Firefox, Chrome virtually tied for JavaScript speed

Google's Chrome now is only a smidgen slower than Mozilla's Firefox on the SunSpider test of JavaScript.

(Credit: CNET News)

On Tuesday, Mozilla released Firefox 3.1 beta 2 and Google released Chrome 0.4.154.33, so it's time for the latest installment of JavaScript performance testing.

Here's the highlight: Though Firefox remains the leader on the SunSpider test, with a score of 2,110, Chrome edged very close with 2,140. A lower score is better; because of some variation in results, the numbers I quoted are an average of several runs.

Firefox and Chrome aren't the only browsers out there, but they're interesting to compare for a few reasons. First, they're both open-source projects launched to shake up the establishment with new ideas about the browsing experience. Second, given that philosophical alignment, they're likely to appeal to the same early-adopter crowd. Finally, both have new JavaScript engines, Chrome's V8 and Mozilla's TraceMonkey, which in the new beta is switched on by default.

JavaScript is used to build sophisticated Web sites such as Gmail or Google Docs, but it's also widely used for more ordinary operations, so faster JavaScript performance is desirable. One interesting possibility Google has raised for Web applications though is to bypass JavaScript altogether and use Google's new Native Client software, a research project that lets Web-based software run closer to the speeds of regular software on a computer.

Chrome is making steady gains in Google's JavaScript test; Firefox is much slower and has a mixed record.

(Credit: CNET News)

SunSpider is only one test, though; Google has its own JavaScript benchmark on which Chrome wins hands-down. A glitch in the first Firefox beta kept me from testing it on Google's benchmarks, but the new beta runs again, yielding a score of 182. That's lower than the earlier Firefox 3.1 beta's 235 score, so perhaps something is still amiss. Either way, it's a far cry from Chrome 0.4.154.33's score of 2,635.

The usual caveats: your mileage may vary; I ran these tests on a dual-core Lenovo T61 laptop with 3GB of memory and Windows XP. JavaScript is only one aspect of Web browsing performance, and indeed of browsers overall. Also, this software is still in beta, Chrome in particular a developer beta. Finally, I apologize to those who've been asking, but time constraints have kept me from trying the latest WebKit builds and Opera.

According to Mark Larson, Google's Chrome program manager, Chrome 0.4.154.33 fixes a crash when opening the Options dialog box on 64-bit Windows and some issues using Hotmail. "Hotmail still does not properly recognize Google Chrome," though, Larson said in his announcement of the new version, though it can be fooled into thinking it's using a more mainstream browser. For details, check the instructions on the release notes.

[manalagi001]

Since you are testing pre-release browsers, why not test the nightly build of WebKit as well, from webkit.org? Webkit is the foundation of Safari and Chrome. Chrome uses a different JavaScript engine though, so a test would be in order here.

You may find that WebKit is even faster than Chrome and FF3.1b, which bodes well for future releases of Safari (and possibly Chrome).

[samkass]

He mentions at the end of the article that he "doesn't have time"... of course, since Safari is the only one on there where he DIDN'T take the 10 minutes to download the latest version, one wonders how much time it would have taken.

Just leave Safari out of the comparison if you're not going to update it!

[Shankland]

@samkass. Sorry for running the earlier numbers on Safari. I figured older numbers were better than none. I re-ran the Safari test today with vesrion 3.2.1 and got the following average score on SunSpider:4,906. So it's a notch better indeed, though not up to the competitors' scores, at least on my machine.

[tundey]

I hope these guys are not focusing solely on JavaScript speed at the expense of other speed improvements (like memory consumption in Firefox and plug-ins for Chrome)

Also, a minor quibble. When you have a separate bar for each version of the browsers, can you put the name of the version right inside the bar? Hard to reconcile all the bar colors with the different browsers.

[jilsonj]

they should have tested opera 10 as well.

[Clarious]

Java script performance has never been Opera strong point, so it is not neccessary to test it.

[markur69]

I concur with the WebKit nightly build. I've been using for about 3 or 4 months in Production environment and its way faster than all the above and more compatible with our own internal testing for XML and Flash... Chrome is barely a Beta and it gets this kind of play just because its Google? Common CNET... Do better testing. Opera also deserves to be considered as it has been around a long time and is competitive in speed and features

[RobWLA]

Google's Chrome is so fast, stable & easy to use - that I find I'm rediscovering the pleasures of web browsing and also rediscovering the pleasure of using my older, slower IBM Thinkpads.

With their slower clock speeds and Internet Explorer crashing so often, these pc's were quickly becoming the bane of my existence.

Not any more.

For every release where Chrome gets better - Windows is fading farther and farther into the background - where it belongs. I always felt it was a shame to spend top-dollar on IBM Thinkpads, only to saddle them with a turkey of an operating system.

Now - I find myself forgetting that I'm even running Windows - and I'm reconsidering my decision to dump this older gear.

Not bad for a free browser.

Rob

[The User]

Rob, let me point out that Chrome still runs on Windows. I don't really see why would it fade into the background just because of some applications running on Windows. Doesn't make sense.

[RobWLA]

Reply to TheUser:

My goal is to use less and less Windows-supplied products as I move forward. Since I don't have the time and energy to move to Linux - this provides my way forward.

You are correct - Chrome still runs on Windows - but Chrome dispenses with 90% of my actual system crashes (as probably Firefox et al would as well). Many times I can't even get windows to terminate a balky IE session gracefully without repeatedly hitting "End Now" and seeing multiple Windows system responses to such efforts.

Moving to Chrome effectively allows me to take one big step away from the annoying Windows experience - no doubt about it.

As I do not agree with Microsoft's design philosophy of bloated products (and bloated methods of dealing with their misbehavior) - replacing Windows applications (from Word to IExplorer) one-by-one over time - effectively reduces my involvement with Windows.

Recently I've moved from Hotmail to Gmail, from IExplorer to Chrome, and eventually I hope to move from Office to Google's or IBM's new web-based office-offerings.

I expect all of these to be less bloated, more friendly and hopefully more stable. We'll see. If so - then Windows inherently will present a more stable - and less intrusive - user experience for me.

This is why I refer to Windows as "fading" into the background - and not "disappearing".

Whether or not all these new products prove less bloated, and more stable, only time will tell.

So my goal of using less Windows products - while still running Windows - will hopefully allow me to achieve greater system stability and ease of use.

[rapier1]

RobWLA

By 'WIndows' product do you mean 'Microsoft' product? Windows is the environment, Microsoft is the company. Companies create products. Environments are where those products are run.

[joetesta70]

Safari is SLOW. That POS.

[tm_anon]

Or you could stop with your google advertisement and take a look at a different OS, such as Linux. I've done plenty of research on Chrome and I've yet to see a reason to switch. Unless they take Google out of Chrome so that my web stats are mine alone I won't even think about using it. A web browser is meant to browse the web, not spy on you. As for Google coming up with its own alternative to java, of course it's going to run faster on its own browser. They have access before anyone else does. They know exactly how the software runs and can build tailor made access directly into the browser, meaning there's no way for FF to compete. Using JavaScript as a benchmark means it's a fair test, nobody has more access than anybody else in the browser wars. Faster single page loading times are a bad way to gauge the abilities of a new browser anyway. What about multitasking? What about expandibility? What about stability? It's great if your browser has all of these things AND speed, but if all it has is speed, then why bother mentioning it? My browser works well, is as fast as any current, non beta version, browser, but it also is a multitasker, checking my email, all my social networks, allowing me to post to my blog without going to my blog, letting me interface directly with every social network I can find, checking my RSS feeds and allowing me close to the same expandibility as firefox currently has because it is based off the same core. I saw 3 articles on this browser through CNET more than 6 months ago, but haven't seen any since Google came out with Chrome. Flock is better than Chrome for several reasons, the main reason being it's not a one trick browser. It allows me to do only the part of web browsing I enjoy, doing all the rest for me. I don't have to go searching various sites for new stories as it brings them to me. I don't have to check my email as it is already built into the browser. I don't have to check the various social networks I'm part of just in case somebody wrote something. I get alerted automatically. I can post pictures to my facebook, myspace, photobucket and many more other accounts without ever going to those various sites, I can look through the pictures my friends have posted without going to their own sites and I can do both without ever leaving the story I was reading. Chrome is for those who want to do everything one bit at a time the Google way. FF is for those who want to do everything one bit at a time their own way. Flock is for those who realize doing things one bit at a time is ridiculous. I do it all at once without ever getting confused, without opening dozens of tabs, without worrying that my browser will crash and without having to keep a dozen thoughts in my head. There's a place for different types of browsers. Flock isn't made for opening up, going to one site, then closing it down, Chrome is, FF is. That's not how I use the internet and I suspect it's not how the majority of users use the internet either. Maybe the future of browsing should try to do something the other browsers haven't.

[MadLyb]

It's called a Line Return.

Next time, I might read your comments if use one every now and then.

[The User]

Now, what's the point of Chome now?

[Clarious]

Replacing IE for normal user, Google has more money and power than Mozilla, that is why it can do it.

[ChrisLang]

It runs on Linux.....

[morlamweb]

Could you please put a caption on the graphs' y-axis that explains just what the numbers mean? For example, what exactly does a score of 2110 on the sunSpider test mean? Is that the number of seconds it took to run the test, or perhaps milliseconds? Or is it a totally meaningless made-up number? And those Google JS tests - apparently a higher "score" is better here, vs. a lower score for the SunSpider tests. What's the unit of measurement for the google tests?

[ktswami]

Come on, Stephen, you do this Chrome/FF circus act and exclude Opera 10 and the lastest nightly Webkit?

Well, here's what it would've looked like.
http://www.favbrowser.com/performance-firefox-vs-safari-vs-google-chrome-vs-opera/

And note that Opera's MAMA shows that ~3.5% of the web use XMLHttpRequest.
http://dev.opera.com/articles/view/mama-key-findings/#xmlhttprequest

[queticomn]

Haha, the usual BIAS c-net does not include Opera Alpha 10.0. an they problye compared FF 3.0.2b with a bunch of add ons installed which chrome doest support.

I dont have the energy or time to mess with windows operating system disinfecting it after infected with its trogens, virii, worms, malware. Yes all stick to Novell's Suse Desktop Enterprize for my Tower. much more reliable then windows.

Bias Bias Bias. Screw google an use the scroogle scrapper at http://www.scroogle.org/cgi-bin/scraper.htm

[ArnieAZ]

I ran the SunSpider test on:
Safari 3.2.1 the current release (3.1.2) result = 2678.6ms
Camino 1.6.5 = 8770.6
Webkit daily build (12/08/08) = 707.8 yes! really fast!!

[AppleSuxLeo]

This article is flawed as it only runs one test. For a real thorough set of tests , see ExtremeTech.
On a battery of tests , Chrome was faster by far. Cnet is for amateurs.

[AppleSuxLeo]

Chrome by a landslide.http://www.extremetech.com/article2/0,2845,2335242,00.asp

[loose_screw]

I have the latest versions of IE (and Maxthon), FireFox, Opera, and Chrome loaded on my ThinkPad. Guess which one is my default browser? Chrome. It's been fast, stable, and intuitive like no other browser I've ever used. It truly makes web browsing pleasurable again.

[RobWLA]

You said it.

See my post above regarding how Chrome single-handedly made me reconsider dumping out my older, slower (but much beloved) Thinkpads.

After years of being flummoxed by endlessly bloated software products from you-know-who, we're finally getting some alternatives here - and Chrome is the reason.

My old Thinkpads are faster, more stable and fun to use like never before because of this product. Can't wait to try Google's web-based office apps on Chrome too.

I've been in IT for 30 years (applications developer - financial services) and am completely fed-up with needless product complexity & dubious "features" - designed from the get-go by Microsoft - just to upstage the competitor's products using a bunch of flashy song & dance routines.

Rob

[Shankland]

There are some limits to my graphing software (right now I'm using Google Spreadsheet because it has shiny colors), but I'll see if I can do better in the future. Thanks for the comment.

[firefoxluva95]

I ran a test on Opera and it came around 8072 ms. Firefox and Chrome were really close, Firefox did seem slightly better. Of course the slowest was IE 8. Chrome is less stable than Firefox 3.1 beta though and hopefully that will be fixed. Also, I noticed that the tabs hang around as processes, even when I closed them. Eventually, Chrome will turn into a hog too. I'd like to see it when it has all those extensions and themes. To rule out lurking variables, I'm running all these on Vista SP1. Safari is just, well, I'm happier with it on my iTouch than on Windows.

[formido]

ExtremeTech said WebKit uses V8. They kind of lose a lot of credibility.

[jordanaustino]

I routinely get an 1880 with chrome... and 2140 with firefox 3 b2 on the SunSpider test, ive run it like three times... so these tests might be messed up

[Jyakotu]

Opera's JavaScript speed is not an issue with me. It still preforms JavaScript pretty fast IMO. I can tell the speed improvements with a Opera 10.00 Alpha 1 build.