Slashdot Mirror
Mozilla Unleashes JaegerMonkey Enabled Firefox 4
An anonymous reader writes "Mozilla has published the first Firefox 4 build that integrates a new JavaScript engine that aims to match the performance in IE9 and reduces the gap to Safari, Opera and Chrome. This is really the big news we have been waiting for all along with Firefox 4 and it appears that the JavaScript performance is pretty dramatic and seems to beat IE9 at least as far as ConceivablyTech shows. Good to see Mozilla back in the game." The Mozilla blog gives a good overview of the improvements this brings; Tom's Hardware also covers the release.
Ironically, the primary site for which I really need a faster Javascript engine is Slashdot. For a heavily-commented article I switch to Chrome.
My understanding of the term Beta is that all features are complete. Has something changed?
Anyone else kinda sad that now Firefox is playing catchup. When no one cared about JS performance, the Open Source crowd was king, then all of a sudden big corporate money was poured into JS performance and now FOSS is lagging behind.
It seems that FOSS can't compete head to head with corporate backed projects, if the corporation actually cares. For example, MS didn't care about JS performance in IE6/IE7 and Firefox was king. Now, Microsoft is trying to compete in the browser space again and IE9 is catching up in features and exceeding Firefox in certain respects.
This is coming from a very long time Firefox user, but I have definitely switched to Chrome for general web browsing. I stick with Firefox for development though because of the large amount of niche plugins specifically tailored for development.
Who cares? You can't make a guess at answering that question? Okay I'll give you the answer: everyone but you.
If this is "really the big news we have been waiting for all along," then we can officially proclaim this as a Slow News Week!. Who didn't expect Firefox 4 to beat IE9 and narrow the performance gap to Safari, Opera, Chrome? Wake me up when Firefox 4 blows them all out of the water!
I need trepanation like I need a hole in the head.
And cheers to the release!
There's no -1 for "I don't get it."
Install ad-block and noscript.
Firefox lagged chrome mostly because firefox cares a LOT more about compatibility, and adding all this crazy JIT compiled JS stuff is hard when you're trying to support all the introspection features which people have been using in firefox.
Anyone know what hardware they were using in the demo video (tom's hardware) to get the 12fps and 91fps comparatively. Is this the kind of performance increase the average user will see or just people with high end systems?
Not exactly everyone but you. I'd guess anyone, like me, who blocks almost all Javascript is going to find an improvement in it's execution speed fairly irrelevant.
Slashdot is quite perky with the last couple of betas. But it's especially disheartening that the video "upgrades" in this most current release fall short on my platform. When viewing the demo page ( http://demos.hacks.mozilla.org/openweb/HWACCEL/ ), I get 1 fps. I get 6 fps when running the same demo on Firefox inside a Parallels Windows XP SP3 VM. The VM is significantly faster... which boggles the mind actually.
So far as I remember, this was an Apple issue not necessarily a Mozilla issue, but still disappointing.
Are you joking?
Check out http://arewefastyet.com/ to see the speeds of several JavaScript engines compared to Mozilla's.
A /. viewer who uses firefox, but hasn't heard of ad-block or noscript?
Todo list:
[ ] Turn in geek card
[ ] Write a will
[ ] Buy shotgun
The linked article is about 4.0b6-pre which is the first version to include JaegerMonkey. The other two links are to articles about the public release of 4.0b5, which doesn't include JM (it's headline feature is really the DirectDraw support on Windows).
4.06-pre isn't currently being pushed to regular beta testers AFAIK.
---- Den ene knappen er powerknapp, den andre er Bender voice knapp "Bite My Shiny Metal Ass"
Kinda surprising that someone around here doesn't get the difference between Javascript and Java.
The acid2 test has the nose of the acid face half a pixel too high.
Is the multiprocessor functionality of chrome optional? That is, I like the fact that a bad page ( usually bad java script or flash ) can only hang one core and not my entire processor.
You mean Javascript? ;)
I care a lot about it. As a partial JS developer it means I can do more without lagging the heck out of my users. I can create more complicated fun stuff instead of doing bare minimum. I can run that extra DOM check that I was scared to do previously looking for items that match some jQuery string. There's just so much more than that and those just sound like the "lazy" stuff.
(It's not about laziness of coding... more: "I can't add this because..." -- I do interactive stuff in Flash/JS/etc. I welcome faster scripting in any language.)
Every time I start to have faith in humanity, I ruin it by driving to work between 7 and 8 am.
YHBT?
Switch back to Slashdot's D1 system.
Unfortunately, there are a bunch of business and finance websites that I use that have JavaShit all over the fucking thing and the only way to see their content is to have JavaShit turned on. And then you have these websites that have so many JavaShit shit on their websites that you can't click on things because they have these JavaShitty popup bullshit that gets in the way.
I think there should be an operating system written in JavaShit that web developers would have to use by law - JavaShit OS. Build it on top of MS DOS. That'll slow'em down.
RIP America
July 4, 1776 - September 11, 2001
Also does the FF3.0 JS engine work on AMD64 linux yet?
All of the above was encrypted with a Quad ROT-13 method. Unauthorized decryption is in violation of the DMCA.
They're getting pretty close to matching chrome and safari: http://arewefastyet.com/ ... and getting there without breaking backwards compatibility horribly like chrome and safari have.
Agreed. They are working on multiprocess, it's called Electrolysis. It's very quiet, so I imagine they're behind schedule. It's also my impression that it's a very small team.
What are you talking about? I've never had a problem with Firefox on AMD64, even with 32-bit support disabled in my kernel.
adblock is a total must-have, IMHO. Also a good feature about firefox is that you can trust it to not send all your traffic informations to Google...
The Wise adapts himself to the world. The Fool adapts the world to himself. Therefore, all progress depends on the Fool.
the only thing Firefox has going for it is adblock and the huge extension repository. Even then, its debatable
Apparently access to source code and the ability to be compiled and run on platforms like BSD and Solaris doesn't count for anything any more.
...for their acceleration. Pretty friggin sad.
It is a MOZILLA issue.
Chrome does not support the HTTPS Everywhere extension. Once it does, I may switch. But privacy is my first concern when using the web.
Buckle your ROFL belt, we're in for some LOLs.
Or Trolling.
So, they introduce a faster Javascript engine. Good for them; they've got a working prototype/alpha of 1/4th of the necessary features to catch up with everyone else, at this point:
* multiprocess functionality
* security mechanisms resulting from said multiprocess functionality
* better thread/tab/etc. management
At this point, the only thing Firefox has going for it is adblock and the huge extension repository. Even then, its debatable: Chrome, for instance, seems to implement most of the extensions I used natively, and does it better than Firefox extensions did, to boot. (Most of which were only necessary to make up for FF shortcomings, like crashing.)
Honestly, the very first thing FF should be working on is multiprocess shit. It's big, bloated, and at this point, somewhat archaic in architecture - the code base is over 12 years old, isn't it?
Perhaps they should just ditch the Gecko/GTK combo and adopt Qt, Webkit and V8, what's the point in trying to compete when they could just assimilate :).
IE 9 betas, I think.
Unfortunately, this is really needed to address the firefox "memory problems". Leaky extensions and the like are a fact of life. With chrome, it's a matter of bringing up the chrome task manager and kill the offending web page. With firefox, I have to restart the entire browser.
The point is that multiprocess, except for plugins (which is already done) isn't a big advantage at all.
They actually implemented it on Fennec, that's Firefox mobile if you prefer, because it would yield an advantage here.
On regular desktops, not so much, in fact, it uses quite some memory. It's not because "others do it" that it's necessarily "teh future embrace or die!".
It also encourage sloppy programming since it's more fault tolerant, chrome tabs crash all the damn time in comparison to firefox which barely ever crashes.
Non-plugin code should _never_ crash, ideally.
adblock is a total must-have, IMHO. Also a good feature about firefox is that you can trust it to not send all your traffic informations to Google...
AdBlock exists for Chrome.
I wish I could mod this flamebait. The early preview speed tests for IE9 compete with Chrome very well.
actually chrome has a real adblock now
but i'd still stick to firefox for other reasons
I hate to break it to you but that's not very common.
To pedantically expand:
Who cares? Everybody except the exceedingly small segment of the Internet populace who does not now, and never ever will, use javascript extensively (either as a consumer or as a development).
According to http://arewefastyet.com/ they have come from being 2x-3x slower than Safari and Chrome a few months ago (v8/sunspider benchmarks) to being within a few percent to 25% of Safari and Chrome, depending on the benchmark.
I think that's pretty impressive - it basically puts them in the right ball game now, and narrows the performance gap to the barely noticeable range for most practical purposes.
It's called competition, and it's what we really, really need in the software industry.
Like all pain, suffering is a signal that something isn't right
Press F11.
I never understood the demand for "tabs on the title bar", personally. It's just 30 pixels (using Win 7, no scaling).
How many toolbars could you possibly have that you can't spare the 30 pixels?
And it usually downloads the adverts before deciding to hide them
-Docvert converts MSWord to OpenDocument, clean HTML
anyone else not really care about Java performance? it works for me, thats all I care about. Until the other browsers have adblock and Noscript they are all dead to me.
Yeah. I'm running Lynx as my default browser. I don't care about JS performance, because I never visit any site designed after 1995.
Web apps? Who needs 'em! Why use a web app when you can download some source and compile your own binary!
Web 2.0? Hah! I'm still on Web 0.3 beta, and I'm lovin' it.
If they do decide to play follow the leader with the tab bar, I hope they include the option to have it the normal way. The taps-at-the-top UI bugs me as I then have to move my mouse that much more to switch tabs.
upon the advice of my lawyer, i have no sig at this time
As a partial JS developer
Which part? Or should I better not ask?
That makes no sense. Unless you were referring to opening [N cores] pages.
-]Phreak Out[-
A /. viewer who uses firefox, but hasn't heard of ad-block or noscript?
Todo list:
[ ] Turn in geek card
[ ] Write a will
[ ] Buy shotgun
you missed a step ;)
It's not completely quiet, the Fennec (Firefox Mobile) beta or alpha version already has it. And they are working on improving on it. Multiprocess is very high on my list though, It it is probably the feature I want the most.
The trick is Firefox is the browser which uses the least amount of memory by some test (there are very little tests being done). This is where multiprocess is going to be interresting to watch, I think that is why they started on mobile.
New things are always on the horizon
Yes, I've had Chrome tabs crash on me many times as well. Actually Firefox Betas are more stable.
New things are always on the horizon
Big advantages for me outside of isolating crashes:
1. Isolating misbehaving webpages from the rest of the browser. In Firefox, runaway Javascript on one tab can slow down or freeze other tabs or the whole browser.
2. Solve the memory leak from fragmentation problem. Closing tabs in Firefox often doesn't decrease the amount of memory consumed.
you missed a step ;)
[ ] ???
alternatively have a can of this drink
You can't handle the truth.
Precisely. It's taking more time for Mozilla to do it because they're not doing it the easy way. They could've had it done quite some time ago if they just made them completely separate processes with a third party task bar. However they're going beyond that. They're pooling some of the resources to hopefully keep the bloat down, and trying to focus the separation on portions of the process which really need it.
The option is available. Right-click a toolbar and uncheck "Tabs on top".
The shareholder is always right.
What?
I have 8 cores. I open 5 tabs, creating 5 processes. One tab crashes. One process crashes. The remaining 4 remain as they are; usually, it's quite simple to regain the crashed page by loading it.
With stock firefox, that usually means pulling each of those 5 pages out of history again, after restarting the whole browser.
As far as "hanging one core and not the whole processor" you do realize that in a modern operating system, processes are not inextricably linked to a core? If your whole system locks due to the browser hanging, that's poor system design (Windows, Mac) or failing hardware.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
Still "hidden" I am afraid - as far as i know. I heard rumor that there's a native-and-real adblock, but I've yet to bother to seek it out.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
I'd agree there, sometimes. I've run Chrome for weeks without a single hickup, as well.
Unlike Firefox crashing, which often requires a 'reboot' lasting up to a minute on a modern machine (if you're using session manager or the like), a single tab requires ctl-R to be hit, once - and it's instantly back (often/sometimes with session data). That's stock, mind you. Firefox, on the other hand, records no such thing, and even with Session Manager will often lose session data (whether it's opened tabs or things typed in the crashed windows).
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
OK, you're right: I should not have said "bloated" and "big". I should have said "slow as fuck and bloated".
Leave your browser open for the better part of the day, and let me know how Firefox performs (vs. Chrome). Night and day difference.
I can't even watch 720p stuff on some of my computers with Firefox - not if I want to be doing anything else. (These are single core systems.) Chrome handles much better in the same (limited memory) environment.
That said, recent Chromium builds have had some pretty severe memory leaks.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
I've downloaded "noscript", but I allow scripts globally, and mark specific websites as untrusted. More websites these days require javascript enabled for best use.
It's just 30 pixels
Common screen resolution on netbooks: 1024x600
Seriously, and that same type of person would complain at the same time how bloated the browser is. It's like talking to a wall.
"Those who would sacrifice essential liberties for a little temporary safety deserve neither liberty nor safety." - BenF
Except that chrome just doesn't let you do that for everything. Even the adblock page says "a few resources might still load before AdBlock can get to them".
Sure. On the other hand, competition is what got us this far. Having only one html engine or js engine that actually improves isn't necessarily a good thing.
It's not about the mouse distance traveled - it in fact takes LESS time to pick tabs that are on top because you can slam your mouse to the top of the screen - you don't need to spend the extra time precisely choosing your Y coordinate, only the X. Same reason it's nice that in Windows, the close button (even though it doesn't physically touch it) can still be hit by slamming your mouse into the upper right of the screen. It's basically a gesture at that point, not a precise move and then click.
WHO NEEDS SHIFT WHEN YOU HAVE CAPSLOCK/ DAMN1
I don't use any toolbars - I just like more viewing space. There's literally nothing on the top bar except the menu button and close/minimize buttons - why not use that space for tabs like Chrome? The title of the tab is the same title that'd be in the titlebar anyway, so you don't really need a titlebar. And yeah, on netbooks, Firefox can be really annoying on some sites.
The main thing, however, is not even the screen space - it's the fact that it's even FASTER (despite your mouse traveling a smidge further) to have the tabs at the very top. This is because it doesn't require precise Y-axis movement to within a 30-pixel-tall space to choose a tab, you just slam your mouse to the top and make sure the X-coordinate is within the 150px or so that the tab takes up.
WHO NEEDS SHIFT WHEN YOU HAVE CAPSLOCK/ DAMN1
The two last links have nothing to do with JaegerMonkey (and only Direct2D + HTML5 Audio), so they are not links giving on overview of this feature. They discuss the latest beta release, not the merge with JaegerMonkey on the trunk / nightly builds.
Beware: In C++, your friends can see your privates!
If your whole system locks due to the browser hanging, that's poor system design (Windows, Mac) or failing hardware.
Or your IE box just got owned and spyware needs to be removed from it.
I'm running the latest build right now (on OSX) and the pure JavaScript C64 emulator is still quite a bit slower than in Chrome but it's still a LOT faster than previous Firefox versions.
My other account has a 3-digit UID.
[ ] Purchase a life insurance policy payable to the FSF?
Why oh why did they move the refresh button to the right side of the address bar? That's so many pixels away...
Most folk'll never lose a toe, and then again some folk'll...
Firefox 4.0 b6 has not been released yet. However, the pre-release versions (Minefield) are available here
2019 is going to be the year of Linux on the desktop.
Several websites I deal with won't even work with Chrome, Bleacher Report and the PogoPlug GUI are a couple of 'em. Don't know why I even left FireFox in the first place. Geez!
I'm pretty sure there's a name for this kind of argument, but it's basically a regular fallacy.
Firefox JS performance has nothing to do with being open source. Javascript did not matter for a very long time, it was only used for seldom tiny effects. That's why Firefox JS engine was not super high performance (albeit still much faster than anyone elses at that time), because it would have been seen as totally overkill.
Things have changed and Firefox JS is adapting. Since it's using an old codebase, it's not as quick (it took Apple quite some time to give birth to Nitro, and Google just plain bought V8 tech - as for IE9, we all know it took them time too, it's not yet there.)
And that's all there is. So much for the anti open source FUD btw.
Also, it pays to look at the development trends. In preparation for Firefox 4.0, Mozilla has been working to improve their tracing engine (called Spidermonkey I believe), and at the same time develop and then tune a JIT compiler. Mozilla has designed these efforts to be complementary, so that they could be merged to get the best of both approaches. The combined javascript engine has been dubbed JaegerMonkey.
Mozilla development first merged these two different methods into the one engine only about two weeks ago, and now they are tuning it. Here is their record of how they have been going:
http://arewefastyet.com/
The trend is very interesting, is it not? The very first test of the combined engine, JaegerMonkey, was between a 10% and 35% improvement, depending on the benchmark.
If the trend holds true to form, JaegerMonkey will be the fastest javascript engine of all in somewhere between one and three weeks time.
So much for the FUD that "FOSS always lags behind". Due to JaegerMonkey and hardware accelerated rendering, Firefox 4.0 has a decent shot at becoming once again the speed king of all browsers.
Except that chrome just doesn't let you do that for everything. Even the adblock page says "a few resources might still load before AdBlock can get to them".
Part of that is actually due to the speed of the rendering and how the extension API is handled. Sometimes extensions just can't get to elements before the browser renders them (in Google's mind the browsers primary function comes first, a good step imo, to prevent all the bloated crap that exists in some FF addons).
I would rather they focus on making things other than JavaShit faster, i try to leave it switched off when I can. I don't use Slashdot 2.0 or Web 2.0 apps so JavaShit speed is pretty much irrelevant to me.
You being the minority exception.
There are several bugs that make FF a pain to use for me (although I still use it because of the plugins).
* Bug 490122 - Firefox periodically becomes unresponsive/freezes: video jerks/pauses/halts; links, tabs, menus stop responding.
* Memory leaks: FF memory use grows to ~1.5GB and then it hangs with 100% CPU usage.
it may be a plugin or GM script issue but I could not isolate it. Seems to happen more frequently when browsing picture-heavy pages (google image seach, galleries, etc.)
Get a descent GC, people! There is no such thing as a memory leak in JS. If there is, the GC isn't good enough.
I know tobacco is bad for you, so I smoke weed with crack.
When Opera released the new JIT engine, Carakan, it was well ahead of the rest with the first public release. No bragging and empty claims beforehand. Just actually delivering on performance. Mozilla, on the other hand, is still mostly talking and bragging...
Clever signature text goes here.
A GC doesn't help you if FF or an extension continually allocates memory and adds it to a persistent data structure, like a list. Most GC's also don't address memory fragmentation issues, which prevent process size shrinkage.
Chrome's approach of having a separate process per tab will help you with both of the above. Yes, it's arguably kludgy, but it's a great workaround. It just works.
Heap partitioning can help combat fragmentation and GC overhead (not to mention many modern GCs handle fragmentation better than the underlying OS, which is all you have with traditional memory management). Data persistence bugs are to be fixed, not worked around, and some better memory profiling tools would be a great help.
I know tobacco is bad for you, so I smoke weed with crack.
You are right, I should have said it that way.
I have a 3 core processor, and typically have about 10 pages open.