used to.

Re:And Opera

[0xFCE2]

Not just lynx - 10 years ago you could run netscape on linux systems which had 4 MB total (+ some MB swap).
I have tried this with X running on a different machine connected via the serial port (115 kBit/s). This may explain the 4 minutes starting time needed by netscape - but it worked. Even java applets worked, though not for very long...
But these days I am willing to spent a lot of memory to have some decent layout - the blink-tag may be very memory-efficient, but CSS is slightly more powerful ;-)

Re:And Opera

[Drive42]

That's what stops me from using Opera. It simply doesn't have all of the functionality I need.
It reminds me of a mac. It's slick as heck, but the wee nerdy obscure apps just aren't as plentiful. Don't need page-zooming! Need cookie profile switchers, torbuttons, and slashdot extensions!

And FF3 is running like a charm, actually. Much faster. Most of the plugins I use work fine if I force the install.

Re:And Opera

[Entropius]

Yes, but it's equally inefficient for Firefox to have to swap this sort of thing in and out when the OS is under memory pressure.

What we really need is a mechanism for e.g. Firefox to use large amounts of memory to speed up page loading when there's plenty of memory, but to optimize for a small memory footprint when I've got ten zillion Gimp windows and Picasa open.

Why should Firefox behave in exactly the same way in two totally different situations?

Re:And Opera

[Anonymous Coward]

Let me put it in the other way.

The raw web data transfered for those pages couldn't be more than 10MB, why do they need 60MB?
They really should be caching the raw data, not uncompressed bitmaps of the pages already rendered.

One feature I like to see is a user set upper limit on the memory size for firefox. Letting it get whatever memory it want is like handing a credit card with no spending limits to your ex-wife.

Re:And Opera

[bunratty]

I think you're referring to caches. They are different from leaks in that caches are deliberately holding on to memory to improve performance. All browsers have caches, because performance would be unacceptable with no caching whatsoever. A leak is different in that it does not improve performance, but is only wasting memory. You can minimize the amount of memory Firefox uses for caching if you want.

Re:And Opera

[neurovish]

189100 4-22:59:07 /opt/opera/lib/opera/9.50-20071102.6/opera

189MB with swap after running for about 5 days with 28 tabs open

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
http://lxr.mozilla.org/seamonkey/source/content/html/document/src/nsHTMLContentSink.cpp

Re:And Opera

[Antique Geekmeister]

Almost all of this is used for advertising and unnecessary dancing bears. If you want to see the difference, use a good quality text-only browser such as Lynx and see what your bandwidth and memory usage are.

Re:And Opera

[emacs_abuser]

But Emacs doesn't display images

Consider yourself corrected. Emacs has been able to display images for quite some time now.

Re:And Opera

[migurski]

Oh, and can we get some ninjas to fucking kill Flash. Seriously, I shouldn't need a bunch of script blocking and flash blocking extensions just to be able to browse the fucking intarwebs without having a seizure.

Why not just remove the plug-in? Your browser may continue to work fine without it, or flashblock.

Re:And Opera

[tepples]

There's really no need for a web browser to use more than 100MB of memory. For people with limited RAM, it'd make more sense at that point to cache the contents to disk re-render them when they're needed. Isn't caching contents to disk what your PC's swap file is for?

Re:And Opera

[sumdumass]

I have seen this too. I used to suffer from crashes and stuff with wild memory usages all being blames on my hardware, addons or whatever. I installed noscript one day and haven't had a crash yet. I see the same memory improvements you mentioned to.

Perhaps there is an issue with how FF2 handles scripts that is more the problem then the scripts itself. I don't know. It is too bad that after Mozilla started hanging out with Microsoft that they decided firefox 3 would no longer support windows 98. I guess my hold out has been for nothing because I will have to eventually abandon FF on a few machines in favor of something I don't have to pay Microsoft for as well as remain compatible with some apps they run.

Re:And Opera

[morgan_greywolf]

MOD PARENT UP! This AC knows what he's talking about.

Re:And Opera

[Anonymous Coward]

Is it just me...
It's just you and everyone else who doesn't understand modern computing.

Re:And Opera

[sumdumass]

Hmm.. Well, I guess by some licenses and ideologies out there, a server reading the mail and presenting a HTML web page with the contents could be considered having fire fox reading email.

Re:And Opera

[forkazoo]

Why not just remove the plug-in? Your browser may continue to work fine without it, or flashblock.

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 and JS. Youtube is an obvious example, but I could live without it. OTOH, there are some places where I am considering applying for jobs that have flash web sites. I'm really okay with swearing off most sites that require flash, but it is much harder to refuse to deal with a website when you are trying to research a company and get their contact information for a really interesting position. (If for no other reason than if I did start working for them, I'd have internal influence for getting rid of the flash!) :)

Re:And Opera

[naasking]

Is it just me or does it seem like 60MB or even 34MB is a LOT of memory for something that browses Web pages?

But a browser doesn't just browse web pages. A browser is a limited form operating system, as it has an execution language (Javascript) and environment (the DOM). A mail client is relatively simple as it's just a texty protocol. A browser is HTML + XML + CSS + HTTP/S + JPG/PNG/GIF/etc renderers + embedded plugins + caches, and in case of FF, it has XPCOM and various other extensible subsystems.

Re:And Opera

[bberens]

I think you neglected to fill in one of the nouns on your Madlibs template above.

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

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:And Opera

[Lord Ender]

Is it just me or does it seem like 60MB or even 34MB is a LOT of memory for something that browses Web pages?

There are very few static "web pages" in use today. Web browsers are more often used to access web applications. The web browser is an application platform. It has its own scripting language (javascript), GUI API, and with the awesomeness that is xmlhttprequests, they are now highly-interactive applications.

Re:And Opera

[cheater512]

There are light browsers around. Ever tried Lynx?

You get what you devote memory to. :)

Re:And Opera

[random0xff]

You know what, your post made me think of something: Why can't I see the processes of plug-ins separately from the browser (and kill them too)?

Re:And Opera

[mollymoo]

Erm, by your definition doesn't every application ever written have the same "memory leak"? They all allocate (or get allocated) some memory which isn't freed till the application exits.

Re:And Opera

[ultranova]

A huge memory footprint is bad but a slow browser is just not worth it.

Well, seeing how Firefox leaks both memory and CPU time, it seems that we have the worst of both worlds.

Re:And Opera

[Myopic]

I'm not following you. Don't you actually have to go out of your way to install Flash? All you have to do to avoid Flash is... nothing. If your browser came with Flash, it should be easy enough to delete it.

Re:And Opera

[daniel_newton]

plugins (dlls) generally run in-process

Re:And Opera

[whitehatlurker]

"And Opera primarily for ."

You misspelled pornography. I don't care how poorly Paradora Flashes you, there's only so much an innocent browser can take. Umm, can you supply a URL so this can be tested?

Re:And Opera

[Blakey Rat]

Don't forget that the DOM model, even for a relatively simple webpage, has got to be at least a couple megs, and that's per-page/frame.

Re:And Opera

[tgv]

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.

Just pray that they're not going to throw Mozilla's resources to this project, because then Mozilla will perform just as well as Dillo...

That was ironic, but I do think that having many people work on a project can lead to "certain" design decisions under pressure (e.g. in an attempt to make it safe to run code that's not been thoroughly checked or to make the interface easier for new-comers) that in the beginning have a modest performance penalty, but in the end bog everything down and can almost not be removed without the whole system collapsing. Anyway, it has Mythical Man Month written all over it, I guess.

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.

Re:And Opera

[RebelWebmaster]

I think that overhaul is something being looked into for the Gecko 2.0 (Fx4) timeframe.

Re:And Opera

[msuarezalvarez]

Let me put it in the other way.

The raw web data transfered for those pages couldn't be more than 10MB, why do they need 60MB?

Do you think that DOM editing of the document tree, for example, should be implemented by actually editing the raw data gotten from the web? Or, for another example, when you double click on a web page, how do you think the word on which the clicked happened is found in the document? I guess you are imagining that the whole document is reparsed, the placement on screen for every single thing is recomputed, reflowed and so on, and then the word in the click coordinates is found? Have you not considered that web browsers developers decided an age ago to do some caching? Don't you imagine that that caching might take more memory that the raw data? Have you even considered the reasonability of what you are asking? How is it on earth that you see yourself as fit to make browser design suggestions?

They really should be caching the raw data, not uncompressed bitmaps of the pages already rendered.

Of course, I'd love to see how a browser designed according to your ideas has to uncompress every image on a web page each time the user moves the scrollbar a pixel down.

Re:And Opera

[JCSoRocks]

To add to what others have already posted - .jpg's suck up memory. Although they're small to begin with, the browser decompresses them and stores the decompressed version in memory. At least they used to a few years back, I'm assuming they still do.

Re:And Opera

[Anonymous Coward]

Right now, with the same 3 tabs open, Dillo is using 1/5th the memory of FF2.0.0.9, and that's just for starters.

oh go fuck yourself. you're comparing dillo, a browser that doesnt even attempt to handle css (slashdot looks pretty much incoherent in dillo), let alone any of the other niceties of a post-1990 browser to a modern fully featured browser. that's like comparing ed to openoffice and declaring ed the winner because of it's superior memory usage.

if you dont want modern features, that fine, but then dont go using a modern browser and whinge about how it's using more memory than browser X with minimal features. if you sat people down with a pc that only had dillo installed, 99.999% of people would ask you "why is you computer so fucking shit it cant even render the internets properly"

and to the mod who gave you insightful, you're a fucking moron and i bet you dont even know what dillo is

Re:And Opera

[nuzak]

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.

Windows has this feature already. But memadvising (using the unix term here) doesn't help when you're wasting memory in the first place. A lot of it is heap fragmentation, which suggests to me that the allocation routines are basically crap.

Re:And Opera

[Anonymous Coward]

"Yet, browsers still seem to make the same class of CPU-memory tradeoffs that made sense on a 68030. For example, I may have ten tabs open in a window. I can only see one of them at any given moment, but the fully decoded images are all sitting in memory for all ten web pages, despite the fact that the page could be re-rendered almost instantly on a modern system."

Actually, they aren't, at least not in Firefox 3. Images that aren't in use for the current tab are flushed. However, this might be really bad for memory fragmentation, which is part of why Firefox gets the bad wrap it does today with "memory leaks".

Sadly, people keep coming up with this shit, when the code's right there. Read it, find a problem, fix it. If you can't find a problem, it's likely there isn't one in Firefox's code, and instead your Operating System's (in particular, Windows XP and it's terrible heap allocation code, which ironically was much improved in Vista but you can't tell due to all of the other crap that's been piled on top of it).

Re:And Opera

[Allador]

If the memory use is because of caches then something is wrong with their implementation, because they grow endlessly.

After a recent thread on this (which I think you and I discussed) I put in some of those settings to reduce FF's memory usage, and it did help, but it still grows endlessly, now just slower.

There's a linear relationship between how long its been running and how big the memory is.

For example, right now FF2 has been running for about a week on my box, and has 17 tabs open. The tabs have been fluctuating constantly over the week.

It's using a physical memory working set of 182MB, and a virtual memory size of 410mb.

Note that this is _massively_ improved since I made some of those changes. Prior to those in a similar situation, it would probably be over 500mb of working set.

Re:And Opera

[belmolis]

Dillo is nice, but the lack of Unicode support really limits its usefulness for me and many other users. I still use ed for some tasks, so I appreciate a fast, simple, approach.

Re:And Opera

[whereiswaldo]

I don't understand why Firefox doesn't include a "disable flash" feature right in the browser, just like it allows you to disable javascript and java applets. Why is Flash treated differently?

Re:And Opera

[MadMidnightBomber]

But Emacs doesn't display images (somebody will probably correct me on this).

Rule #94 - just because something is a very bad idea, doesn't mean there isn't an implementation of it for GNU Emacs.

Re:And Opera

[Anonymous Coward]

BBC via Dillo - http://www.clanjones.org/bbc-via-dillo.jpg
BBC via Firefox 2 - http://www.clanjones.org/bbc-via-firefox.jpg

Firefox wins.

Re:And Opera

[fatphil]

Congratulations, you've bought the lies, I hope you didn't pay too much for them.

Anything which takes memory from the system and never returns it is not a cache. Anything which keeps everything indiscriminately is not a cache. They may have intended it to be a cache, but the bodgers behind it are evidently incompetent, and not worthy of the mantle 'programmer'.

And why on earth would I be interested in the browser configuration that I've already done and which appears to be able to return 34MB to me if firefox has already leaked several hundred megabytes, and continues to leak more with every single new page that I visit.

Re:And Opera

[fatphil]

Nope. It had already been stated that the growth was without bound.

Re:And Opera

[bunratty]

Memory use is a combination of necessary memory use, caches, memory fragmentation, and memory leaks. What you're describing sounds like memory leaks, not caching. The good news is that in Firefox 3, Mozilla developers' "extensive testing shows an occasional leak here and there and we are working to fix those, but in general we aren't seeing many leaks anymore." If you see any way to reproduce any memory problem in Firefox 3, please report it with the set of steps used to cause the problem.

Re:And Opera

[ThirdPrize]

What the world really need is a text/command line web browser. Something that looks and feels a bit like "Word star/perfect". Rather than clicking on hypelinks you type in your destination into a command line and it converts all images into ascii art. Maybe even colour as well.

Re:And Opera

[ESqVIP]

And even then some scripting is safe, some is not, so there are rules that the code has to implement

I use NoScript, so no Javascript in-use here 99% of the time, either...

That doesn't make the JavaScript engine disappear, and the DOM is still generated for every document. And you can't really strip those at all, since both are a core part of XUL. The point is, disabling JavaScript won't really reduce memory usage on Firefox, and probably on any other browser. I don't think Opera or any other browser will switch to some alternative renderer that does not use the DOM or the box model when you disable JavaScript or CSS.

like pop-up blockers, password managers, warnings on insecure pages,

A pop-up blocker is a passive device, simply refusing to execute certain code, it saves CPU time, not the other way around. Similar for pointless warnings about page contents.

And again, I have the password manager disabled, so it should not be using any resources.

There's certainly a bit more to a popup blocker than simply one conditional. Some context on the page must be stored to know when a popup is authorized or not (i.e., certain events give temporary permission for websites to open popups), but certainly it shouldn't really be a noticeable difference anyway. But if it does bother you I'm pretty sure those features are mostly JavaScript XUL bindings (if I recall, Fx2's password manager is written in C++, but Fx3's is written in JS) and should be fairly easy to remove. Easier than your promise of removing parts of the HTML parser.

Re:And Opera

[bunratty]

It sounds like you've got a serious problem with your configuration if Firefox "continues to leak more with every single new page that I visit." Try the steps I posted earlier: Download Firefox 3 beta 1 and create a new profile. If you can then reproduce any memory problem, post the steps to reproduce it so the problem can be fixed.

Re:And Opera

[fatphil]

My claim, and the claim before mine, is that Firefox 2 leaks.
The counterclaim by the firefox fanboys, including most involved with development at the time, is that it doesn't leak.

Firefox 3 is irrelevant to that argument.

However, if behaviour has changed in Firefox 3, then it kind of supports the original claim about Firefox 2, no?

Hmmm. 2MB just got gobbled up pulling up the list of my messages, and loading the Post Comment page. I'll never see that 2MB again until Firefox is killed...

Re:And Opera

[bunratty]

I agree with you that Firefox 2 does leak memory. Mozilla developers agree that Firefox 2 leaks memory. I don't see anyone claiming that Firefox 2 doesn't leak. If anyone does, they're obviously incorrect. Do you think you're saying something controversial? It seems like you're stating something that every reasonable person agrees with. There's no reason to start an argument, as everyone agrees already.

Developers are now working on Firefox 3. Do you see any major problems with that version? If so, create a new profile and give steps to reproduce the problem. If you cannot, I don't see the point in discussing the matter further, as the problem would seem to be resolved.

Re:And Opera

[fatphil]

"I don't see anyone claiming that Firefox 2 doesn't leak."

What planet have you been on?

People have been calling the behaviour a feature and claiming it's not a leak for yonks. A 2 minute web search for ``firefox doesn't leak memory'' finds these immediately, for example:

http://internetducttape.com/2006/12/02/how-to-fix-the-firefox-memory-leak-firefox-hack/
http://blogs.chron.com/techblog/archives/2006/02/firefoxs_memory.html
http://weblogs.mozillazine.org/ben/archives/009749.html

Re:And Opera

[bunratty]

What you're describing is a common misconception of Ben Goodger's blog about the bfcache. Let me give you a few snippets from that post: "All versions of Firefox no doubt leak memory" and "What I think many people are talking about however with Firefox 1.5 is not really a memory leak at all. It is in fact a feature."

So, yes Firefox does leak memory. On the other hand, some people see the extra memory used by caching and mistake it for a memory leak. Some people see the memory wasted due to memory fragmentation and mistake it for a memory leak. And there are actual memory leaks also.

I'll repeat it again so it might sink in this time. Mozilla developers have never denied that there are leaks. If anyone says that Firefox doesn't leak memory, they're obviously incorrect. Go argue with the people who are stupidly claiming that Firefox doesn't leak memory. Don't argue with me, as I agree with you completely.

Re:And Opera

[bunratty]

If Firefox's memory grows without bound, that is a memory leak. Firefox's caches have bounds on them, and will not grow without limit. It looks like you're seeing memory leaks and thinking that the caches have been designed stupidly so that they grow without limit. Let me repeat: if you see Firefox's memory usage increasing without limit, that is a bug that should be fixed. File a bug report with the steps to reproduce the problem so it can be fixed.

Re:And Opera

[fatphil]

You don't agree with me completely as you think that developers have never denied that there are leaks. They have.

Re:And Opera

[bunratty]

Where have Mozilla developers denied that there are leaks? I've never seen such a statement from them.

Re:And Opera

[fatphil]

After scores of firefox related stories on /., and dozens of complaints about the leakage, some of which are gainsaid, some of which are countered with links to mozilla's bugzilla page, I'm 100% sure I've encountered sayings like "Not releasing memory is a feature, not a bug". Quite which bug report, I haven't a clue, there are hundreds of thousands of firefox bug reports.

Incidentally, with nothing to do with this discussion at all (just doing some research for an article), I discovered a perfect example of a killer memory leak: http://www.blacksmithinstitute.org/ten.php with javascript blocked, there's an iframe with a meta refresh that reloads itself. Bye bye RAM - one way, no returns. If you do have FF3, then it might be worth seeing what it's behaviour is. I'm not installing FF3 until it becomes part of the distribution I run, which may be many months.

Re:And Opera

[bunratty]

There are many thousands of Bugzilla accounts. Anyone can sign up. Just because you saw something in Bugzilla doesn't mean a developer said it. Most comments are not from developers, especially in bug reports with many comments about a hotly debated topic. It also depends on the context; in some bug reports, the reporter may be mistaking caching for a memory leak. In that context, not releasing memory is a feature, not a bug. That doesn't mean they're saying that Firefox has absolutely no memory leaks whatsoever. Sorry for the confusion.

I've had your perfect example of a killer memory leak open for over two hours in Firefox 3 beta 1 on Windows XP. You weren't kidding when you said it reloads itself; it's constantly downloading and rendering the iframe. Memory started (of course, after a full day of using Firefox) at 91 MB Mem usage and 84 MB VM Size. The memory use is now still at 91 MB Mem usage and 84 MB VM Size. You can file a bug report on Firefox 2 if you like.

Re:And Opera

[fatphil]

Yes, it's hard to know who's actually a developer, and who's simply interested enough in the development to enter discussions on the buglists. Typical average user isn't botherred, I'm sure. (I'm not, 99% of the time to be honest.)

Thank you for the test in FF3. I did send a 'this page doesn't render correctly' feedback report to the FF crew but if it's fixed in 3, then let's wait for 3. Of course, the 'bug' is google's asking for a redirect to the very same page, that's just inane. I raised it as a bug with the author of NoScript too, as I had 'forbid meta refresh redirects in noscript elements' enabled at the time.

Re:And Opera

[tzot]

Because, "ubiquitous" as Flash may be, it's not part of "standard" web browsing.

JS is managed by Firefox code, therefore there is a built-in setting whether to enable it or not.

Flash is managed by an extension; it's the Flash extension's (or another extension's, as in the case of Flashblock) job to offer a toggle setting.

I thought it was obvious, but, well, obviously it wasn't. Just imagine a freshly installed Firefox without any Flash plugins installed; the power-user wannabe looks at the settings, checks the "Enable Flash" checkbox, and then writes a post in their blog complaining about the useless setting in Firefox ("I enabled Flash in the Settings, but still could not see any Flash content!!1")

Yeah but it's still beta

[jayhawk88]

I'm sure that low memory usage bug will be fixed by the first release candidate.

Re:Yeah but it's still beta

[diegocgteleline.es]

The things is, there're real efforts to reduce memory fragmentation (which allows to free memory) that have been developed recently and are not included in this beta.

Re:Yeah but it's still beta

[beh]

Actually, that 'low memory bug' has already been fixed - I've downloaded the beta and installed it on WinXP - after looking at 2 pages, Firefox 2 memory usage was at about 45MB; Firefox 3's memory usage was up to about 750MB after less than 5 minutes (and the same 2 pages; in two tabs, just as with Firefox 2.0.0.9), completely bringing the machine to a crawl (1GB mem; and apart from Firefox, Outlook, Eclipse and SquirrelSQL were open)...

I'm reverting back to Firefox 2 for the time being, and will file a bug report once I have some more time to find out what's causing the issue...

Re:Yeah but it's still beta

[Anonymous Coward]

I'm reverting back to Firefox 2 for the time being, and will file a bug report once I have some more time to find out what's causing the issue... Good thing Firefox 3 has a fast uninstall. I might switch to it just for that reason.

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:Yeah but it's still beta

[unfunk]

Wow, something's definitely very wrong there; I currently have fifteen tabs open in FF3b1, and it's "only" using 230MB...

Mind you, back when I was a lad, Netscape 4.0 ran fine on Win95 with 24MB RAM total...

Re:Yeah but it's still beta

[jnaujok]

Really? Because I can do it in one page. I tried out FF3 yesterday. Opened our local search page (about a 4K page with very little other than text) and let it sit for 2-3 minutes while I worked on something else. Suddenly the machine slowed to a crawl, the drive went to 100% on, and by the time I could finally get Task Manager open a minute or two later, Firefox.exe was at 635MB and climbing.

I can reproduce it every single time.

Re:Yeah but it's still beta

[Drive42]

And what was the content of the pages you were loading? Plain-text HOWTOs? Straight HTML webpages where the most demanding element was a line of animated gifs and the dreaded blink tag?

Re:Yeah but it's still beta

[Anthracks]

Then file a bug and attach the page as a testcase.

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:Yeah but it's still beta

[jnaujok]

I had the same problem, and ran a "test". Basically, after you go to any page with a form field on it, Firefox goes out and tries to update the "Phishing" list. About 2.5 hours later (52 minutes of CPU on a 2.4GHz machine) and a peak of 450MB of memory for Firefox, it finishes. After that, this problem doesn't occur.

It's already been multiply posted to bugzilla.

Re:Yeah but it's still beta

[owlstead]

Which pages? Have you looked at the DOM (document object tree)? You can do so with the developer extension. It was mentioned that there was some faster dynamic HTML as well. That's all nice and dandy, but if the page keeps adding to the DOM in a Java Script loop, this means that the DOM of that page will grow to immense proportions. Maybe we'll need some kind of monitoring device for misbehaving pages, but as it stands, pages can use any amount of memory and CPU cycles, in any browser. And of course, due to browser differences it may be that some pages react differently to pages.

Fortunately I saw that there is an option to safe your pages (without firefox doing so automatically because it thinks it has "crashed" when closed by shutdown), so it is easier to close the thing when I want my CPU to be fully available (playing games, in other words). The memory usage and CPU usage of browsers during games can really irritate the hell out of me, just like the video jitters when an email arrives during a game.

Re:Yeah but it's still beta

[operagost]

The same thing happened to me. I had two tabs open and I wasn't even using it when I felt the system slowing down and heard the hard disk grinding. 1.1 GB of RAM in the VM.

Re:Yeah but it's still beta

[RebelWebmaster]

For more information, see this blog post from Asa Dotzler.

Re:Yeah but it's still beta

[pravuil]

Yeah, I'm running 2.0.0.9 on Fedora Core 8 atm and it's running at 65k MB. It's been on for the past couple of hours and I've browsed hundreds of pages within that time. I have several tabs open and it's still at 65k MB. No tweaks with about:config, just the default configuration. I've toyed around with 3.0b and it's pretty much the same. Still has some bugs to it like rendering PNG format correctly but the memory is pretty much consistent.

Re:Yeah but it's still beta

[Anonymous Coward]

Really? Because I have different memories of web browsing in the same era. If I came across a site with particularly heavily compressed jpegs, winamp would stall for a second or two!

Re:Yeah but it's still beta

[Anonymous Coward]

That's why they call it a leaked beta...

Re:Yeah but it's still beta

[cathodion]

yeah i had the same thing... had to kill Firefox to regain control over my PC (WinXP sp2). Some serious memory leak if you ask me.

Re:Yeah but it's still beta

[Anonymous Coward]

A friend of mine tried FF3 for Linux. He opened it and just let it sit there - didn't even load a page. Within a few minutes, it was using 40% CPU and 90% of installed RAM - on a machine with 2GB RAM! There's reasons why it's still in beta.

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:How are they measuring?

[vally_manea]

From what I've heard this is common practice at MS. When the app gets minimized it releases/caches a lot of memory. There was a story on a MS person's blog but I'm to lazy to find it out.

Re:How are they measuring?

[Lemmingue]

The Working Set (physical memory) size will drop, but the memory consumption (Private Bytes, Virtual Memory) will be the same. When a window is minimized, Windows mark the memory pages as candidates to be relocated in case of memory shortage. When you restore IE focus the Working Set size will return to the previous size.

Task Manager sucks, use Sysinternals' Process Explorer.

Re:How are they measuring?

[wile_e_wonka]

I don't believe this is only IE--when I listen to music on my computer on an airplane, I minimize the music program because it uses less memory, and therefore less battery (from my tests, Winamp Lite minimized uses the least RAM when minimized of all the players I tested. It was better than Foobar and mplayer, etc).

In fact, I just tried the same thing with Opera--it dropped from 60,000 to 11,000.

I don't think it's an estimate--I think the program really uses less RAM when minimized.

Re:How are they measuring?

[MyLongNickName]

Do this experiment.
1) Browse with IE for a few minutes.
2) Look at the memory usage
3) Minimize IE.
4) Look at the memory usage
5) Maximize IE
6) Look at the memory usage.

Compare #2 and #6. I've found that #6 is lower than #2.

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.

Re:How are they measuring?

[BlowHole666]

It probably uses less memory because when the window is minimized Windows does not have to spend the time to update the GUI. So all of the GUI data can be taken out of memory. So the actions of the GUI are still taking place but the memory holding the redrawn objects etc is not needed. It should also execute faster due to not have to refresh the screen.

Re:How are they measuring?

[nxtw]

How does minimizing a program / using less active memory cause less power draw? I would imagine that the power you save would be from not having to render the visualizations, resulting in less CPU work. Memory chips aren't going to be shut off or anything like that.

Re:How are they measuring?

[jandrese]

That behavior drives me crazy about Windows, so much so that I will often resort to forcing the page file down to 2MB just to keep it from swapping my applications out when I Minimize them. I hate having a system with 2GB of memory and having to wait 30 seconds for it to page some application back in (slow laptop HDDs don't help) just because it thought I might want a lot of free memory for some reason.

Re:How are they measuring?

[empaler]

Embrace, extend, extinguish

Re:How are they measuring?

[astrosmash]

So, what's your analysis? Why is #6 less than #2, and what does than mean?

Where does the memory go when an application is minimized? Where does the memory come back from when you restore the application?

Hint: the application isn't doing anything. It's an OS trick.

Bonus Question: Why do you think Firefox disables this trick?

Re:How are they measuring?

[MyLongNickName]

My assumption is that that is related to managed memory. An object is marked as being ready for disposal, but is not actually disposed. When a minimize event occurs, garbage collection occurs.

Now, I could really be off on this, but I have not seen anyone give a reason for this behavior that explains everything I see. If you have a better conclusion, I am all ears :)

Re:How are they measuring?

[Threni]

Do this experiment.
1) Buy enough ram so you now have 4GB.
2) Just use whichever damn browser you like.
3) Go outside and get on with it (your life, that is).

Observe which procedure (this, or that detailed in the OP) brings more joy to your life. Frankly I don't give a shit if Firefix uses a few hundred meg, because I earn enough to buy a few hundred megs in the time it would take me to do the original procedure.

Re:How are they measuring?

[TheSpoom]

Uh... They're not exactly extinguishing it if they're allowing free, unencumbered download of the product. They just bought SysInternals a while back so they distribute their stuff now.

Re:How are they measuring?

[fatphil]

Extra bonus question:

If what is happening is an OS trick, and firefox is a userspace application that disables this trick, then why is a user-space application interfering with the running of the operating system?

Re:How are they measuring?

[quanticle]

To echo a previous reply, the reason #6 is less than #2 is that when the application is swapped to disk and swapped back, the OS doesn't fully restore all objects that the application had in memory. It only restores the objects that the application is using, and doesn't restore the rest until the application requests some swapped objects. Given that most browsers are coded in C++, rather than Java or C#, garbage collection is a non-issue.

Re:How are they measuring?

[MyLongNickName]

3) Go outside and get on with it (your life, that is).

Perhaps you should do that instead of replying to posts you don't give a shit about?

Re:How are they measuring?

[justthinkit]

When I suspect my browser (Opera) is using too much RAM I restore Task Manager, note the total memory usage (at this point I could care less about how much it says Opera.exe is using). Then I close Opera and wait. And wait. And wait. After a minute or two I get a figure that is usually 2 to 3 times what Opera.exe is/was taking. In my mind this is the figure that should be used for each browser.

Re:How are they measuring?

[Ctrl-Z]

Given that most browsers are coded in C++, rather than Java or C#, garbage collection is a non-issue.

Here's where you lost me. Are you saying that there aren't garbage collectors for C++?

Re:How are they measuring?

[wile_e_wonka]

I'm not sure it would draw less power, but I was trying to think of every possible thing to save battery for the ride. My thinking was that RAM chips require power to retain memory, so remembering more stuff might require more power.

I also shut off every unused and unnecessary program and service, whether or not it appeared they were using any CPU. I also dimmed the screen all they way and had it turn itself off after three minutes of nonuse (not surprisingly, this seemed to do more to save power than anything else)--I don't have a computer that lets me play music with the lid closed.

Re:How are they measuring?

[Rallion]

Why don't you just disable the page file altogether?

Re:How are they measuring?

[Threni]

> Perhaps you should do that instead of replying to posts you don't give a shit about?

I'm taking a break from work, by browsing the web using Firefox 2.

Do you think more people would use Firefox 2 if it didn't use so much memory? How many more people? How much time is the problem worth?

Re:How are they measuring?

[empaler]

It was mostly a joke while pointing out that it wasn't Sysinternals any more, but free downloads can be a part of the 'embrace' and 'extend' parts of the process. ActiveX is a free part of IE, for instance.
Later on they might make it a "full-featured separate product" ($$$) or slap WGA on it.

Re:How are they measuring?

[Anonymous Coward]

My thinking was that RAM chips require power to retain memory, so remembering more stuff might require more power.

You're on track with the other stuff (especially about the screen) but this was off. Yes, RAM chips require power to function, but they are always on, even if that part of the memory is "empty". (Actually, with power-hungry FB-DIMMs it might be good to be able to disable one of two sticks at will; but most laptops have just one unbuffered SO-DIMM anyway. One powersaving measure in the ingenious Pentium-M was its ability to turn unneeded portions of its big L2 cache off on-the-fly.)

However, in a way you did the right thing, though. Minimising memory usage helps prevent the OS from needing swap (and waking up the disk) and that saves some power.

Re:How are they measuring?

[bunratty]

Because users complained that Firefox was too slow after un-minimizing.

Re:How are they measuring?

[tepples]

If what is happening is an OS trick, and firefox is a userspace application that disables this trick, then why is a user-space application interfering with the running of the operating system? The app is not interfering. It's using an API call that when written one way says "use the trick less aggressively" and when written the other way says "use the trick more aggressively".

Re:How are they measuring?

[jafac]

Well - Microsoft may be distributing Sysinternals products; but Marc Russinovich's Blog postings are becoming less and less frequent.

AND - Sysinternals used to distribute the source on some of their tools. No longer. It's out there. But it's not legal.

Re:How are they measuring?

[jafac]

Process Explorer; no workee on Vista - must use Process Monitor - which is GREAT if you want to monitor file and registry IO, not so great as a replacement for the kinds of things Process Explorer did (A Task Manager replacement).

I'm also not a big fan of the logfile format for Process Monitor. I liked the old csv format. In any case - for doing this kind of process monitoring stuff in Vista, I've resorted to writing my own vbs/WMI scripts to get this (performance) data, because it's just not coming out of Process Monitor in a useful format.

The roughly equivalent task in Pet Ownership is like having to do your own canine proctology (without rubber gloves) because Purina bought all the veterinarians to make them work on test animals at the food plant. It's nice that Purina ships free flashlights with the dogfood though.

Re:How are they measuring?

[Anonymous Coward]

That behavior drives me crazy about Windows, so much so that I will often resort to forcing the page file down to 2MB just to keep it from swapping my applications out when I Minimize them. I hate having a system with 2GB of memory and having to wait 30 seconds for it to page some application back in (slow laptop HDDs don't help) just because it thought I might want a lot of free memory for some reason. Why not just run a quality operating system? If you're running 2gb of ram you likely have a decent processor. Try running Ubuntu as your OS and put Windows inside of VirtualBox (gpl'd and included in Gutsy).

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:How are they measuring?

[dakameleon]

It may be an estimate, but since they're all estimated by the same tool, it's a roughly valid comparison. Relativity is the point here I suppose, as opposed to absolute lower memory usage.

Re:How are they measuring?

[jandrese]

One does not have a choice about the operating systems much of the time. I run FreeBSD on my home machine, but my work makes me jump through hoops if I want to run anything other than their "trusted" XP install on a machine. It's not worth my effort, plus they've been seduced into running Exchange, so I have to deal with the bloated and slow piece of crap that is Outlook.

Re:How are they measuring?

[jandrese]

Because then Windows gives you grief about turning it off every time you boot the machine.

Re:How are they measuring?

[w_crossman]

Are you talking about Process Explorer? The page says that it works with Vista.

Re:How are they measuring?

[Lemmingue]

Nope. Firefox is a C++ application using Mozilla's XPCOM. XPCOM uses a reference count scheme to control memory lifetime.

Re:How are they measuring?

[Maian]

No, it's that they don't always use them. It isn't trivial to just plug in a GC for an existing project. Besides changing your memory API to utilize the GC, there can be subtle bugs that result after the switch, which costs QA time and money. And even then, a GC isn't always the best solution.

Why? Because there are performance downsides to using them. Many C++ project opt to use reference counting instead, and manually make sure there aren't any cyclical references. I think the Gecko engine uses reference counting for most things, and only uses a GC for its JavaScript engine. I could be wrong though. Microsoft's COM model is all based on reference counting. I think its newer .NET framework uses a GC though.

FYI, I use that GC you linked for many of my C++ projects. It works pretty well for my purposes, but my projects are all relatively small.

Re:How are they measuring?

[Anonymous Coward]

I bet it goes up if you minimize Firefox...

Re:How are they measuring?

[jesser]

Operating systems use large amounts of memory for two things: application memory and caching reads from the hard drive. Normally, this is a good thing, as it speeds up file access using RAM that otherwise wasn't being used at all or wasn't being used actively. But sometimes it isn't very smart; for example, on my Mac, grepping through a few thousand 1MB log files can result in applications being paged out.

Forcing the page file down to 2MB sounds like a kludgy way to keep your application memory in physical RAM. For one thing, it means that if you try to run too many applications at once, you'll hit out-of-memory errors (which often result in application crashes). Perhaps there's a way to limit the page cache or tell your operating system to prefer using physical RAM for application memory rather than the page cache?

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).

Re:How are they measuring?

[mattgreen]

Process Explorer works fine on Vista; I use it all the time.

Re:How are they measuring?

[jandrese]

I wish there were a way to tell Windows not to page out your applications. I've looked and the only thing I've found is a setting that is already set to "prefer applications".

Re:How are they measuring?

[quanticle]

I'm not saying that there aren't garbage collectors for C++, I'm saying that its not likely that any project of significant size will use such a garbage collector, mainly due to the performance penalty over using non-garbage collected code.

In other words, if Mozilla wants a GC for Firefox code, why don't they write the thing in Java or C#? It'll probably as easy to rewrite the app in a language designed around a GC as it will be to put in a C++ garbage collector.

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

[Red Flayer]

Firefox 3.0 b 1:

Loading a five pages into the browser - 38,644KB
Loading a single page and leaving the browser for 10 minutes - 63,764KB
Loading 12 pages into the browser and wait 5 minutes - 62,312KB

I wonder what would have happened had he loaded 12 pages and let the browser sit for 10 minutes -- would the memory usage still be less than the single page/10 mins test?

Seems to me that memory usage must still spiral under 3 beta, otherwise how would the single page/10 min usage be less than the 12pp/5 min test? Sure, it's not as bad, but that number really caught my eye... more testing is in order if I can get some time away from the in-laws over the holiday.

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

[CastrTroy]

Also, seems like a pretty crappy test to me, especially considering that most of the complaints with Firefox are with memory leaks, and not memory usage from opening a few pages. What happens after an entire work day of using the browser? Is there a significant difference in memory at that point? People who open their browser and look at 10 pages, then close it again will rarely ever have a problem with memory usage in Firefox. However, those of use who leave it open for days at a time, doing web development, and constantly looking at new pages are the ones who need to worry. It's like comparing the performance Visual Studio 2003 to Visual Studio 2005 on a project that only has 5 classes.

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

[What'sInAName]

Indeed. Personally, now that Firefox will save tabs, I no longer have much concern about memory leaks. Of course it would be good if they were fixed, but practically speaking, it doesn't have a material impact on my usage. I always close Firefox at the end of the day and reopen it in the morning.

However, if they can make Firefox even quicker, that would be something I'd appreciate!

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 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.

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

[ILongForDarkness]

Kind of like MS, they could release a "perfect" OS, and the internet would still be full of blogs of people bitching about Windows. People remember the bad and forget the good. If there computer needs a reboot once a month, that is want they remember, not the 29 days were they had 50 programs open simultaneously and still had a responsive GUI. This is where IMHO the software industry needs to change, we are always in such a rush to add new features that we risk releasing things too early.

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

[bunratty]

You can find the results of a more extensive memory test here.

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

[rrkap]

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.

Well, a good start would be fixing the high memory consumption.

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

[bunratty]

Here we go again... could you explain to me how I could see "high memory consumption" in Firefox, significantly higher than other browsers use? Please give a detailed set of steps for how I could reproduce the problem on other computers. If you do, I'll file a bug report so the problem can be fixed.

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

[Deliveranc3]

Be better than Opera.

You won't be able to brag until you do, plus Opera users are very open minded. When firefox gets better than Opera our 3% marketshare will jump on board!

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

[Rufty]

That's the problem, Microsoft couldn't release a perfect OS if you paid them.

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

[ILongForDarkness]

But I do pay them, and still I don't have my perfect OS :)

That said Linux, MAC OS, or UNIX isn't perfect either. One might be best for one type of use, another for a different one etc. The problem is users want to do a variety of things on their workstation, so it is a pain to have an architecture/scheduler/GUI that works for the whole gambit. I think Windows comes the closest for an average user, probably Linux or UNIX for the techi that likes to hack at a lower level. By definition the people with expertise are a minority, so a UNIX like OS has a long way to go before it will dominate the desktop. Easier for server land, because by definition, server admins are techies.

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

[AVee]

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. Let's see if I can burn some karma here...
It's not that hard to fix, they should simply drop the current code and use WebKit instead. And then convince the world they really did change something to the engine this time, instead of just changing the name, or the UI, or the pug-in system. But they indeed lost a lot of my trust when they wrapper a new UI around an existing engine, called it firefox instead of mozilla and claimed it became really fast. Especially because they did it not all that long after MS dis the same thing with Windows XP.

Really, Firefox is progress only if your previous browser was IE, of course this goes for a lot a people, hence it's succes. But if you are used to Opera, Konqueror or Safari it's a step back on most aspects, especially speed and standards compliance. I will try Firefox 3 when it's done, but it has quite a but of catching up to do and I doubt they will make it. Besides that, it is a pointless exercise in extending the life of an old codebase, if the same amount of energy and money invested in Gecko would have been spend on writing a totally new engine they would have had a far better browser by now. Netscape did not open-source the thing because it was so very good, but because it had became a real PITA.

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

[BZ]

> otherwise how would the single page/10 min usage be less than the 12pp/5 min test?

Is the single page one of the pages loaded in the 12-page test? The article doesn't say.

This article is a great example of how to write up your experimental results without allowing anyone to ever verify them, in fact. ;)

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

[BZ]

> It's not that hard to fix, they should simply drop the current code and use WebKit instead.

It's been considered. That would be a much larger undertaking than you think, though.

> Netscape did not open-source the thing because it was so very good

Sure. Of course the "thing" Netscape open-sourced wasn't Gecko. Gecko was written after people looked at the thing Netscape open-sourced and decided that it sucked.

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

[rrkap]

Firefox has two memory consumption problems. Right now I'm using firefox 3.0b1. With this comment open, it is consuming 67,928K vs. IE 7 at 41,924K, so it has significantly higher memory usage during normal operation. However, the real problem is difficult to reproduce. If I leave Firefox 2 (which is what I usually use) on overnight, about one time in 10, it will be consuming 500+ megs of memory and half the CPU when I get back into work in the morning. I don't have the time to track down this problem in detail so that you can reproduce the problem, but it's definitely there and is frequent enough to be annoying.

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

[ErikTheRed]

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.
...
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.

A large part of their problem is the vehement denials issued by the Firefox team regarding the memory leaks over the years. We all knew they were completely full of shit. I verified this myself - I tried uninstalling, deleting my profile and any left-over program files, reinstalling, and loading one of my favorite pages - Nagios Service Detail view. This is a reasonably long page with very few graphics that auto-refreshes itself every few minutes. After 24 hours, I'd have Firefox using about 1GB of RAM.

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

[flakier]

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.

It would be quite easy to flip the negative perception IMO. When they are sure it is all fixed (you can use the browser heavily for days w/o restarting etc...) then release the fixed version. Now, instead of the release notes consisting of "Security and stability fixes" there would be a nice banner saying something a tad less nebulous like, say "WOWZAAZZZ, lookee here we've finally gotten around to fixing that big RAM wasting issue everyones been bitching about for years! WOOHOO Download now and tell all yer friends!"

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

[bunratty]

First problem: When I start Firefox 3 beta 1 and IE 7 and do nothing but go to your comment to reply, IE 7 takes 56 MB Mem Usage and 43 MB VM Size, but Firefox takes only 44 MB Mem Usage and 34 MB VM Size. So I can't reproduce that one.

I cannot reproduce the second one because you haven't given instructions for how to do so, and I never see it on my own.

I'm not saying that the problem is not there or is not annoying. I'm simply saying that to get the problem fixed, someone needs to file a bug report. To file a bug report, we need a set of steps to reproduce the problem. That's where we keep getting stuck, because I can never reproduce the memory problem, no matter what I do. If you still think there's a problem that needs to be fixed, all you need to do is point out what it is, specifically, by giving a set of steps that will allow others to reproduce it on their computers. Until someone gives such a set of steps, you cannot expect the problem to be fixed.

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

[bunratty]

There have been no "vehement denials issued by the Firefox team regarding the memory leaks," unless by that phrase you mean "politely asking for how to reproduce memory leaks." If you have a way that others can reproduce the problem, please file a bug report so the bug can be fixed!

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

[Anonymous Coward]

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).

(Asbestos suit on.)

It's a C++ app. The C++ memory model is just too complex: new/delete, copy constructor, overloaded assignment operators ... they'll be finding memory leaks in this until the day it dies. Even C++ programmers I know don't claim to be able to write perfect C++ code, and this is one of the most complex programs I've ever used. It makes Linux look like hello-world.

I find it interesting that C++ never really caught on in the free software world like it did in the proprietary software world. The biggest examples of free C++ we have (Firefox, Qt) descend directly from code which was once proprietary (Netscape, Qt).

Now, I suppose by not having written a better web browser I'm also part of the problem. But I wonder: if all the people with garbage-collected functional languages think they're really good for solving complex problems in fewer lines of code with no leaks despite hard-to-determine object lifetimes, why isn't there an awesome web browser written in SML or Lisp or Erlang or IO by 5 guys over the weekend that's 10 times faster and smaller and more extensible than Firefox?

True, if the Mozilla guys had started with a garbage-collected language (since it seemed like they basically started over in 1998 anyway), they could have spent all this time tuning their GC to make performance awesome, instead of fixing bugs for correctness. But now that they've sunk the time into it, assuming you don't want to ever look at the source code, isn't it quite possible that Firefox is a pretty good performer?

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

[Maian]

Just like Mozilla couldn't release a browser that fixes its memory reputation? *sigh* Prejudices are bad, my friend.

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

[bunratty]

The catch is that there is no "it" to fix. There are lots of little problems, and no way to know when all the problems are fixed. Even if all known problems are fixed, there could be problems that no one has reported. Besides, if they ever did say all memory problems were fixed, people would complain that they're again denying that Firefox has memory problems. It's a catch-22, isn't it?

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

[phreakhead]

There are definitely some left! Right now Firefox is taking up 1.5GB RAM and 2.2GB virtual of memory on my Mac with about 10 tabs open across 2 windows. If I force quit and then restore the same session, it take up about 0.7GB RAM and 1GB virtual (and that will just keep climbing as long as I leave it on). If that's not a memory leak I don't know what is...

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

[stony3k]

It may just be that the web pages you have opened are not static - that is, there may be some javascript activity going on (and this is very common especially for ads). Some flash may be getting loaded, etc. This could very legitimately be increasing your memory consumption without there being any memory leaks.

Too many of us, especially old timers, assume that web pages will not change once loaded, but that is no longer true. If you leave a web page on for 5 hours, there may be 100 different ads displayed on that page - all of which get cached and increase the memory consumption.

And the trend is towards even bigger and more complex web pages which means that even more memory will be consumed by all browsers. That said, I hope Firefox devs can fix the memory issues - that would be another vindication that the open source model really does work.

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

[flakier]

I see what you're saying. Still, I think with some due diligence the majority of all these little problems could be fixed...enough of them that popular opinion could be fixed in terms of it. IMO, it would almost be enough for the "team" to simply make an anouncment to the effect that yes the end-user experience is in fact suffering at the hands of various memory allocation issues and we are going to be putting a strong focus on fixing this in the next n builds. Put a bounty on it even...$100 bucks to the first AutoIT script that will put fooMozilla product at >=100MB private bytes (or whatever was deemed excessive).

Perfect...of course not. But if they fixed even 80% of the RAM issues I would probably be less likely seen using Opera.

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

[bunratty]

They've already said "Our extensive testing shows an occasional leak here and there and we are working to fix those, but in general we aren't seeing many leaks anymore." It seems like they've done everything you've suggested (except for the bounty), and yet you still seem unsatisfied. Isn't that proving our point?

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

[onefriedrice]

> Form the tests that the developers have been running, most of the memory leaks in Firefox itself seem to be fixed...

The memory leak issues in Firefox have really surprised me. Yeah, so C++ doesn't come with automatic garbage collection. That doesn't make memory handling too much harder since the one of the first systems a team of programmers (who aren't idiots) will lay down is the system to handle the allocation and deallocation of memory, be it garbage collection of smart pointers or whatever. Or, if the original base code was such a mess that it's taken this long to implement fixes for these issues, perhaps the Firefox hype got misplaced. I remember people were somewhat surprised that Apple adopted khtml rather than gecko for WebKit, but they clearly knew what they were doing.

So you're right. I like Firefox and it's my default browser because I have a lot of RAM on my system, but perhaps I have become a little bit disenchanted, and perhaps my confidence in the mozilla developers to manage the code and fix issues has waned. Not to say they're bad coders, because I think the code itself is the biggest problem.

not to point out the obvious

[wwmedia]

firefox 3 doesn't have any plugins yet, last i checked it was plugin writers who were blamed for all the memory issues by Mozilla

btw i did same test in IE7 and Opera9 and only got 30-40MB usage

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.

Re:not to point out the obvious

[Mr. Underbridge]

last i checked it was plugin writers who were blamed for all the memory issues by Mozilla

Which to me sounds eerily similar to Microsoft blaming 3rd party software for taking down the operating system.

Re:not to point out the obvious

[Ash-Fox]

firefox 3 doesn't have any plugins yet

All my old plugins, flash, vlc's movie plugin and so on still work in this Firefox 3.

Re:not to point out the obvious

[ToLu the Happy Furby]

You can enable extensions not explicitly marked as compatible with Firefox 3 beta by going to about:config and adding an entry for extensions.checkCompatibility : false. I'm running the same extensions and usage pattern as with Firefox 2 and performance is MUCH improved, especially AJAX performance on Gmail and shutdown/session recover speed. Of course, it has only been one day since my last FF restart. FWIW, I'm running about 8 extensions and have about 50 tabs open across 5 windows; currently on my 2GB machine Task Manager shows Firefox 3 using 235MB, where in the past Firefox 2 would easily consume ~450MB or even 600MB+ under similar workload. (Of course in the past I only checked Task Manager once FF's performance became noticeably slow, so this is not necessarily a good comparison.)

Another point regarding your IE7 and Opera9 tests: as far as I know, all modern browsers choose to allocate more or less memory depending on how much memory the OS reports as available (certainly Firefox does), so users on different boxes can show very different results.

Re:not to point out the obvious

[MMC Monster]

If an extension starts allocating megabytes of memory, should a browser stop it? How much is too much? Should this be a limit set by the user (god help my dad, in that case)?

If an extension crashes, I'm not even sure if the browser is supposed to stay open, or crash with it. Certainly an OS should not crash regardless of what a browser does.

Re:not to point out the obvious

[the pickle]

>last i checked it was plugin writers who were blamed for all the memory issues by Mozilla

Which to me sounds eerily similar to Microsoft blaming 3rd party software for taking down the operating system.

Except "taking down the operating system" is very different, both in severity and root cause, from leaking memory. If you're going to allow extensions to run as part of the browser, you don't really have any control over what they do with the browser's RAM usage. An OS has the luxury of being able to partition processes off into their own memory spaces if it so desires, but to do this within the constraints of a single application and within the limitations an OS imposes on how much low-level memory management apps can do on their own is up there on the difficult-to-impossible scale.

p

Re:not to point out the obvious

[CastrTroy]

What about a task manager extension for firefox that shows how much memory each extension is using? Seems like it could be useful. I mean, we know how much memory firefox in general is taking up, but it would be nice to get a breakdown of where that memory is going to.

Re:not to point out the obvious

[Stewie241]

Except that using a lot of resources and causing a crash are very different... There are certainly third party apps that use lots of memory in Windows... that is pretty clear. But the argument is that the Operating System should be smart enough to not let those apps bring the system down. It might be useful to be able to limit the amount of memory that plugins can use... using resources (i.e. memory) is very different from causing a crash.

Re:not to point out the obvious

[bunratty]

No, plugin writers are not being blamed for all the memory issues. At any point in time, the total memory used by Firefox is a combination of necessary memory use, memory used for caching, memory fragmentation, memory leaks in Firefox, plugin memory use, extension memory use. Pointing to any one of them and saying it's the sole problem, or even the main problem, is simplistic. Certainly memory leaks in extensions are the reason for a significant percentage of memory complaints.

Re:not to point out the obvious

[Anonymous Coward]

Can't be done.

You probably were foolishly deceived into thinking of extensions like an extension, something that hooked into well define extension points. (I wonder where you got that idea?)

They aren't.

Think of an extension as a run-time patch, which is much closer to the truth. There's no separate name space for extensions. This means that extensions can step on each other, since they're all operating in the same giant JavaScript mess.

When it comes to the UI, the browser uses a system called "overlays" which are basically a method for specifying an XML patch. It literally patches the browser's internal XML representation of the UI with new elements.

Two things there: first, the UI is represented by an XML tree at run time. (And you wondered why Firefox took a lot of memory?) Secondly, it actively alters the internal structure. There's no way to separate memory usage into what the extension added and what Firefox started with, because it's merged into a single final "patched" version of Firefox.

Firefox extensions don't so much extend the browser as they actually alter them. One side effect is that extensions can't be toggled on or off without restarting the browser.

Ultimately there's no way to track how much memory they're using compared to how much Firefox is using, because the final result in a "patched" Firefox with the original Firefox XML and JavaScript merged with all the extension XML and JavaScript.

Re:not to point out the obvious

[Ambidisastrous]

Since users tend to treat Firefox as sort of a mini-desktop environment of its own (at least I do), I guess the "right" behavior when extension XYZ crashes would be for Firefox to toss up a simple alert saying, "Extension XYZ just hurled." Possible options:

• Keep going without it
• Try to restart the extension (is this even feasible with current codebase?)
• Restart Firefox (I know this capability is built in)
• E-mail a traceback to the developer (this could actually work really well if it was integrated into the extension system)

For memory limits, I would think that should be handled by the OS, or Firefox talking directly to the OS, not the user. So when malloc() finally fails, it would be nice for Firefox to squash the extension that called it (with the above notification), rather than seppuku. I guess. Because who knows, maybe there is or will be an extension that actually should be allowed to use insane amounts of memory -- like an image or video editor, maybe.

Expanding on the FF-as-OS idea, it would be nice to have extensions that work like kill, nice and top, as another poster mentioned.

Re:not to point out the obvious

[644bd346996]

If you can come up with a way for the web browser to know how much memory is reasonable for a plugin to use, please tell us.

Re:It's about time...

[Tesen]

Hey it might get someone laid, you know they could flash out the changelog at a local bar and say to a chick: "I did that! I reduced the memory usage of Firefox!" And when that hot sexy blonde rolls her eyes, they'll probably end up with something that is a cross between a horse, an elephant and Dick Cheney.

Tes

FFX3 uses up all memory and causes thrashing

[t35t0r]

Can anyone else verify this on Linux?

https://bugzilla.mozilla.org/show_bug.cgi?id=404688

Re:FFX3 uses up all memory and causes thrashing

[roblarky]

I can confirm this has happened on WinXP SP2. Happened two times in a row. Checked Task Manager, mem usage > 670MB and VM size was 1.6GB, it took 25 minutes to be able to get TaskMan open and kill the process (I had unsaved code in VS.NET, wasn't about to reboot). After I killed FF3, I saved my work and re-opened (this time I reduced the process priority to below normal..); everything was fine for about 15 minutes then I saw the HDD light start going crazy again. I left TaskMan open and saw FF start eating RAM and VM, killed it and went back to FF2. Rob

Re:FFX3 uses up all memory and causes thrashing

[Anonymous Coward]

Fx3 Beta 1, testing without swap from a Knoppix live CD in a VM; because the build process is a PITA and those wacky mozilla people aren't with 2004/64bit computing yet!

Top shows a steady 141MB for firefox-bin. An lsof shows 2 handles each for formhistory and urlclassifier3 sqlite databases!?!? I've not assigned any spare memory to the VM (total 256MB) -- it's still responsive.

What kernels are you guys running? Anyone reproduced on a BSD?

Re:FFX3 uses up all memory and causes thrashing

[f8l_0e]

I had the same thing happen to me. Restoring the session when I started firefox again brought the problem right back. I killed the task and brought it back up again and chose 'New Session' and everything went back to normal.

Re:FFX3 uses up all memory and causes thrashing

[Anonymous Coward]

I confirm it on Ubuntu 7.10. Mem usage just climbs until the machine freeses in a swap-thrash frenzy. I have reverted to FF2 for my normal usage.

Re:FFX3 uses up all memory and causes thrashing

[zoid.com]

It does on my PPC Powerbook. It used 1.5 gig in a matter of 45 seconds.

Re:FFX3 uses up all memory and causes thrashing

[Anonymous Coward]

https://bugzilla.mozilla.org/show_bug.cgi?id=404645

Seems urlclassifier3.sqlite (nsUrlClassifierDBService.cpp) stores gzipped data in it's moz_classifier table. Kinda makes me glad I disable superfluous crap like this (google suggest and co).

Why didn't they contribute an upstream patch to support compressed databases in sqlite? This is why I don't work in software anymore; it makes me want to club baby seals to death or something!

Re:FFX3 uses up all memory and causes thrashing

[Anonymous Coward]

Confirmed VM lockup after extended inactivity.

Re:FFX3 uses up all memory and causes thrashing

[dankelley]

Happened to me in OSX. I killed it, using a force-quit operation, and it seemed to be fine the second time. Since it lacks plugins, though, it's hard to see why anyone would use it when safari-3 is out, and is so much faster. (Yes, I like the "places" idea, but that's not enough to encourage me to use something that may freeze up my system.)

Re:This is irrelevant

[Pazy]

eh.. I read slashdot all the time and I have 1gb (2*512's).

This explains a part of it (memory fragmenting)

[Anonymous Coward]

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

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

----

Netscape 9.0.0.3 (0% unpatched)

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

----

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

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

----

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

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

----

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

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/

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

apk

Re:This explains a part of it (memory fragmenting)

[wwmedia]

pitty opera doesn't have a firebug type plugin (for javascript debugging)

then i would have no need for firefox at all during the development stage

right now i have firefox open and its using 170MB ram (which is not too bad, ive seen worse :(, and is possibly due to all the dev extensions i have) but compare that to a vmware images i have running now ( windows xp and suse10.3) both are running comfortably with 120MB allocated RAM each

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

It MAY have one... I am not sure though. There ARE "addons" for Opera too!

(& they're called "Opera widgets").

Perhaps there is one like you mention OR, perhaps not... but, no time like the present eh?

Perhaps it's waiting for YOU to create it (why not, if you have the skills, you know?).

APK

P.S.=> In any event, the analysis done that illustrates memory fragmentation creates uncontiguous blocks of RAM which FireFox allocates & uses & HAS TO GROW QUITE A LOT to operate, was interesting... it goes a long ways towards making some points I had in a discussion I had with Dr. Mark Russinovich & many arstechnica people on the Windows IT Pro website in fact circa 2003-2004. Nothing like proving a GOOD point, vs. the "industry gurus"... & Windows VISTA's showing thusfar (also a RAM eater by way of comparison to Windows 2000/XP/Server 2003) under Dr. Russinovich's tenure @ MS is another... apk

Re:This explains a part of it (memory fragmenting)

[Ash-Fox]

I don't use Opera for two reasons (I use Firefox):

• Doesn't work with some sites I visit (banking)
• No equivalent to Google Browser Sync

Re:This explains a part of it (memory fragmenting)

[Tranzistors]

VMware memory usage is rather tricky. I have one dedicated VMware server and guests eat only a 1/3 of maximum memory (reported by utility "top" and gnome system monitor). But that doesn't mean memory is enough (the machines start to slightly choke on that memory usage).
The VMwares' web management tool says all memory is eaten up (guests have ~95% of RAM). Go figure.

Re:This explains a part of it (memory fragmenting)

[Ash-Fox]

Perhaps it's waiting for YOU to create it (why not, if you have the skills, you know?).

Seeing how small projects like Firebug can consume huge amounts of time for it's developers, I doubt that there are many people who want to dedicate themselves to a single project like Firebug when they could be doing *insert thing they aspire more to doing*.

I have to say, I don't think this attitude of "don't like it? Go make it yourself!" makes a product you are promoting look good.

In any event, the analysis done that illustrates memory fragmentation creates uncontiguous blocks of RAM which FireFox allocates & uses & HAS TO GROW QUITE A LOT to operate, was interesting...

Seeing how K-meleon uses the same rendering engine as Firefox and yet uses barely any memory, I wonder if the whole XUL GUI framework that Firefox uses is the cause of these memory issues.

Nothing like proving a GOOD point, vs. the "industry gurus"... & Windows VISTA's showing thusfar (also a RAM eater by way of comparison to Windows 2000/XP/Server 2003) under Dr. Russinovich's tenure @ MS is another... apk

This is like gibberish now...

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

"I have to say, I don't think this attitude of "don't like it? Go make it yourself!" makes a product you are promoting look good." by Ash-Fox (726320) on Wednesday November 21, @09:37AM (#21434487)

I think what was being said was "if you have the skills do it" but if not, then writing the devs who make Opera widgets is the next best thing - one may be interested in creating it is why I mention this.

"Seeing how K-meleon uses the same rendering engine as Firefox and yet uses barely any memory, I wonder if the whole XUL GUI framework that Firefox uses is the cause of these memory issues." by Ash-Fox (726320) on Wednesday November 21, @09:37AM (#21434487)

It may be something the guys from FF can use then. If you're concerned, I'd point them to the article noted & see if it helps the FF team.

"This is like gibberish now..." by Ash-Fox (726320) on Wednesday November 21, @09:37AM (#21434487)

To each his own. That's just an opinion, & you're entitled to it, but I found the article from pavlov.net far more useful than your rather useless comment, which has helped nobody by its mere utterance.

Re:This explains a part of it (memory fragmenting)

[fimbulvetr]

It's not just a js debugger:

It tells you the size and speed of download of all elements on the page.
It gives you a dom layout.
It lets you modify html, js, css on the fly!
It lets you right click an element and instantly find its source.
It lets you click a little symbol next to each and every css element so you can enable/disable it on the fly to see what is causing what.
It lets you view inheritence of any element in terms of css...

I just thought I'd mention this, I have no vested interest in firebug outside of my own use for it to substantially decrease web-programming stress.

Re:This explains a part of it (memory fragmenting)

[compro01]

is probably the MOST standards compliant browser under the sun (not counting HTML dev tools)

standards compliance isn't worth shit if the rest of the web isn't even close to complying with the standards.

last time i tried opera, i had some really freaky problems with some sites, like a menu appearing upside-down and on the wrong side of the screen! never managed to figure out what was causing that on that site (an internal page for at college)

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

I've had things happen "weird" in firefox as well. Your point's not restricted to Opera, or FireFox (or netscape & ie even). The point about standards is exactly what you noted: Who's following them? How many pages are IE specific?? Using an ASP.NET app might be an example here - IE will work just fine with them, because ASP.NET apps are geared to work with it specifically/mostly, but if you try to do them with FF or Opera & especially if they use an ActiveX control, it can mess up largely compared to how it runs in IE, server-side driven, or not (as ASP.NET apps are).

Re:This explains a part of it (memory fragmenting)

[darthflo]

Have you tried "free" yet? Ought to be installed everywhere, so just try it (-m outputs in megabytes for nice readability. Especially check the -/+ buffers/cache line.)

Re:This explains a part of it (memory fragmenting)

[Tranzistors]

Nice!
Checked with 'free' - VMware only uses ~ 15%, the rest is cache. :)

Re:This explains a part of it (memory fragmenting)

[Drive42]

Ugh. The collection of Opera widgets is depressing. Go to widgets.opera.com. Go to addons.mozilla.org.

Opera:
Popular: Touch the sky (weather), SimAquarium (really?), Analog Clock, Stay Secure (app to remind you how secure you are according to a bar graph), Google Toolbar (this isn't the official Google client. I've tried this...this...hideous thing. Don't think I've ever uninstalled something faster.), Video Downloader, Pandora Widget (admittedly cool), Sketchbook, Spirograph, Torus (block game), Functions 3d, Scientific calculator

Firefox:
Popular: Adblock, Noscript, Flashgot, Video Downloadhelper, IE tab, Download Statusbar, Stumbleupon, Greasemonkey, Firebug, Forecastfox

Ah, personally I'd choose Firefox. I lean towards the more technical side of things, anyway.

Opera:
Highest Rated:Touch The Sky, SimAquarium (really? again?), Artists Sketchbook, Torus, Staysecure, Pandora, Analog Clock, Dotoo (todoo list), VideoDownload Helper, Google toolbar (REALLY?), Spirograph, Functions 3d

Firefox:
Reccomended (not all of them, and this list isn't based on user reivews necessarily): ChatZilla, Greasemonkey, Clipmarks, Stumbleupon, Del.icio.us, WebDeveloper, Adblock Plus, Firefox Bookmark Synchronizer, Foxytunes, Firebug

I'm sorry, but there's more functionality in firefox's library of extensions. To be honest, sometimes I think that because my browser can: turn tor functionality on with a click, swap cookie profiles in a second, choose which sites I want to give permission to run scripts on my pc, open an Internet Explorer tab without firing up another browser window, do god knows what with Greasemonkey, initiate an irc or an ftp session, download both embedded objects and embedded video, control itunes, access a gmail drive, and a whole slew of other useful things from within the browser, that the memory hoggage might be worth it. It's not necessary, by any means, and it requires a quit and restart every so often, but, really, where else are you going to be able to do something like this? There's no need to reinvent the wheels and write a widget for a closed-source browser when all the wheels are already attached to an open-source one.

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

"I'm sorry, but there's more functionality in firefox's library of extensions" - by Drive42 (444835) on Wednesday November 21, @12:45PM (#21437307)

And, more bugs/vulnerabilities: Recall the "GreaseMonkey" one (since you mentioned it)?

That's JUST NAMING 1... there's others.

(NOW, above all else: Opera, as is (faster & more secure), has, outta-the-box/oem stock, more features in it than Mozilla/Netscape/FireFox (what have you from that family) than all of them!)

One NICE one in Opera, that's pretty "addicting"? Is SPEED DIAL!

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

"Ah, personally I'd choose Firefox. I lean towards the more technical side of things, anyway." - by Drive42 (444835) on Wednesday November 21, @12:45PM (#21437307) Don't you mean you lean to the INSECURE side of things? Or, is this data incorrect when comparing Opera to FireFox:

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

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

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

----

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

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

----

Perhaps you meant you like to surf SLOWER (as well as less secure), per this data:

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

AND, yes others (most recently in Javascript parsing speeds), right here:

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

To each his own... You claim to be more on the "technical side of things" & your choice indicates you are not very technical regarding security, OR speed/efficiency.

Re:This explains a part of it (memory fragmenting)

[Anonymous Coward]

"I've had things happen "weird" in firefox as well. Your point's not restricted to Opera, or FireFox (or netscape & ie even). The point about standards is exactly what you noted: Who's following them? How many pages are IE specific?? Using an ASP.NET app might be an example here - IE will work just fine with them, because ASP.NET apps are geared to work with it specifically/mostly, but if you try to do them with FF or Opera & especially if they use an ActiveX control, it can mess up largely compared to how it runs in IE, server-side driven, or not (as ASP.NET apps are)." - MYSELF IN MY LAST POST PARENT TO THIS ONE

Same with ActiveX controls, or JAVA .jar files, you name it, between browsers of all kinds noted (Opera, IE, Mozilla/FireFox/NetScape) here (or, interpreter engines they use e.g. -> Sun JAVA vs. Microsoft)...

(Forgot to mention that as well)

APK

All of these windows-centric tests

[SCHecklerX]

What I really care about is how it has changed on the linux platform (where I've never had an issue with it). Is it going to be an improvement there too?

Re:All of these windows-centric tests

[jesser]

Most memory leak bugs in Firefox have been cross-platform, and we haven't switched allocators on any platform yet, so I'd expect the results to be similar on Linux.

That's before it goes mad...

[Anonymous Coward]

I installed it this morning. Twice so far it has gone mad, chewing up all the memory on the machine. I found with top that it was taking 80+% mem on a 500M machine, and that's with 2-3 tabs open and only four or five pages of history. This of course drives the machine into swap thrashing and makes it difficult to even do anything to kill it.

Talking of which, it won't die. Three of the four times I have shut it down with the X on the window border, I have been left with a zombie firefox-bin process which I have to kill by hand. And this is which switching between versions 2 & 3 tying to figure out exactly why a drop-down list doesn't work in FF3 (it does perfectly in FF2).

This is on Ubuntu Gutsy. Maybe the Windows version is more stable, but the Linux one is not useable for me. Let's hope they fix it before release.

Re:That's before it goes mad...

[techno_dan]

Windows XP is the same, I had a few tabs open, and it was sitting in the background. 4 hours later, my machine started acting up. I looked at the memory usage in process explorer, and Firefox was consuming 800MB+ of ram, while the Resident size was over 2G. This is on an XP sp2, with 1 gig of ram. Still needs a lot of work.

Re:That's before it goes mad...

[Anonymous Coward]

You are confused. First if your RSS is over 2GB then of course you are consuming 800+ MB of RAM. Because RSS is how much is actually in memory. But if you've only got 1 GB, then you of course didn't have an RSS of 2 GB.

12 pages? Who has 12 tabs open?

[140Mandak262Jamuna]

Come on, this is not a fair test. When I go to bookmarks/open all in tabs in a folder, I usually open anywhere between 18 to 30 tabs. In fact the first thing I do is to open all the editorial cartoons bookmark folder under "open all in tabs". By the time I am done with email, I will have the 21 cartoons ready to be perused.

BTW I never found old FireFox's memory consumption as annoying as intransigence of some sites in refusing to support Firefox and the lax/laisse-faire coding for IE only. May be because at work I usually have a couple of four processor 16GB machine for development/testing. I used to have a dedicated 2GB machine exclusively for Firefox. But that old machine's hard disk started squealing with an annoying noise so I had to throw it away. Even at home with my puny 512MB 4 year old desktop or the 1GB 2 year old laptop I get by without any serious memory issues.

Re:12 pages? Who has 12 tabs open?

[ultramkancool]

Indeed. In my tipical browsing session, I have so many tabs, they're hard to manage (so, I use tab mix plus. I can have it use multiple rows instead of a scrolling tab bar, I also have it so I can use my mouse wheel to switch tabs.) I usually have a few windows open too. My machine only has 256 MB of ram, but, I've never had a problem on Gentoo, Ubuntu or even XP.

Re:12 pages? Who has 12 tabs open?

[mollymoo]

I've just read you post three times and I'm pretty sure you just said you had to throw a computer away because the disk started making a funny noise. You didn't really mean that, did you? No. Nobody with enough nouse to have development and testing as part of their job would do that, surely. Or are you a .NET developer?

Wow

[Anonymous Coward]

Wow, that sure is some sloppy programming...How can a simple web-browser use almost 100Mb of memory?

Re:Wow

[Ash-Fox]

Wow, that sure is some sloppy programming...How can a simple web-browser use almost 100Mb of memory?

The problem is one of your assumptions, Firefox is not a simple web-browser.

not an issue

[SkankinMonkey]

Memory usage really isn't a huge issue for most end-users. Sure if it was sucking up 800 meg with 2 or 3 tabs open people would complain but right now people are just starting to get used to the idea of tabs much less use 12 of them. The memory usage now is hardly a system stopper for most people who only run their browser and mail client and maybe an office suite and picture viewer.

Re:not an issue

[tji]

It becomes an issue when the memory grows at the alarming rate shown in the test. Considering the browser is the one app I have open all the time, and keep open for days/weeks at a time, memory usage does become a concern.

Also, considering some other applications I often have open are memory hogs too (Microsoft Word / Excel / Powerpoint, Apple Mail, VMware Fusion) memory efficiency becomes more concerning. Even with 2GB of RAM, I run into problems at times.

Re:not an issue

[FromellaSlob]

but right now people are just starting to get used to the idea of tabs much less use 12 of them.

Speak for yourself. I've been running 100-200 tabs on a regular basis for at least the last year.

Re:not an issue

[Anonymous Coward]

I don't understand that attitude.

Are you proud that Firefox took the computing revolution, which literally put supercomputers in everyone's home, and reduced those supercomputers to slow, unstable, browse-only webTV like appliances ?

Most of the Firefox defenders who make the "memory doesn't matter" argument, already know the following:

1) Other programs (Opera) do largely the same task in a 10th the memory, and half the CPU
2) A substantial portion of computers out there have less memory than can run Firefox well

Given that these facts aren't new to you, why can't you mentally address the "Firefox is a fat hog" issue ? I think some of you guys love Firefox the way Apple users like Apples, and then somehow otherwise rational people end up defending the corporate shenanigans of Steve Jobs.

I mean, it's not like your post made me STOP NOTICING that my computer is slow when I run Firefox. Instead, I am forced to notice a disparity between what my computer does when I use firefox (become crippled) and what "SkankinMonkey" posted on slashdot. I mean, I know that when I click on the URL bar after the browser has been open a while, the harddrive light will come on for 15 seconds or so, and then another 10 seconds after that I will get a cursor and can type in a new URL. You, or at least the majority of people promoting the "memory doesn't matter" mantra, know that also.

I know that as I am sitting here, I have a 3 MB jpg image that needs to be downsized for a customer's web page, and I know that to do the simple ImageMagick "convert" command I need to kill this Firefox first. If Firefox is not there, it will take a couple of minutes. If Firefox is running and has been for several hours of browsing, then forget it, it will take at least 45 minutes. By the way, I did all those "about:config" settings that are supposed to reduce you memory usage, and have no extensions loaded. Memory usage does go down when I reduce the cache size, but it doesn't solve the problems, mainly because it is probably due to memory fragmentation.

I think there is a similarity between the Firefox Anti-Defamation League and the hard-core LISP bigots. Here are some of the similarities:

1) both groups deny that the fact that their software is obviously inefficient in use of the computer matters
2) as a corallary to the above, both groups disparage computers that have less than cutting-edge benchmarks
3) as a corallary to the above, both groups drift into a contemptable attitude towards those with less economic resources
4) I suspect, but cannot prove, that membership in both groups is associated with an extremely empowering early experience using the technology in question

I think it would behove you to examine the history of LISP cultism. On comp.os.lang.lisp you can see people predicting the return of LISP given that all computers will soon X amount of RAM, for about the last 15 years. All computers in use today have specifications that once caused LISPers to announce the return of the Messiah. Think about that.

Re:not an issue

[kardar]

On Linux, at least, it seems to happen with more than just Firefox. I just randomly checked my "top", and the browser I use, Opera, is consuming 424megs resident and 738 virtual. As I have 2 gigs of RAM, I usually restart Opera when it tops 1.0 g or more virtual. The memory usage will only continue to grow over time.

Firefox does the same thing. The complication with Firefox is the fact that it spikes to 100% CPU usage from time to time. So you've not restarted the app for a while, it's eaten up your memory without your knowledge (of course you're used to keeping your Linux system up for months, why not one of your favorite apps, right?), and it's swapping like there's no tomorrow, now it pegs your cpu on top. Pretty please?

This is in fact the last time I remember having anything even slightly resembling a BSOD on a Linux system in the past few years (with the exception of some ATA driver changes and a mismatched kernel). I actually went Ctrl-Alt-F2 and logged in as root, killed Firefox. This was a couple of weeks after Opera went free, so I switched and never looked back. At least with Opera, when you restart (which you eventually have to do), you're back where you started. It's no sweat to restart. I do it often.

I still use Firefox sometimes, but only because my cookies are turned off on Opera, and so I only use Firefox todo something that requires cookies. Nothing else.

This is a problem with all web browsers (at least on Linux), but the two bigger problems with Firefox are the 100% CPU utilization bug and the fact that you can't restart - out of the box - and take up where you left off. Fix those, and the memory thing is more or less irrelevant.

Just the other day, my brother called me and asked me to investigate his laptop, and why their web surfing slowed down. Did he have a virus? Well, turns out he tends to put his laptop to sleep, and keep Firefox open for days or more. His wife does the same thing, on another account. Problem solved, although it didn't occur to me right away. Restart Firefox. I downloaded and installed Opera for them and told them to use it instead.

Anyway -- you have to restart the app eventually. Priority #1 is to remember where you were out of the box. #2 it would be nice to not have the 100% CPU utilization bug. If someone can figure out how to do garbage collection without a restart, that would be brilliant - pure genius.

What it all boils down to is that with Opera, your CPU doesn't peg in strange ways, and you can very comfortably restart and take up where you left off. It may not seem like a big deal, but in fact, these two things really cut to the heart of the matter. If you forget to restart or put off restarting Firefox because you're doing something important and don't feel like engaging in a massive bookmarks session when you fire it back up, but you push it too close, your memory gets eaten up (which is more likely when you have more stuff open - which is when you're least like to want to restart), and then your CPU unexpectedly peaks, and it's already swapping like crazy, then your wonderfully stable Linux system can go into a semi-freeze like state, swapping and pegging your CPU. Your mouse doesn't move, your menus don't respond -- well actually, it's just that the boomerang is a thousand miles away (which complicates the situation) -- you get the idea.

As much as I love Linux, and am a strong supporter of Open Source stuff, I just don't like Firefox. I'm fine with Opera, and then if that doesn't work, I just go straight to IE (which is not very often).

Compared to usefulness? Compared to total RAM?

[Anonymous Coward]

How can someone be worried about -/+20MB memory usage of application that actually does something with it then be using Vista which does not and uses a hole lot more...

Well opinion or not I wouldn't mind a browser using 400MB (20% ca on most of my machines) if it meant snappier history/rendering with less CPU usage.

Re:Compared to usefulness? Compared to total RAM?

[webmaster404]

What Firefox needs to do is in about:config have an option that would be "Use RAM saving rendering" have it set as true but if you have loads of RAM just set it to be false, good for people running Firefox on slower systems and good for people with 4 gigs of RAM

Re:Compared to usefulness? Compared to total RAM?

[tepples]

What Firefox needs to do is in about:config have an option that would be "Use RAM saving rendering" have it set as true but if you have loads of RAM just set it to be false, good for people running Firefox on slower systems and good for people with 4 gigs of RAM Why an option? Can't an app just have the operating system check the amount of system-wide allocated memory and automatically shift into RAM saving mode based on how much is available at program startup time? Or does monitoring system memory consumption need administrator privileges?

Re:Compared to usefulness? Compared to total RAM?

[webmaster404]

Because it depends on what you are trying to do. If I am just web browsing and nothing else, I would want it to use full RAM no matter if I just have say 200 MB free, if I am going to do other things in addition to web browsing such as putting a music video on while I code and compile, I would want it in RAM saving mode even if I have 600 MB free at the time that I launched it, and to put a process managing RAM will just use up more RAM. Options are good, they let the user choose, that is what makes KDE popular is that you can configure most things easily, you can do the same in GNOME but it requires more hacking.

Re:Compared to usefulness? Compared to total RAM?

[tepples]

I would want it to use full RAM no matter if I just have say 200 MB free, if I am going to do other things in addition to web browsing such as putting a music video on while I code and compile, I would want it in RAM saving mode even if I have 600 MB free at the time that I launched it Then it should periodically poll free system memory and invalidate some caches if it starts to fill up.

and to put a process managing RAM will just use up more RAM. So would putting one more SWF advertisement or even one more image on a web page. When trading off complexity of managing scaling of performance, quantitative statements matter. So how much more RAM are you talking about?

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

[sreekotay]

They should add the "Virtual Memory Size" column in Task Manager and use that for comparison. It better reflects actual memory consumption.

The "regular" "Memory Size" column is the "working set" - so its possible that IE or FF 2/3 took more during page loads that hasn't been reclaimed by the OS because no one neededed it.

To see how this can be bogus, try minimizing all the windows for an app and watch "Memory Size" shrink as the working set is paged to disk. "Virtual Memory Size" won't change. See here for more info.

Additionally, one (probably) should disable toolbars/extensions - depends on what you're trying to test, of course, but IE's more likely to have some bogus BHO or toolbar installed by third parties (like Google, Yahoo, AOL, etc.) that are actively sucking RAM; which affects the steady state.

One could argue that's just the real world - but the intent is to compare *browser* efficiency?
----
graphically speaking

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:This is irrelevant

[Anonymous Coward]

and I read slashdot and I have 256mb (2*128) in my laptop. 100mb vs 60mb means a lot to me.

OLPCs laptops have only 512mb.

What a stupid "test"

[JeremyGNJ]

Seems like the author is playing up to some feature in FireFox 4 that releases un-viewed pages from memory after a certain amount of time.

I bet if he re-clicked on each of the 12 tabs after the 5 minutes was up, that memory usages would go back up again.

"using less memory" isnt always desirable. I have 4 GB of RAM in my system and i'd rather if the applications USED THAT RAM, to keep application response "instant", rather than un-caching stuff, only to pull it back into memory again when I want to see it.

Re:What a stupid "test"

[rustalot42684]

"echo 0 > /proc/sys/vm/swappiness" You will need to be root (obviously).

Re:What a stupid "test"

[niceone]

"using less memory" isnt always desirable. I have 4 GB of RAM in my system and i'd rather if the applications USED THAT RAM, to keep application response "instant", rather than un-caching stuff, only to pull it back into memory again when I want to see it.

A good point! And firefox does seem to take this into account. I am running Firefox 2 on Ubuntu (right now!) on a Thinkpad T20 with 256MB of ram - it works fine.

Re:What a stupid "test"

[QuoteMstr]

Don't do that. "But I want to keep my applications in memory!" you might say. That's wrong. Virtual memory systems these days basically use main memory as a cache for the disk. It doesn't matter whether a page came from a file, an anonymous application allocation, or anywhere else. The kernel automatically keeps the most frequently used blocks in RAM and pages everything else out to disk. By using 0 for swappiness, you defeat that automatic management and force the kernel to treat application pages specially. You don't want to do that.

Re:What a stupid "test"

[Tranzistors]

It would be much nicer if OS would take care of what should be cashed.
FireFox could just write unused data to some temp file and OS cashes this file.
If there is lots of unused RAM, this should work quite fast.
If there is shortage of RAM, other applications won't choke.

Re:What a stupid "test"

[Kjella]

Maybe because they can be dense as brick? One of my favorite sports is
1) Fire up a lot of applications
2) See a movie
3) Go back to my applications, and realize they're all swapped to disk and slow as hell because my memory is full of extremely useless file caching. A lot of the time it's like that, you only need it once. And at least in C++/Qt which is my favorite, I haven't seen any way to hint to the OS that this file SHOULD NOT BE CACHED. So it's a lame workaround for a lame system.

Re:What a stupid "test"

[QuoteMstr]

The system doesn't page to disk because it feels like it.

If the kernel is paging your programs out, it's because something else is paging data in. The correct solution is to fix the offending programs*, not to lobotomize the VM.

* I'm looking at you, updatedb(8).

Re:What a stupid "test"

[Anonymous Coward]

Disable the swap file. It's useless. I disabled it on my Windows laptop with 448MBs of RAM and on the linux server with 128MB RAM. It's been working fine for couple of years now. If you're into games, you can get 2GB or 4GB ram and they'll work fine. No need for a swap file, and especially if you're using solid state drives.

Re:What a stupid "test"

[mpcooke3]

You may just be seeing the effect of the contents in the VM disk cache being changed. Eg, previously your applications seemed very responsive because all the files they were accessing were in the VM disk cache and after watching a movie they no longer are.

I have a similar problem with postgresql, which relies on the the VM disk caching for most of it's database caching - the problem is similar to what you are seeing in that you lose all fine grained control of what your memory is used for potentially resulting in a big slowdown because important stuff was ditched from the VM disk cache for something you know doesn't need caching.

Re:What a stupid "test"

[rtaylor]

Ideally you could tell the operating system that the memory is useful but not terribly important and the OS will decide how much to keep around and how much to free based on memory availability and activity.

Applications should not be expected to dynamically change cache sizes dynamically based on system usage patterns.

Re:What a stupid "test"

[Pathwalker]

A lot of the time it's like that, you only need it once. And at least in C++/Qt which is my favorite, I haven't seen any way to hint to the OS that this file SHOULD NOT BE CACHED.

fadvise with FADV_SEQUENTIAL|FADV_NOREUSE on the file handle should do what you want.

And Links

[4D6963]

is using 14 MB (Ubuntu)

Re:And Links

[Drive42]

You know, 14mb seems like a hell of a lot for this.

Re:And Links

[4D6963]

Yup, but it has no tabbing, does it? So I had to run an instance of links for each page.

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:Yet flash....

[guruevi]

Same happens on Mac OS X with a PowerPC G4. Adobe's plugin will eat up 100% cpu for something as simple as a video or a game while other things like Java plugins or real video doesn't take anything.

Re:Yet flash....

[Anonymous Coward]

I use Opera on XFCE4. After the switch, there's no reason to look back.

Re:Yet flash....

[SilentUrbanFox]

If you're a GNOME user try epiphany ('epiphany-browser' in Debian and Ubuntu's repositories.) It's still Gecko, but I've generally had better luck with it performance wise compared to Firefox.

Re:Yet flash....

[Solder Fumes]

Never saw that problem, actually. Automatic Flash plugin installation system has worked flawlessly on all the systems I've tried for the past year or so, which is a pretty diverse mix of AMD and Intel systems and Linux distributions. There may be something messed up on your system.

Re:Yet flash....

[BZ]

Uh... What's the problem with compiling it? Apart from needing a compat autoconf version (which is always installed anyway) and a number of -devel packages (not really surprising, now is it?), it's a single make command...

Vista

[Anonymous Coward]

It takes a lot to load on Vista and has a splash screen about canceling or allowing something....

Re:This is irrelevant

[webmaster404]

Because mostly on Windows, most people's RAM is stretched to the limit, if a simple program that people use every day (Firefox) will decrease memory usage, then they can focus on speed and in the end, if Firefox can be 2X as fast as IE, Konqueror(and by extension Safari), and Opera people will switch to it. And I actually have around 512 MB on both my Laptop and Desktop with the Desktop currently running Xubuntu and my laptop running Ubuntu 7.10 happily. And when Linux can resurrect a "dead" system like a crashed Windows system that someone may give you for like $10 that happens to have 256 MB of RAM on it and a slower but usable processor like a Pentium III, Linux can run fine on it however, if FF runs slowly, most people have little need for a computer if they can't browse the web with it.

Memory AND speed issues

[keraneuology]

If I don't shut down Firefox when I leave work for the day my system will be at a dead crawl in the morning - it shouldn't do this. (The only other program that acts like this is MS Streets & Trips). I am annoyed that Firefox is painfully slower to load certain pages - I do a lot of work for an in-house Quickbase application and MSIE blows firefox out of the water performance-wise, to the point where the same page in MSIE will load 3-5 times faster than it will in Firefox.

Re:Memory AND speed issues

[0123456]

"If I don't shut down Firefox when I leave work for the day my system will be at a dead crawl in the morning - it shouldn't do this."

Let me guess: your company runs virus scans or similar on your PCs overnight, so Windows swaps out all your applications to cache the files the scanner has just finished scanning?

That's why I found that quitting and restarting Firefox was much faster than leaving it running overnight. It's not Mozilla's fault that the Windows disk cache was designed by a retard.

Re:Memory AND speed issues

[keraneuology]

I'd considered something like that, but at times I'd come in and find that Firefox had grabbed over 300Mb of memory to sit there and do nothing. Granted, shutting down Firefox isn't that big of a deal - I saw it cause the problem and found a painless workaround. But it is still a symptom of something that isn't quite right: an application shouldn't progressively grab more and more resources as it just sits there doing nothing.

As for global warming, by running the PCs I'm generating heat in this one tiny room so the furnace doesn't have to burn all of the extra fuel to heat the entire building.

Re:Memory AND speed issues

[mgblst]

Really? I regularly leave Firefox open for days, with about 20 to 30 tabs open at a time. It is up to about 180megs at the moment (after 3 days), and I constantly open and close tabs - 100s a day. I have found that I need to close ie down quite a lot. I do have no-script enabled, so maybe that makes a big different. Everytime I use ie it seems so slow.

Re:Memory AND speed issues

[DLG]

I agree that Firefox's overnight memory usage makes it necessary for me to shut the browser in the morning if I forget to do so at night. Weekends are even worse.

On the other hand, while I am not going to run benchmarks, I have found javascript/dhtml behaves slightly faster on firefox in the development work I do. That may be because I code on firefox and test on IE6/7 (firebug makes it just too easy to develop on Firefox for me to switch over). Obviously in the end I have to make sure that IE behaves well with my code, so I do tend to find ways to get the performance to be equivalent, however I really do find that both the type of html and the size of pages makes a big difference.

Re:Memory AND speed issues

[Anonymous Coward]

Fuck you, hippie. I burn as much of my trash as I can just to increase my carbon footprint and piss you retards off. Plastic sacks and all. You should see the lovely polluting black smoke it throws off ;).

Lick my carbon footprint and kiss my ass.

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.

Re:Won't be going back

[Nazlfrag]

Betas can be surprisingly stable, but I'd give it more than a cursory glance before recommending it widely. Until you've played with it for at least a week, you won't see its weaknesses. That said, I'm sure responsiveness and memory usage are better because they were two of the glaring flaws in the package and an obvious site for improvement. As for the rest, well who cares if other little issues remain, if they did fix those two glaring flaws and kept the rest tight it should be the de facto browser fairly soon. Finally we can code to W3C not Redmond. Netscape may have lost many battles, but the war isn't over by a long shot.

the blog is brain fart

[computerchimp]

Nothing wrong with making a blog post about 5 minutes worth of work that really amounted to nothing. Blogs are always full of brain farts.
But, then posting this on slashdot...thats an entirely different type of stinky brain fart.

Was this a slow news day?

CC

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:Memory still an issue for me...

[j_sp_r]

You can try removing your profile folder, reimport your bookmarks (I had to use restore, import grinded to a non-working halt)... Maybe that's the problem, I don't know but it works nicely now (including some plugins, use extensions.checkCompatibility=false)

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.

Re:Memory still an issue for me...

[Anonymous Coward]

Everybody has these problems, but the open source zealots don't want to admit them.

If a site heavily relies on javascript, after a while, firefox takes 2 seconds to switch/close between tabs, make links clickable etc...
It's a pain.

Re:Memory still an issue for me...

[owlstead]

Will you guys stop complaining and post the freaking URL's and configurations already? Jeez, I am not saying that the user should start to program or anything, but this is the 7th or 8th post saying that it still has problems, but don't let anybody find out for themselves. For all we know, you've enabled the option to pre-load pages once in FF, and forgot to put that back.

There is *no* browser that has any memory quota for web pages afaik. So if a page has problems, it might very well be that 800 MB is used in an instant, especially if dynamic HTML is running much faster in the browser (as seems to be the case for FF 3).

Oh, and try it with other browsers as well. If they inhibit the same problem, it is definitely the page that's causing the problems.

Re:Memory still an issue for me...

[Anonymous Coward]

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.

Yeah, I notice that too when I do *cough* research */cough* that includes pages with a couple hundred high res images. I'd suggest limiting the scope of your research, or you may want to invest in a program to download all of those images instead of trying to view them all at once in a browser. Try something like wget.

Re:Memory still an issue for me...

[cyphercell]

you're a dork, this whole thread is open source zealots bitching about problems with firefox. FWIW I find that Outlook 2007 is just as bad. that's with about thirteen tabs open in firefox and outlook sitting there waiting to send and recieve.

Re:Memory still an issue for me...

[Anonymous Coward]

Most people do not see that kind of out of control memory usage and a lot will depend on what pages you go to, the number of tabs you have open, and the Firefox plugins you are using. That's a lot of variables to narrow down your issue. I would advise first disabling or removing all of your plugins. Then identify which web pages cause the endless creep in memory usage.

With multiple plugins and going to sites like digg and slashdot, my memory utilization tends to hover around 100 - 180 MB. That's still ludicrous, in my opinion, but is hardly 900 MB. Google reader tends to cause Firefox to consume large amounts of memory but I never see my memory usage go beyond about 180 MB.

Good luck!

Safari on Windows

[TheDrewbert]

12 tabs, 10 minutes, 133megs, but still stable and fast. FF is still my main browser though

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

Yes, but...

[supabeast!]

How much RAM did the Firefox 3 box have free after leaving it running a few hours?

Final conclusion:

[Anonymous Coward]

All three leak like the filthy, smoldering piles of shitty software they are. Try Opera - sure, it has its bitchy moments, but atleast it's properly programmed.

Re:Final conclusion:

[Anonymous Coward]

Sure, when they GPL it.

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!')

Re:Opera 9.5

[Ash-Fox]

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.

The issues I had were Javascript related, not browser identification (yes, I have tried).

Opera Sync: The upcoming Opera 9.5 has syncing between accounts too.

Which doesn't encrypt the information from the hoster (like Google browser sync, which uses RSA encryption), nor will it synchronize passwords or cookies - which is important to me.

Yes, I played with the Opera 9.5 alpha.

Re:Opera 9.5

[Nicolay77]

Fair enough. U_U

I would like for them to add session syncronization too.

Re:Opera 9.5

[pi8you]

You'd like that, but then you go and sync at work while forgetting about the porn tabs you have up at home...

CPU usage concerns

[mkuczara]

To be honset i can live with big memory consumtion, but i cannot agree on cpu usage. Quick test - go to [mmorpg.com] site (widely known website), then check cpu usage (im sitting on win xp). I get almost constantly 50% usage by firefox.exe process

Re:CPU usage concerns

[Pope]

So what? CPUs are like RAM; they're wasted if they're not being used. Why is everyone so bloody concerned about keeping their free memory open in these days of cheap RAM and virtual memory? It's not like we're all using System 7 or something like in the bad old days.

Re:CPU usage concerns

[Cheesey]

Installing noscript and flashblock might help, since Javascript and Flash are the two ways in which websites can waste your CPU time. Executing scripts correctly must also involve executing badly written scripts that burn CPU cycles, and the whole "Web 2.0" is full of badly written scripts.

Re:CPU usage concerns

[Anonymous Coward]

I'm getting spikes up to 40%, but it's averaging out to about 5%. :/

Re:CPU usage concerns

[bunratty]

Did you notice the Flash animation on the right side of the page? That's what's consuming the CPU, not Firefox. You can read more about the issue in the Firefox CPU usage article in the MozillaZine Knowledge Base.

Re:CPU usage concerns

[jesser]

CPU usage affects power consumption and therefore battery life. So it's bad if an application is using 20% CPU all the time, even when you're not using it.

I'm feeling the same

[Nicolay77]

But I'm testing Opera 9.5 Beta and I have 42 tabs open.

Checking RAM usage, it's using 237MB right now, as reported by Process Explorer.

why?

[Anonymous Coward]

Why are you leaving your computer turned on overnight? Modern operating systems have a save session function. We aren't talking about a server, but a desktop, right? Turn it off, help save energy, help lessen air pollution. If you have to work at home, just get a laptop and tote it around with you then.

Re:Why?

[Explo]

Well, even though I have 2.5 gigs of RAM, it's still not infinite and the more carelessly individual applications use it, the sooner I have to either get even more RAM or ditch some badly-behaving applications. Of course with basic usage (just a mail client, browser, some text editing etc) it may take a long time to bump into the limits, especially if one doesn't have the machine running 24/7 or shuts applications that aren't currently used. However, I'm personally an amateur photographer (image processing takes a lot of memory), generally don't close applications when I'm not going to use them for a while and use a wide variety of them as well. Thus, any memory leaks or inefficient use of memory will show up fairly soon in my case.

(Kind of ironically I tried lately switching to Opera because of a crasher bug in Seamonkey/Firefox and found myself switching back a few weeks later, because in my case Opera was very stable, but leaked a lot of memory, so I ended up in the reverse of the situation I tried to escape from; for me the memory usage of FF/Seamonkey always stays in the range of a couple of hundred MBs even if it's been running for weeks, but Opera hogged about one GB in a few days and didn't release the memory until I restarted it...)

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...

Re:Just did this test in linux

[Tango42]

Why are you comparing to an old version?

Re:Just did this test in linux

[Adult film producer]

Tons of people still use FF 1.5 and the automatic updates in Fedora 6 don't include FF 2 ..

Re:Just did this test in linux

[Tango42]

Nevertheless, it's not a very good comparison. There is no way to tell if the improvements are from changes made between 1.5 and 2.0, or 2.0 and the beta of 3. If there were many improvements made between 1.5 and 2.0, and then some of them were undone between 2.0 and 3, you would still see a great improvement between 1.5 and 3, despite 3 actually being a bad upgrade.

Re:Just did this test in linux

[A beautiful mind]

Seldom mentioned, but good news for computer users who want low energy usage:

If you've got a kernel that is >= 2.6.21, if you've got dynticks enabled, Firefox 3 fixes a few things and the result is that the browser doesn't generate as many ticks as before.

Re:Just did this test in linux

[Blakey Rat]

None of those columns have labels... can someone let me know what those numbers mean? For instance, in the first row, does "175M" indicate 175 MB of memory used?

Re:Just did this test in linux

[xehonk]

GP posted default-formatted output from top.
the headers are:

    PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND

Re:Just did this test in linux

[Blakey Rat]

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.

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.)

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.

Re:Regrettably It Also Locked Up!

[Anonymous Coward]

Hmmm... I wonder if maybe that's why THEY CALL IT A BETA. Ugh...

Lockups bigger issue

[BlueParrot]

Memory leaks are of course always bad and should be fixed, however, I have to say that a much more pressing issue is the tendancy for the interface to lock up ( especially on less powerful systems ) if one tab gets stuck loading or has to deal with a poorly coded javascript.

Mind you it is perfectly possible that the two issues are related, and since my knowledge about the inner workings of firefox are, to put it very mildly, limited, I suppose I can't really judge what kind of changes would be hard to implement and what the security implications would be etc.. I would however argue that the browser's interface failing to respond when it encounters a poorly written/ bloated webpage is a more devestating problem than a larger than necessary memory consumption.

Still giving issues

[kryliss]

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

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

Re:Still giving issues

[improfane]

Your taskbar is the same size as mine and Quick Launch has PSP it is uncanny!

Meta

[ObsessiveMathsFreak]

I'm sorry, but what did your reply have to do with the grandparent post? Shouldn't you have started a new thread instead of hopping on a +5 funny first post?

Re:Meta

[jZnat]

He should also learn how to use Slashdot's signature feature rather than spamming his website, but apparently he must be new here.

Re:Meta

[sreekotay]

oops - yes, and yes...

Re:Meta

[MrAnnoyanceToYou]

Actually, it depends on whether he had enough karma to get above the default 1+ filter everyone has....

It's actually mildly funny that the comments system is still borked IMHO. It's been a long time, and the comment ranking system still rewards highly fp's and other gaming.

Re:Meta

[Anonymous Coward]

But the editors suggest that one should reply to an existing thread rather than start a new one. Why, I don't understand.

sounds like a poorly written extension

[sweatyboatman]

I am going to guess that you have a couple dozen extensions installed on firefox and most of them you don't ever use (or even think about). Get rid of the extensions you're not actively using and see if that helps both the memory and speed problems you're seeing.

Alternatively, you could use Opera.

Re:sounds like a poorly written extension

[keraneuology]

IE Tab, IE View, PDF Download, Adblock Plus, User Agent Switcher, Bandwidth Meter and Diagnostics, BugMeNot, ChatZilla, DictionarySearch, FireFTP, FlashBlock, ForecastFox Enhanced - nothing else.

Re:sounds like a poorly written extension

[bunratty]

According to the list of problematic extensions, IE Tab has a memory leak. Perhaps you should uninstall that extension.

And lynx is using...

[Chemisor]

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
688 bob 20 0 483m 88m 26m S 0 2.2 1:47.84 seamonkey-bin
735 bob 20 0 21060 3792 2316 S 0 0.1 0:00.00 lynx

I think that whichever number you take, lynx is the clear winner. Eat your heart out! Though I am a bit at a loss to explain the 21 megabyte virtual image. Maybe it uses that novel hundred-buffering technique...

Re:And lynx is using...

[ultranova]

I think that whichever number you take, lynx is the clear winner. Eat your heart out! Though I am a bit at a loss to explain the 21 megabyte virtual image. Maybe it uses that novel hundred-buffering technique...

Well, libc alone is over 6 megabytes in this machine. Add in libraries for encryption, screen handling, locales, etc., and it adds up fast. Of course most of these are shared between multiple processes, so it's not like it's actually consuming 21 megs.

600+MB real memory usage

[MarkoNo5]

In OS X, I had to kill it when it used more than 600MB of real memory. I opened slashdot.org and dpreview.com and left it open for 15-30 minutes. Pathetic.

HTML vs XHTML

[Midnight Thunder]

I would be curious to see how they both compare on badly marked up, but renderable HTML, and XHTML. Normally XHTML should require a smaller memory overhead, than HTML, because HTML needs so many special case handling in order to handle badly marked up websites.

Memory is not as important as Stability!

[Danathar]

Although I'm sure there are some Slashdotters who run Firefox on a 350 Mhz PII with 256Mb of memory, that is really not issue for me. Most people with a recent PC probably have over a Gig of memory and more like 2 onboard.

CPU utilization where the browser all of a sudden is sucking down 100% of your CPU or of a single core and/or crashes are just as important (or more). More than likely the memory leaks have related browser stability issues that can be addressed with single fixes but if the browser continues to have runaway CPU issues and crashes it will not matter HOW small a footprint of memory it uses.

Re:Memory is not as important as Stability!

[Anonymous Coward]

In general, if you have a piece of software that runs like crap and you want to make it better, the first thing to look at is memory.

It is rarer for an unoptimized nested loop to kill a program, than it is for a loop with a malloc() and free() that should be outside the loop (so the same memory is just reused) to kill a program by memory fragmentation.

Once a program uses less memory, it is easier to debug.

I suspect, and some people have done analysis that supports this, that firefox's main problem is memory fragmentation. It is has tiny chunks of memory scattered around the place, and can't allocate medium sized chunks between the small ones.

I think one of the things that the firefox programmers will have to start realizing, is that they need to write less code. The project generated a lot of enthusiasm which swamped it. I think if they see a way to solve a problem by deleteing code, versus writing 100 lines to work around it, they will write the 100 lines because somewhere in their heads they are thinking "more firefox is better". Finally, they should have a way to write plugins in C, so that the ones that are important could be written well. They could leave the current XUL in there for amateurs and weekend "software engineers" to fiddle with.

I don't think most of the people in the firefox project realize how bad it has gotten. I suspect that many of them think that work on firefox will be a positive thing on their resumes . . . right now, I view it like having worked for Enron or SBC/ATT.

Re:Memory is not as important as Stability!

[dropadrop]

Although I'm sure there are some Slashdotters who run Firefox on a 350 Mhz PII with 256Mb of memory, that is really not issue for me. Most people with a recent PC probably have over a Gig of memory and more like 2 onboard.

I don't know what you do with your work machine, but unfortunately mine does not accept more then 2GB of ram and it's far too little to let a browser eat up as much memory as the operating system does. Firefox can easily use 500MB of ram after its been running for a week or two, that's more then my operating system uses when it starts up. I would need to run other apps too, email, Office, a music player, ssh sessions, an IDE a Windows or Linux virtual... Suddenly it matters how much memory your apps use. It's not that long ago when 2GB of memory seemed like quite a lot, then what happened? Developers got a "Most people probably have a gig or more memory" idea?

Re:now after 5 minutes

[MarkoNo5]

It seems to be speeding up. Now it sucked up 500+MB in about 5 minutes. No extensions installed. Think Java and C# are slow memory hogs? Try Firefox! They should give it codename Cartman.

Whiskey tango foxtrot.

[urcreepyneighbor]

2 used 103,180KB and 3 used 62,312KB. IE used 89,756KB. Someone care to explain why simple browsing consumes so much damn memory?

Re:Whiskey tango foxtrot.

[Spy der Mann]

Someone care to explain why simple browsing consumes so much damn memory?

Because each rendered page, and I mean the pixels, is kept in memory. Uncompressed.

Re:Whiskey tango foxtrot.

[Anonymous Coward]

Except, if you add up the size of all those pixels, you don't even get a 10th of that amount of memory.

Rather than do the math, just try opera. It is not going to the disk or reloading the page when you switch tabs, so they must be in memory. It's memory footprint is about 1/10th of Firefox's.

Is it time to re-evaluate using Firefox as the flagship of open source ? I'm beginning to think that it just gives ammo to the people who claim the "bazaar" method of development inevitably leads to crappy disorganized software.

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

[644bd346996]

If memory isn't in the working set, is it really hurting performance? For the vast majority of users, I think not. And the difference between browsers is definitely too small to care about. Besides, it isn't like anybody is limited on swap space.

Re:This is irrelevant

[someone1234]

4GB? Are you nuts? 1GB must be enough for everyone.

Okay, so I tried it...

[BobMcD]

Based on this and other encouraging news, I slapped it on the old desktop at work - where it subsequently tanked. Not sure why, and I don't have time to deal with it today, but it pegs the CPU at 85-100% and runs upwards of 800 MB's of RAM. I tried waiting it out, no joy. I tried dumping all of the cache, uninstalling, clearing IE's cache, rebooting, reinstalling. Still no dice. Even on a blank page (open new tab, close other tabs) it uses this level of resources.

My desktop is unusual, to be sure, but firefox 2 has always been fine. Perhaps later I'll find the link. Meanwhile, what should I have expected, it IS a Beta...

Meanwhile, FF3 Beta 1 is working perfectly on my laptop, and is much, much lighter.

Loading lots of images

[Anonymous Coward]

They should also try loading bunch of images and swf's at the same time. Sample of this site is http://canz.net/

Time for a new name!

[Shivetya]

and no, it wasn't meant as a joke

Re:This is irrelevant

[Count_Froggy]

Not all of us have the luxury of going out and just buying more memory.

I run into memory problems easily on my work machine, WinXP w/1gb ram. And getting this much was difficult! A typical work session will have:
- Firefox 2 with 3-5 windows into different email and PeopleSoft instances
- WinSQL logged on to 1-2 database instances
- Excel with nVision (PeopleSoft) add-ins, commonly 2-3 relatively small sheets open
- xplorer2 lite (file manager)
- Palm Desktop (PIM, Scheduler)


My personal WinXP/PCLOS dual boot is hardware maxed out at 2gb ram. I'm looking forward to the lower memory requirements of FF3.

memory usage

[Anonymous Coward]

Firefox 3 on my old MS 2000 PC at work took up 280MB of memory in twenty minutes. I had four tabs open, iGoogle, wordpress blog site, and two screens of an internal ERP system.

Considering this old system only has 324MB and I had email and Excel open already, I was stopped dead. I know it's a beta, but please. Flood the memory in 20 minutes?!?!

Currently Opera with eight external tabs - 14 MB. Firefox 2 with two tabs external three tabs internal- 142MB.

About Time

[WebmasterNeal]

It's about time they fix the memory leakage in Firefox.

Don't take it anymore

[bunratty]

I don't see that memory usage remains a problem for most users. It's just the vocal few who are having memory problems. The main problem is that these users assume this is part of the "normal" experience of using Firefox, so they complain that every user must also be seeing the same thing. They take no steps to fix or report their problems, as they consider the problem to be "well-known" and think developers must be idiots for not being able to see it.

If you're still having serious problems with Firefox, try creating a new profile and installing the Firefox 3 Beta. If you still have problems, discuss them on the MozillaZine Builds forum. If the problems do not get resolved, just switch to another browser. It's not normal to experience serious problems when browsing, so I don't see why anyone accepts it as part of the "normal" experience.

I agree that the damage to Firefox's reputation is already done. I've found that no matter how many reports come out that Firefox doesn't have a severe and obvious memory problem, the few reports that show a problem are the ones that become popular. If any of them just included instructions to reproduce the problem on other computers, those reports would be productive. Somehow, they always seem to leave that part out.

Re:Don't take it anymore

[rrkap]

I don't see that memory usage remains a problem for most users. It's just the vocal few who are having memory problems. The main problem is that these users assume this is part of the "normal" experience of using Firefox, so they complain that every user must also be seeing the same thing. They take no steps to fix or report their problems, as they consider the problem to be "well-known" and think developers must be idiots for not being able to see it.

Firefox is normally pretty good, but it has intermittent machine bogging memory consumption problems. I've experienced these problems on all three computers I use regularly, one running windows 2k, two running XP, with really different hardware and with no extensions installed. My wife, who runs vista, also experiences the same thing. That being said, having to restart firefox every couple of days isn't that big of a deal and it is more than made up for by it sucking less than IE in other ways, but it is a bug.

Re:Don't take it anymore

[bunratty]

I'm not saying it's not a bug. I'm saying that because no one is filing a bug report, it is not getting fixed. I'm sure that many people cannot figure out how to reproduce the problem, because as you describe it is intermittent. On the other hand, out of 130 million users, someone should be able to figure out how to reproduce it.

Firefox's problem:

[Anonymous Coward]

Javascript.

You don't write huge portions of a browser in javascript. Seriously, wtf are they thinking?

I'm sure if the entire codebase was in C and only ran javascript present in webpages that the browser would be lean and mean, fast and efficient.

I will not recognise comments on the insecurity of C. There is only insecure and poorly written C code. Relying on a platform on a platform on a virtual machine (like java) to enable you to write lazy unsophisticated code is the lamest fucking excuse I have ever heard.

If Firefox dumped XUL and javascript and actually wrote some quality real code for a browser in C, then everything would be just fine.

Re:Firefox's problem:

[Anonymous Coward]

But . . . do you want the kind of people who thought XUL and javascript were good ideas trying to write good code in C ?

Re:This is irrelevant

[petermgreen]

Anyone reading slashdot already has 4GB or more
I call bullshit.

desktop versions of 32 bit windows do not support more than 4GB of physical address space meaning ram usable to the OS is limited to some figure below that (exactly how much depends on the exact hardware configuration). Even if you are one of the minority that uses an OS that supports more the chipset may not. For laptops the situation is even worse, even if the chipset and OS support 4GB of ram 2GB sodimms are not cheap and I'm not aware of any laptops that support more than two modules.

On a system with 2GB of ram (which is the most many users can easilly and economically install) 200MB is 10% of your ram, that is not negligable.

Not convinced

[edelholz]

After Opera crashed on me for the first time as I was searching for "Firefox" (must not have much self-esteem), I installed the FF beta and proceeded to read the comments to this and the other firefox-related story. Wondering about some sluggishness, I fired up the task manager and was surprised to see a 99% cpu consumption and a memory consumption that was rising by megabytes per second... while the beta wasn't doing a thing but displaying two slashdot stories with comments. Not even the fancy new comment system. I killed it at around 900 mb. The uninstall procedure is nice, though.

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 and in CVS. I guess no one's updated the project web page yet.

Tags

[Kingrames]

Whoever manually adds the tags to these sites misspelled one.
I'm sure that should say "whatisopera" not "whatiopera"

Just a heads up.

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

[hedwards]

I would say so, it might not hurt the majority of /. posters, but for anybody that has only 512mb of ram on XP, it definitely causes problems relatively quickly.

IE does the same thing, my parents, prior to me educating them on the issue, would leave IE open for days at a time, by the end the computer would be so sluggish that there was pretty much nothing one could do but close the program.

So no, this isn't an exotic problem, often times non geeks will be hurt far more than those that are more inclined.

Completely bogus

[Anonymous Coward]

I downloaded the firefox 3 beta and tried to restore my bookmarks. Firefox took 95% of my 4GB of RAM and hung the system pretty well; I had to kill -9 the main process to recover my system. I won't be using version 3 until I know the memory use has at least stabilized.

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

[sreekotay]

In addition to what Hedward replied, I was mostly saying that the "working set" makes it difficult to compare apples-to-apples because you don't know the "real" state of the memory for one app vs. another: allocated and actively in use, allocated, but safey "page-able", or freed, but not reclaimed by the OS

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

[Anonymous Coward]

He's on Vista. Thats "Memory (Private Working Set)" vs. "Commit Size". (aka "Memory - Private Working Set" vs. "Memory - Commit Size", for some reason different names in the "Select Columns" dialog) At least I assume that's the column being measured, it's not clear from the screen shots.

And given that FF2 keeps a very large percentage of its virtual memory in its working set (I'm currently at 136/149 MB), this is still very interesting. It would be great to know both.

PCs != the world

[tepples]

There's no point RAM sitting there, not being used. It might as well be used to speed up page loading etc. rather than doing nothing. But Firefox should still be able to work without the extra RAM if we want free software to run on more handheld devices.

Re:This is irrelevant

[Anonymous Coward]

Why is memory use of a browser even relevant?

o Multiuser systems where several up to many users are
using a browser. If you have 100 users on a box, where
everybody uses a browser, it does make a big difference
whether these browser processes on average need 40MB, or
400MB.
o systems on which you are running other memory-insensive
applications
o most of the time its not memory actually needed by the
browser to render the pages the user currently views,
but some sort of memory leak or memory access error.
Some time later, the browser then crashes.
o swapping greatly decreases performance of various
applications, say, compilers.
o if the developers can't even manage memory properly,
what else is broken?

Thomas

Efficient Firefox? Finally!

[Jim Efaw]

Wasn't the original point of Firefox (pre-1.0) that it was a rewrite that was supposed to use less memory and be faster than any other browser, as opposed to the browser in Mozilla Communicator/Suite (now Mozilla Seamonkey)? I have a friend who uses Seamonkey constantly and still swears it's faster than Firefox. On the other hand, I'm running several Firefox extensions, and whatever speed Firefox (2.0) is right now, I think it's worth it. (Opera tends to be very fast, by the way. I just like my extensions too much.)

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

[coolfrood]

That's a bogus argument. Virtual size is absolutely unimportant as far as the performance is concerned. You have a fixed amount of memory in the system and if Firefox takes up a lot of real memory, other parts of the system will feel the pain. The fact that the *real* memory usage of FF 3.0 is low means that it is not being greedy about using system resources. Every process has the full VM size to play with, so looking at the VM size doesn't really tell you much about what effect that process will have on the rest of the system.

I'm not impressed

[Anonymous Coward]

Who cares if its faster or uses less memory? I care more about FF working correctly and rewrites of rendering engines to not be synonymous with breakage of existing sites.

Theres still quite a lot of CSS related breakage I would have hoped they would have addressed in the ALPHA stage before releasing a BETA quality browser into the wild.

I guess I'll be having more fun submitting bugzilla reports to people who are interested only in closing tickets as quickly as humanly possible while finding any insane reality defying reason to do so.

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:This is irrelevant

[Anonymous Coward]

Or 3.5 GB after memory mapped IO takes its share of the 32-bit address space :(

Opera Mini doesn't read web either

[tepples]

Well, I guess by some licenses and ideologies out there, a server reading the mail and presenting a HTML web page with the contents could be considered having fire fox reading email. Some versions of Opera don't read web either. Opera Mini uses a proxy server to translate HTML into something that Opera Mini can read.

Irony

[Anonymous Coward]

Hmm... I downloaded Firefox 3 beta 1 yesterday after reading Slashdot. Everything did look fast, and I was happy. I have made my peace with Firefox using a LOT of memory, and if everything is fast, I don't care.

Then I closed everything except 1 terminal of URxvt (also, I run Windowmaker) and left for my home. I came after around 22 hours and opened Slashdot to read this story as the first thing. Now the ironical part, just when I started scrolling down, Firefox starts to crawl. My hard disk starts making noise non-stop and I have to go to tty1 to kill firefox-bin. "top" shows me that firefox-bin with 1.2 %CPU and 99.4 %MEM. The other process it kswapd0.

Then I come back and see this thread (by that time only GP had posted and I start replying to him. I type 1 sentence and ka-boom! Everything stops again... this time I wanted to take a screenshot but even executing "Run" command took 2 minutes. So I go back to tty, kill firefox and go to a restuarant.

As of now, it is running fine, as I am able to type my reply. Draw your own conclusions, but I am back to opera for normal sites I visit.

PS: I use Firefox with Adblock plus and noscript, nothing else. I use Firefox for only those site which I don't want my boss to know :) This includes Slashdot.

Feta

[Anonymous Coward]

And then how effective would karma whoring be, huh?

Opera 9.5 is not out yet - you are running a Beta!

[Anonymous Coward]

You should have said so.

ROFL

[Anonymous Coward]

So ummm do they still blame in on BAD addons? Or do they finally admit the truth that Internet Explorer is better?

And now watch the Chinese censors (sorry moderators) mod this down reaching a censorship orgasm.

Firepenguin vs. Fireborg

[Zero__Kelvin]

I think Firefox needs a new naming scheme that makes explicit what platform it is running on, since people so often say "I'm running Firefox and it sux thusly" without identifying the platform. Then we would mostly "I'm running Firepenguin and it works great for me" and "I'm running Fireborg and it suxors." Then we could simply say to the latter group: "I'm afraid you implicitly gave up any right to stability, security or reliability as you chose an OS that quite has no commitment to these concepts." Yes, I know some people have experienced issues on Linux boxes as well, I'm merely saying that the number is certain to be quite disproportionate.

That being said, many posting here have a little knowledge about how memory use and performance correlate. A little knowlege is a dangerous thing.The assumption that less memory use is better is absurd. If I have 1.5 Gig free and my browser is spending its time conserving memory rather than rendering pages that is VERY WRONG. If it is Linux, use the memory you need and let the Kernel handle swapping, etc. Don't WASTE it, but don't waste memory by not using available memory either." That is a much bigger waste.

Conserve memory when memory needs to be conserved , and use more memory when plenty of memory is available.

A slightly off topic aside, hopefully eye opening to some, exemplifying how a little knowledge can be a dangerous thing:

I once requisitioned 16 Megs of RAM to add to the 16 Meg I had back in Win95 days. The typical person with minimal knowledge who thought he had a lot of it was trying to block my order, because he added 16 to his box and it still said 0% free so he concluded that "Windows just allocates and uses all the memory you give it." ROTFLMAO. I tried to explain to him about swapfiles and show him that the swapfile size decreased in proportion to the added RAM and why that was a HUGE win from a performance perspective, alas to no avail.

He told me he would be watching for my P.O. to block it, and I told him to have fun watching it go right through. I explained things to the CEO, who was not a complete moron, and he did in fact ultimately watch with gritted teeth as it got processed under direct approval of the CEO of the company. I figure we spent 10 times the money wasting time trying to address his ignorance than the actual RAM costs. The guy was quite an ASSet to the company :-)

Re:Firepenguin vs. Fireborg

[Zero__Kelvin]

What cracks me up about you Bill, is that you actually think that spending your new free time browsing Slashdot and modding accurate and informative posts as trolls because they happen to point out your company and your OS sucks, will somehow have a prayer of saving your drowning garbage of a company :-)

Re:now after 5 minutes

[hununu]

Yeah, I experienced exactly the same. back to safari... there's nothing as good as it.

Content differs?

[Anonymous Coward]

Welcome to the year 2007, where the User-Agent header has no effect on the content served to browsers.

Yeah, those results are accurate alright. The author mentioned nothing about analyzing the page content returned by the ZDnet server for differing User-Agent headers. Returning browser-specific content is a common practice used by many commercial websites.

Re:This is irrelevant

[LokiFoo]

Offtopic? Moderator is on tryptophan already. I didn't know they celebrated Thanksgiving in New Zealand!

(posted with a measly 2gb)

Comment removed

[account_deleted]

Comment removed based on user account deletion

Realworld Test: MySpace and Bebo

[meehawl]

The sites these "testers" choose are hardly able to get any browser really worked up. They need to hit some of the uglier MySpace and Bebo pages loaded with all kinds of crap and see what happens. Preferably the stripper ones with lots of huge Adds banners and porno music videos. That's like natural selection for browsers, either they make it through the first few minutes of opening them or they just die, mercilessly.

My personal best with Firefox 2 on Windows was 780MB of RAM "Memory Used" and 785 "Peak Memory" on a 1 GB laptop before I killed that browser sessions like the mad dog it was.

What About The Swap File?

[Plekto]

I've noticed that under Windows, applications tend to not release their virtual memory/swap file allocation. As in hardly ever.

I can run a firewall or AV program, for instance, and short of rebooting, there's just no way to reclaim it. Hence the memory leaks. ie - it's a glitch in Windows itself that so far I haven't been able to find a way around. Even memory manager programs won't free or touch this. So you get into silly situations where you have 1 gig virtual and 200 megs physical despite having 1 gig of ram.

Re:Firefox Memory Leaks, C++ Memory Leaks

[BuildMonkey]

I'm the lead for a group of over 20 engineers and programmers that produces public safety communications equipment and have direct experience with memory leaks large scale C++ projects AND Firefox. The system we sell includes embedded devices, a Linux/Oracle/Apache server for management, and management terminals. Management and monitoring is done via web browser, either on Windows or Linux. We officially support both Firefox and IE, and both leak memory. We strictly control what is installed on the management PCs - no browsers extensions, no ActiveX components, no Firefox Ad-ons. Typical usage is to open the browser and monitor one page continuously for weeks until an alarm is shown.

Let me be clear, both Firefox and IE leak memory so badly that even on management PC with 2GB of RAM we have to require the end user to restart the browser every week. We are monitoring FF3 and looking forward to reduced memory leakage. In the referenced article it discusses reducing memory fragmentation. OK, that's a worthy goal but first fix the memory leaks. Memory fragmentation and memory leaks are related, but different beasties. A memory leak almost always results in fragmentation, but fragmentation can happen simply from an unfortunate memory allocation/deallocation pattern.

Regarding C++ and memory leaks: over 2 1/2 years we've worked on the embedded code, which is pure C++, we have hunted exactly one memory leak. And that leak turned out to be from the OS. We use Boost smart pointers, RAII, exceptions, and exception safe code. We have no trouble with leaks or fragmentation, despite a fairly high turnover rate and a customer base that would quickly notice memory leaks requiring reboot of the embedded devices.

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:Experiencing RAM/CPU Problems? Try Again.

[gottabeme]

So...this bug (which I encountered but since I'm on a 400 MHz laptop with 128MB of RAM, found it not worth investigating) was caused by...downloading a file? Downloading a file? What is Firefox doing with this file that could possibly cause it to go bonkers like that?

Re:Experiencing RAM/CPU Problems? Try Again.

[lpangelrob]

Thanks for the post. I wasn't even aware it could be a network issue, and on an office T1-or-better, that killed my system real quick.

I'll try FF 3 Beta 1 again, and hope to see better results this time! I liked what I saw initially.

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

[Anonymous Coward]

Sounds like a feature to me. Best to run it regularly then :).

Re:Firefox Memory Leaks, C++ Memory Leaks

[msuarezalvarez]

Have you considered allocating resources to work on the problem of hunting those leaks and fixing them? One of the two browsers you mention provides you with full sources so you have what to work on. You seem to be one of the many people extracting value from Firefox: maybe you could put some value back...

cos of XUL runner?

[Anonymous Coward]

is the underlying XUL runner to blame?

song bird is a music player based on the same xul runner and it eats up memory similarly.

May be firefox needs to be built from ground up? like Opera

Re:This is irrelevant

[JCSoRocks]

I'm happy with my 4GB of RAM and 64bit Vista! Although I agree that memory is getting cheaper, and I'm fortunate enough to have a ton of it, that's still no excuse to just let programs consume as many resources as they can. I'm not saying FireFox is unreasonable, but I am saying that sloppy coding is getting far too common. We as IT / SE types need to help put a stop to some of the bloat that comes about as a result of lazy coding. ...Not everyone has a brand new machine. My mom's is 5 years old - she wants to surf the net too.

Has anyone compared to Firefox 1.5

[Dr.Who]

With Firefox 1.5.0.12 on Windows XP SP2 and six windows open the VM size is less than the Memory size.

Mem Usage Peak Mem Usage VM Size
68,388 K 69,356 K 59,276 K

I wonder if there are some shared memory segments, or memory mapped I/O.

No time to flush out with SysInternals. I have a turkey to prepare!

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

[Schmodus]

Agreed, I doubt anyone has upgraded their harddrive in order to increase their virtual memory pool.

Re:Firefox Memory Leaks, C++ Memory Leaks

[nuzak]

He's not qualified to track down leaks, since he's run across so few of them. He is perhaps qualified to rewrite the whole steaming pile with modern C++ idioms and libraries that don't leak in the first place. Some problems you can't engineer out, you can only not engineer them in.

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

[jon_burgess]

Much as I love firefox I have to say that for me it still has memory problems. It isn't the firefox process itself but the resources it hogs in X.

Right now I'm running the firefox-3.0b2pre nightly builds and firefox is taking 236MB of virual memory. X is using 3560MB and xrestop shows the firefox pixmaps as being responsible for 2714M of that.

Seamonkey

[First Circle]

About a week ago I installed the latest Seamonkey. Since then, I've been using it as my primary browser. Speed and memory consumption are like night and day vs FF2. My only real complaint at this point is the relative dearth of extensions and themes for it, but maybe that situation will improve. I have encountered few, if any, rendering issues with Seamonkey, and the overall browsing experience is as good as FF in my opinion.

I've been running the thing for days now w/o a restart, and I currently have 11 tabs open and have had about half of them open for at least 24hrs and the other half for maybe 3-4 hrs. Under Linux (Ubuntu 7.10 x64) I see:
    PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ PPID SWAP COMMAND
14334 eric 15 0 220m 113m 24m S 0 5.6 26:47.68 14330 106m seamonkey-bin

In contrast, an average session with FF2 quickly eats up 700m-900m.

Seamonkey imported all my FF bookmarks, no problems, and I have the important (to me) extensions running - NoScript, Adblock+, Forecastfox, Chatzilla, plus a retro Orbit theme, and I'm a happy browser.

Why?

[John Musbach]

Why are we worrying about ram usage when we have multiple gigs of ram these days?

Simple.

[Anonymous Coward]

When "working" and "many" turn into "done" and "any" then the general populace will be not only satisfied but will in fact jump for joy. FF is a great piece of software and memory problems are often hard to kill, but there isn't a memory problem in existence that can't be found with a good profiler, so I'm going to keep pointing at the FF dev team (which is way smaller a group of people than the group of people having memory problems, I might add) and say: Let us know when those words change.

Re:Simple.

[bunratty]

Why do you insist that every memory leak be fixed? Not all crash bugs in Firefox are fixed, so should users be dissatisfied with Firefox because sometimes it crashes? Likewise, if Firefox leaks so slowly that all users are able to run it for months without restarting, why bother with the last few leaks?

And besides, users will never be satisfied, even if all leaks are fixed. I see users complaining about normal memory use all the time, often when they mistake memory fragmentation for a memory leak, or a memory leak in an extension with a memory leak in Firefox. Some users will accuse Firefox developers of simply denying that Firefox has memory problems instead of fixing them.

Uh, how about 'focus'

[wilec]

Uh, how about 'focus', though I will agree that some time critical issues like media burning need other methods, same argument could be made for network apps, oh damn here we go again.

Wabi-Sabi
Matthew

Re:Firefox Memory Leaks, C++ Memory Leaks

[ir]

the firefox org is making millions (see previous slashdot articles). why don't they fix their own damn problems? they can certainly afford to hire programmers.

Re:This is irrelevant

[Anonymous Coward]

Also, something I noticed is that they aren't keeping any pages except the one you're on in RAM now. If you open a ton of different websites, they aren't loaded into RAM until you click their tab and then there's a massive wait (3 - 5 seconds) for it to appear. Also, if you try to close some of those tabs really fast, this delay is incurred each time and it becomes very tedious to close a large number of tabs. Granted, I know there's "Close Other Tabs" but what if you have a few you still want open, not just one. Anyways, that's a major setback IMHO, but the general speed increase and low RAM usage is definitely a plus.

Re:Firefox Memory Leaks, C++ Memory Leaks

[msuarezalvarez]

The point is, it is not their problems, but his.

He is the one who is having problems extracting value from a product (which was handed to him as a gift) when using it in a particular way, which may very well not be the one targeted by the Mozilla Foundation. Just as with every other piece of free software out there, he is very much allowed to go and fix what bothers him in order to do what he wants. If you look at the Linux source, you'll find drivers for things you've never even imagined existed, and they are there because someone wanted, in order to do their business (whatever that may be), Linux to behave in specific ways in which it did not behave, so they saw to that to be changed (they wrote it themselves, they paid someone to write it, whatever)

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

[TFGeditor]

Uh, okay, so maybe there is some hidden reason, like the parent poster is Cowboy Neal's brother-in-law, but I cannot understand for the life of me why the parent is modded "Offtopic."

Someone enlighten me.

Re:Yeah, but they're still SI prefixes.

[tzot]

By 65k MB, I tend to believe that somehow you meant 650 MB, not 65000 MB.

Re:This is irrelevant

[tzot]

I call bullshit.

You do well calling bullshit on "anyone reading slashdot already has 4GB or more", however you shouldn't be so absolute in your next statement:

desktop versions of 32 bit windows do not support more than 4GB of physical address space meaning ram usable to the OS is limited to some figure below that (exactly how much depends on the exact hardware configuration).

Since I hope that you neither omit Windows 2000 and Windows XP from the "desktop versions of 32 bit windows" set, nor are you ignorant about PAE, I have to assume that you talk about incompatibilities with drivers that make Windows report a little more than 3 GiB or RAM available even if there are more; I was of the same opinion as you are, yet a sysadm at a client site shut me up by showing me a system with 6 GiB of memory running WinXP. It was a Dell blade server, can't remember which model though. System properties showed 6 GiB of memory installed. And yes, it felt and was sluggish-- but it had more than 4 GiB of usable memory.

I agree, wrong method: should not use Memory Usage

[Anonymous Coward]

The memory usage that is measured is determined by "Memory Usage" column of Windows Task Manager. It is NOT the correct method of measuring memory usage, because it only indicated how much memory is "active", not the total memory allocated by the program.

To get how much memory allocated, the correct column is "VM Size" in XP or "Commit Charge" in Vista. Actually programs such as TaskInfo are able to reduce the "Memory Usage" of any program (e.g. 5 MB only for Firefox 2). Hence, "Memory Usage" is not a good measure for memory usage.

A better memory comparison test is available at:
http://www.kejut.com/browsermemory

Re:This is irrelevant

[petermgreen]

Since I hope that you neither omit Windows 2000 and Windows XP from the "desktop versions of 32 bit windows" set, nor are you ignorant about PAE
I quote from that MS link

"Windows XP (all versions) 4 GB of physical RAM*"

"* Total physical address space is limited to 4 GB on these versions of Windows."

Since it is the total physical address space that is limited usable ram ends up somewhere between 3GB and 4GB depending on how much space the chipset reserves for IO.

2000 advanced server and all editions of server 2003 do support more ram but I do not consider those desktop operating systems

a sysadm at a client site shut me up by showing me a system with 6 GiB of memory running WinXP. It was a Dell blade server, can't remember which model though. System properties showed 6 GiB of memory installed. And yes, it felt and was sluggish-- but it had more than 4 GiB of usable memory."
That finding contradicts the MS page you linked to and all experiance I have seen from people attempting to use XP32 on boxes with more than 4GB of ram.

It sounds like either it was running XP64 or someone had hacked out the limitation.

Re:This is irrelevant

[tzot]

Like I said, I was of the same opinion like you and that was the one case that I was contradicted. I forwarded to that sysadm (hopefully his email address is still valid, haven't had contact with him for some time) the link for the article, he might reply himself or back to me; if the latter, I'll copy the text here.