Chrome currently fastest browser, according to benchmark

Futuremark, the company known for its benchmarking software, has released a new version of its Peacekeeper software that's made specifically to test the performance of web browsers. This new version has been made to use the latest HTML5 standards. The program can be used on any Internet connected device which means browsers on the PC, Mac, smartphones and tablets can be measured.

Futuremark states that since the first version of the Peacekeeper was released in March 2009, the program has tested 3140 different browser versions. While Apple's Safari browser was the fastest when the benchmark launched, Futuremark says that Safari is now the slowest of the web browsers that are tested on the PC and Mac platform. Google's Chrome has been the fastest web browser tested on Peacekeeper since May 2009 for the PC and Mac. However, it notes that the Opera web browser has been closing the speed gap.

The most popular web browser among those who have used Peacekeeper is Mozilla's Firefox. Nearly 30 percent of all of the software's benchmark results come from versions of Firefox.

In terms of their browser performance, Apple's iPad 2 is the fastest among tablets and smartphones, according to Peacekeeper. The Samsung Galaxy Tab 10.1 tablet is second followed by the Acer Iconia W500 in third place, Apple's iPhone 4S in fourth place and the Samsung Galaxy S2 smartphone in fifth place.

Report a problem with article
Previous Story

Apple's policies on leaking information leaked

Next Story

Windows 8 Metro to change Control Panel to PC Settings?

72 Comments

Commenting is disabled on this article.

All browsers these days are fast enough.

Speed used to be a problem for some browsers. Specially for internal apps. We had an internal application here where the user had the options to show all results from a search without any pagination. There could be thousand of results presented in a table.

This page was slow to load and scrolling was extremely laggy under IE 6 and 7. IE 6 could take up to 1 minute to draw the page and scrolling was a slide show. Opera was taking about 10-15 secondes to draw it and scrolling was smoth. Firefox (think it was version 2 not sure though) was taking around 20-30 secondes to draw it and scrolling while not smoth was not a slide show.

This feature was simply not usable under IE 6. When the client told us it was too slow we had no choice but to answer the problem was his browser and not the application and if he wanted to use the feature to show all results from a search without pagination he had to use another browser (the page was a simple well coded table without any js or special css positioning).

Back then speed did actually mater for internal apps where clients had specific needs you generally don't find in well designed public web applications avalaible over internet.

But today all browser are close enough.

It's all about preference, even with benchmarks. My preference lies with quick load up times, something Firefox has never had during the entire time I've used it. Takes at least 4 seconds to start for me, and what feels like years on older computers. This is why I don't put Firefox on anyone's computer anymore too.

Maybe it may not seem like such a big deal to others, but it's important to me at least. I just like responsiveness...

1st Time Chrome User as of 3 days ago. Long time Firefox User.

I stayed away from chrome because when I first tried it, there were no extensions I needed and Roboform didn't work with it.

I installed versions 15 with adblock, roboform, and 4 or 5 other extensions. My firefox has about 15 to 20 extensions which I'm very comfortable with.

Overall, Chrome is MUCH faster than Firefox. That said, I miss some of the extensions like Noscript, using my fonts always, security tweaks etc. Chrome has definitely come a long way but it's still only my general news viewer, the "work" gets done from Firefox.

I use LogMeIn everyday, all day in my browser which is FireFox, the LogMeIn Extension is simply amazing and makes things such as full screen work perfectly. I have tried this in Chrome but for some reason LogMeIn does not work as well, maybe they need someone to make a better LogMeIn Extension for Chrome? who knows...

Yes i will admit that when i first turn my PC on, FireFox can take about 15-20 second to open up, but considering all the extensions ive got, its worth it!

Ive never "got" what all the fuss is about Chrome, and maybe im being stupid here but i really do think its made for people that just want to browse the internet. Not developers / people like me who like to customize everything about their browser? /puts flame suit on

WarioTBH said,
I use LogMeIn everyday, all day in my browser which is FireFox, the LogMeIn Extension is simply amazing and makes things such as full screen work perfectly. I have tried this in Chrome but for some reason LogMeIn does not work as well, maybe they need someone to make a better LogMeIn Extension for Chrome? who knows...

Yes i will admit that when i first turn my PC on, FireFox can take about 15-20 second to open up, but considering all the extensions ive got, its worth it!

Ive never "got" what all the fuss is about Chrome, and maybe im being stupid here but i really do think its made for people that just want to browse the internet. Not developers / people like me who like to customize everything about their browser? /puts flame suit on

Chrome has about as many addons as Firefox does - plus it's even more developer friendly because it can run userscripts without the need for an additional add on.

Wakers said,

Chrome has about as many addons as Firefox does - plus it's even more developer friendly because it can run userscripts without the need for an additional add on.

Maybe so but for LogMeIn, its still useless and thats pretty bad imo.

I've heard about this benchmark. I read an article by someone from Mozilla IIRC when it was released. They said the test wasn't very extensive. Things might have changed since then, though.

But I like the educational aspect of it. Even people that don't understand Web technologies can see what the test is doing and have it explained to them.

well it finished before 5 min on ie9 hmm, firefox 8 on desktop is dam stupid and not finishing on second try aswell.

im not going to trust a benchmark who starts with false info says five minutes to finish yet its 20 mins passs and still it isnt half.

now iv done sunspider and cause i only use 2 firefox and ie9 ie9 is faster on that.
and in my SGSII phone stock browser is slower than iphone 4s with few points but using opera it beats the **** out of it 1600 points and using firefox 8 i get 1400..

I've actually just started using Safari for Windows because someone recommended the new version to me. I'm very impressed. It certainly doesn't feel "slow" - although I don't think it is quite as fast as Chrome at loading javascript content. Then again, it doesn't have any of the annoying issues that Chrome currently has.

I would be using Opera but since upgrading to 11.52 that browser gives me lagging and stuttering.

In other news, mediocre but popular browser surpasses all others in a useless benchmark that is out of touch with the real Internet - news at 9

+1 for chrome. I used to use Firefox before they fast forward versions. Last version I used was 3.6. and after that version 4,5,6,7,8,9 came out which are all version 4 for me. e.g 4.0.4 ... 4.0.8. I would probably wait until version 5 and meanwhile I am using chrome

S3P€hR said,
+1 for chrome. I used to use Firefox before they fast forward versions. Last version I used was 3.6. and after that version 4,5,6,7,8,9 came out which are all version 4 for me. e.g 4.0.4 ... 4.0.8. I would probably wait until version 5 and meanwhile I am using chrome

i would say Firefox 7 was more like Firefox 5 as Firefox 7 is when they introduced the much improved RAM use to Firefox as that alone was a solid boost to the browser as a whole.

so basically...

Firefox 4/5/6 is pretty much Firefox 4, where as Firefox 7/8/9 are Firefox 5 but with more performance oriented changes between versions.

I really like Firefox and the fact that it doesn't belong to a corporation such as Google, apple, Microsoft etc. but I think they should fix things fast before they keep loosing their share. first I still think the performance is an issue even with Firefox 8.0.1 which I believe is the latest stable release, performance is worse than 3.6 and chrome 15 of course. secondly I am a web developer and for some memory management feature of chrome it works better with heavier web application less crashes and more stable and responsive and thirdly for html5 support it is behind chrome and that said it is behind safari too but its not true. you can check the score in http://www.html5test.com/

johnnyftw said,
Firefox, Chrome, Opera, IE9.

[Windows Phone 7], Android, iOS, Blackberry

Windows, Apple, Linux

too many people, too many preferences.

Here I fixed it for you.

last time i checked opera was faster... i dont see the chrome hype seriously... firefox its the best overall experience, from customization, addons and realworld usage moderate ram and cpu usage.... opera its a cpu hog and chrome its a ram hog... i find firefox the better balance between both

eilegz said,
last time i checked opera was faster... i dont see the chrome hype seriously... firefox its the best overall experience, from customization, addons and realworld usage moderate ram and cpu usage.... opera its a cpu hog and chrome its a ram hog... i find firefox the better balance between both

+1 i could not agree more as i to think in real world Firefox is pretty much at the top or damn close to it. (i suspect Opera's main problem is it just needs GPU acceleration to it which it should have had ages ago now as that's one area it's behind but considering it don't even support that, it's not bad at all)

Firefox (especially in v7/8+) just has a all around smoother experience. i would imagine it's more pronounced for those of us running Windows Vista/7 with 2GB of RAM as the browser's RAM use stays are much more reasonable levels where as Chrome etc can burn though RAM quickly if you tend to have 10-20tabs open which it's fairly common for me to have around 10 tabs open.

day2die said,
Since when did WebM, Theora, and WebGL become standard?

It's supported in Chrome and Firefox web browsers but it's a option as a download for Internet Explorer 9/10.

Can't imagine what that machine is like they used for testing, but Chrome on ANY of my 8 machines is easily the slowest browser of the four (5 sometimes) I use, not to mention it is just set up dumb as he**!!

Guess you could say I use more than those browsers also, as I DON'T use Firefox, but do use Palemoon and Waterfox (on 64bit machines). Am also playing with that Comodo Dragon browser only have most of their default settings changed. Have used Chrome, but if I'm going to use that variant of a browser, I'd much rather use SRWareIron.

Personally,
Opera still outclasses them all when it comes to the "feel" of speed on any of my machines. Speed is probably close to the last reason I use a browser anyway. NONE of them are so slow, I even think about it twice.

johnnyftw said,
Crazy but, I find Chrome much more responsive and smooth then firefox..

makes me wonder if Chrome works 'faster' for those with fast PC's?

because on my PC, which is pretty much as slow as dual core's get (AMD Athlon X2), Firefox seems to feel all around smoother. especially since it's RAM use is much better vs Chrome and when you got a 2GB of RAM system that seems like it could become noticeable.

Only thing I like about chrome is the bookmarks sync and the fact I can use the chrome marks app on my phone and get the bookmarks since google who creates android AS WELL AS CHROME does not have an actual chrome broswer with bookmarks support on it

Lil Wayne said,
I went from FF to Chrome. I find Chrome much faster and it doesn't freeze compared to FF.

i never really have freeze issues in Firefox 8... my PC is nothing fancy either, AMD Athlon X2 with Radeon 5670 GPU.

also when people say Chrome is faster i assume you need a mid-to-higher end CPU to notice it? , because on my PC (which is about as low end as you can get for a dual core) i think Firefox is overall smoother plus extensions etc are icing on the cake.

but i might give Chrome another go just for kicks as it's been a while since i last tried it.

I guess it depends on user preference....I find chrome pretty responsive, however, I'm slowly finding myself migrating over to IE9 more and more...although I'll never use FX again, I think it still boils down to user preference....whatever works for you works for you, no matter if there's a poll for it or not (although I don't remember participating in such a useless benchmark). Meh...

obiwankenobi said,
I guess it depends on user preference....I find chrome pretty responsive, however, I'm slowly finding myself migrating over to IE9 more and more...although I'll never use FX again, I think it still boils down to user preference....whatever works for you works for you, no matter if there's a poll for it or not (although I don't remember participating in such a useless benchmark). Meh...

Yeah totally man. I find myself using IE9 more and more too. But chrome is pretty light. I dont see how people can state Firefox is more responsive.. considering firefox has indeed strayed from the "lightness" they were back in the day. I dont see it or feel it. Chrome is fantastic. I use it as a daily driver in between IE9. I stopped using firefox long ago because of the road it went down. I dont take anything away from firefox, i still consider it a honest and decent web browser.. it just hasn't been that great on my lenovo laptop.

johnnyftw said,

Yeah totally man. I find myself using IE9 more and more too. But chrome is pretty light. I dont see how people can state Firefox is more responsive.. considering firefox has indeed strayed from the "lightness" they were back in the day. I dont see it or feel it. Chrome is fantastic. I use it as a daily driver in between IE9. I stopped using firefox long ago because of the road it went down. I dont take anything away from firefox, i still consider it a honest and decent web browser.. it just hasn't been that great on my lenovo laptop.

+1

Well said. These days people use more than one web browser that suites their needs and not what others think. I find Chrome relative fast compare to Internet Explorer 9 in some benchmarks and websites. Other than that I enjoy and love Internet Explorer 9/10 over the competition. Choice is good for consumers. Without choice we will be the same boring and dull and web browsers won't be where they are right now.

Wow, not sure where to begin...

1) The tests dock your score if you don't support WebM and Theora, which are not a standard any more than any of the other codecs. - Fail

2) The Iconia W500 tests are running IE10, but the numbers are either in non-accelerated mode or from Windows 8, that are running with reduced performance GPU drivers. I couldn't gather any more information, as it isn't provided.
*(Why this is weird, is that a 2007 Acer Atom based Netbook, running IE9 on Windows 7 gets a higher score, yet it is a MUCH slower computer than the Iconia W500.)
-Fail

3) The HTML5 tests, 3/7 on IE9. 2 of these are failed because of WebM/Theora, and the other 2 are failed because of Web Worker processes. IE supports background script threading without the need of a separate API, which is why when Microsoft added them to IE10 'preview' they were reluctant, but did it to keep the 'tests' happy.
*(On IE, the browser engine is basically a compiler for everything, not just a JIT for Javascript, so all recursive and background scripts are already threaded independently of other UI and HTML elements.
-Pass

4) WebGL - Holy crap batman, really? This is NOT a standard and is HIGHLY dangerous, as it allows the website to directly access and run code on your GPU. Let me put it this way, what if Microsoft released a WebDirectX standard that gave Web sites direct access to your GPU, would anyone be running it as a required feature and 'test'? Nope.
*(The irony to WebGL is that it gained traction from Google and Apple, when IE9 was still in pre-beta, as Google COULD NOT get Chrome to do the graphical HTMLl5 tests at even 1/100th the speed of IE9. So their 'reply' to Microsoft was a WebGL fish animation, shoved into an HTMLl5 Canvas, that they called their HTML5 test. And people were foolish enough to think it had anything to do with HTML5. Yet even as the 'trick' was noticed, with the trick being WebGL, people then moved on to think WebGL was a GOOD idea.

Security experts have demonstrated and warned against WebGL, as it opens the door wide open for malware and at the very least the Web site's WebGL could fry a GPU by knowing exploiting flaws in GPUs that OpenGL expose.

WebGL is not a standard and is highly dangerous.

Even one of team's OpenGL gurus wrote some WebGL code that can exploit a few ATI cards that literally will fry them during their own security work and contribution to this subject.

When security leaders are saying WebGL is more dangerous than unfettered ActiveX of 10 years ago, maybe the world should listen, and STOP USING IT in a freaking 'test' of performance.

-Massive EPIC Fail

5) With the actual tests, on just my test laptop here, Chome flat out pegs one core of the CPU, leaves the other core alone, and yet returns numbers lower than IE9. On the other hand, IE9 threads to both cores on the test with neither Core going over 10%.
*Yet Chrome gets nearly double the score because it supports non-standards like WebM and WebGL, which makes this test 'worthless'.

-EPIC Fail

----

If arbitrary items are used for testing, where a browser's score is reduced for the lack of support of something that is NOT a standard, the test is virtually worthless.

---

We might as well be doing comparative analysis of IETestDrive HTML5 tests, at least they are truly pure HTML5/CSS3 standards, at least they are standards based. However, because Microsoft wrote the tests, people are reluctant to use them. Which is fine, but when other tests like this one, and Sunspider (designed specifically for WebKit engines) are used, people applaud the test even knowing that Sunspider admits it favors WebKit engines.

---

Let's demonstrate this theory out further, and as it also shows where Chrome is 'tricking' certain Graphical tests.

http://ie.microsoft.com/testdr...mance/FishBowl/Default.html

Simple test, should get 60fps (assuming this is your LCD refresh cap)

IE9
10/100/500 Fish 60FPS
CPU Usage Core1 - 40% - Core2 - 20%

Chrome
10/100 Fish 60FPS *Note it reports 20fps on 500 fish
CPU Usage Core1 - 100% - Core2 - 5%

Now notice the actual 'fluidity' and draw speed on Chrome, it is not 60FPS, it is choppy like you are running Crysis on a low end machine. However the test via Chrome reports 60FPS.

The IE9 test is truly 60FPS fluid.

---

Here is where it is interesting, moving over to an older system with an AMD X2 4600 Dual Core, which is significantly slower CPU, yet has slightly faster GPU than my test laptop.

IE9
10/100/500/1000 Fish 60FPS
CPU Usage Core1 - 90% - Core2 - 10%

Chrome
10 Fish 40 FPS
100 Fish 15 FPS
500 Fish 5 FPS
1000 Fish 1 FPS
CPU Usage Core1 - 100% - Core2 - 0%

This is interesting, as you can see where the GPU features of IE9 are valid and kick in for people on lower end hardware.

So if this was a real world HTML5 App, Chrome is unusable for this 'user' in the real world, yet IE9 works just fine for this user.

One more...

http://ie.microsoft.com/testdr...mance/ParticleAcceleration/

Chrome
Max FPS 24, Average FPS 19
CPU Usage Core1 100% - Core2 40%

IE9
Max FPS 60, Average FPS 60
CPU Usage Core1 90% - Core2 20%

Lower end AMD X2

Chrome
Max FPS 8
Average FPS 3
CPU Usage Core1 100% - Core2 60%

IE9
Max FPS 60
Average FPS 60
CPU Usage Core1 90% - Core2 50%

So even though the CPU dropped by over half between the two systems, the IE9 users would have the same experience, and a very fluid experience. The Chrome user can 'get by' on the higher end system being far more CPU bound, but on the AMD X2 system, the test/website is absolutely unusable.

These are tests, but HTML5 tests, and if we are building HTML5 to do RIA, user's need consistent speed, or it makes HTML5 worthless for Chrome users on lower end machine. Sure a nice i5, i7 does ok with Chrome, but users with the AMD X2, are left behind. As for IE9 users, they get the same experience, the developer doesn't have to compensate, nor does the quality of the RIA website, and nobody gets left behind.

Just for a kicks, look at this 'consistency' across hardware...

2007 Acer Atom 270 Netbook
IE9 Particle Test - 30fps solid

WP7 (7.1 Mango)
IE9 Particle Test - Max 10FPS, Average 7FPS
*This one is very interesting, as it is the same exact page, a few less pixels, but for the end user is usable, and faster than Chrome running on a AMD X2 4600 with an NVidia GTS 250 GPU.

(As for my testing, I ensured the Window sizes were the same size for each browser on each machine. *Exception Netbook is lower screen size, and WP7 is lower screen size.
Chrome both levels of GPU acceleration enabled, latest release, and newer test build were tested with neither having an advantage over the other. Laptop )


So if WP7, an old Netbook, and two mid level systems run the HTML5 tests I picked at random from IETestDrive run at 60fps or acceptable levels, yet Chrome is all over the place being so CPU bound, that it is worthless for graphical and higher end HTML5 content.

With this...
Is Chrome really the future of HTML5?
How will web developers deal with the disparity of performance?

My guess, IE9/IE10 will take over the HTML5 RIA world for web developers if Google and Chrome doesn't do something, they and other browsers will once again start losing to IE, as developers will need the IE performance in handling the higher end RIA content and HTML5 graphics.


*So how good is this test? Worthless....

thenetavenger said,
Wow, not sure where to begin...

1) The tests dock your score if you don't support WebM and Theora, which are not a standard any more than any of the other codecs. - Fail

2) The Iconia W500 tests are running IE10, but the numbers are either in non-accelerated mode or from Windows 8, that are running with reduced performance GPU drivers. I couldn't gather any more information, as it isn't provided.
*(Why this is weird, is that a 2007 Acer Atom based Netbook, running IE9 on Windows 7 gets a higher score, yet it is a MUCH slower computer than the Iconia W500.)
-Fail

3) The HTML5 tests, 3/7 on IE9. 2 of these are failed because of WebM/Theora, and the other 2 are failed because of Web Worker processes. IE supports background script threading without the need of a separate API, which is why when Microsoft added them to IE10 'preview' they were reluctant, but did it to keep the 'tests' happy.
*(On IE, the browser engine is basically a compiler for everything, not just a JIT for Javascript, so all recursive and background scripts are already threaded independently of other UI and HTML elements.
-Pass

4) WebGL - Holy crap batman, really? This is NOT a standard and is HIGHLY dangerous, as it allows the website to directly access and run code on your GPU. Let me put it this way, what if Microsoft released a WebDirectX standard that gave Web sites direct access to your GPU, would anyone be running it as a required feature and 'test'? Nope.
*(The irony to WebGL is that it gained traction from Google and Apple, when IE9 was still in pre-beta, as Google COULD NOT get Chrome to do the graphical HTMLl5 tests at even 1/100th the speed of IE9. So their 'reply' to Microsoft was a WebGL fish animation, shoved into an HTMLl5 Canvas, that they called their HTML5 test. And people were foolish enough to think it had anything to do with HTML5. Yet even as the 'trick' was noticed, with the trick being WebGL, people then moved on to think WebGL was a GOOD idea.

Security experts have demonstrated and warned against WebGL, as it opens the door wide open for malware and at the very least the Web site's WebGL could fry a GPU by knowing exploiting flaws in GPUs that OpenGL expose.

WebGL is not a standard and is highly dangerous.

Even one of team's OpenGL gurus wrote some WebGL code that can exploit a few ATI cards that literally will fry them during their own security work and contribution to this subject.

When security leaders are saying WebGL is more dangerous than unfettered ActiveX of 10 years ago, maybe the world should listen, and STOP USING IT in a freaking 'test' of performance.

-Massive EPIC Fail

5) With the actual tests, on just my test laptop here, Chome flat out pegs one core of the CPU, leaves the other core alone, and yet returns numbers lower than IE9. On the other hand, IE9 threads to both cores on the test with neither Core going over 10%.
*Yet Chrome gets nearly double the score because it supports non-standards like WebM and WebGL, which makes this test 'worthless'.

-EPIC Fail

----

If arbitrary items are used for testing, where a browser's score is reduced for the lack of support of something that is NOT a standard, the test is virtually worthless.

---

We might as well be doing comparative analysis of IETestDrive HTML5 tests, at least they are truly pure HTML5/CSS3 standards, at least they are standards based. However, because Microsoft wrote the tests, people are reluctant to use them. Which is fine, but when other tests like this one, and Sunspider (designed specifically for WebKit engines) are used, people applaud the test even knowing that Sunspider admits it favors WebKit engines.

---

Let's demonstrate this theory out further, and as it also shows where Chrome is 'tricking' certain Graphical tests.

http://ie.microsoft.com/testdr...mance/FishBowl/Default.html

Simple test, should get 60fps (assuming this is your LCD refresh cap)

IE9
10/100/500 Fish 60FPS
CPU Usage Core1 - 40% - Core2 - 20%

Chrome
10/100 Fish 60FPS *Note it reports 20fps on 500 fish
CPU Usage Core1 - 100% - Core2 - 5%

Now notice the actual 'fluidity' and draw speed on Chrome, it is not 60FPS, it is choppy like you are running Crysis on a low end machine. However the test via Chrome reports 60FPS.

The IE9 test is truly 60FPS fluid.

---

Here is where it is interesting, moving over to an older system with an AMD X2 4600 Dual Core, which is significantly slower CPU, yet has slightly faster GPU than my test laptop.

IE9
10/100/500/1000 Fish 60FPS
CPU Usage Core1 - 90% - Core2 - 10%

Chrome
10 Fish 40 FPS
100 Fish 15 FPS
500 Fish 5 FPS
1000 Fish 1 FPS
CPU Usage Core1 - 100% - Core2 - 0%

This is interesting, as you can see where the GPU features of IE9 are valid and kick in for people on lower end hardware.

So if this was a real world HTML5 App, Chrome is unusable for this 'user' in the real world, yet IE9 works just fine for this user.

One more...

http://ie.microsoft.com/testdr...mance/ParticleAcceleration/

Chrome
Max FPS 24, Average FPS 19
CPU Usage Core1 100% - Core2 40%

IE9
Max FPS 60, Average FPS 60
CPU Usage Core1 90% - Core2 20%

Lower end AMD X2

Chrome
Max FPS 8
Average FPS 3
CPU Usage Core1 100% - Core2 60%

IE9
Max FPS 60
Average FPS 60
CPU Usage Core1 90% - Core2 50%

So even though the CPU dropped by over half between the two systems, the IE9 users would have the same experience, and a very fluid experience. The Chrome user can 'get by' on the higher end system being far more CPU bound, but on the AMD X2 system, the test/website is absolutely unusable.

These are tests, but HTML5 tests, and if we are building HTML5 to do RIA, user's need consistent speed, or it makes HTML5 worthless for Chrome users on lower end machine. Sure a nice i5, i7 does ok with Chrome, but users with the AMD X2, are left behind. As for IE9 users, they get the same experience, the developer doesn't have to compensate, nor does the quality of the RIA website, and nobody gets left behind.

Just for a kicks, look at this 'consistency' across hardware...

2007 Acer Atom 270 Netbook
IE9 Particle Test - 30fps solid

WP7 (7.1 Mango)
IE9 Particle Test - Max 10FPS, Average 7FPS
*This one is very interesting, as it is the same exact page, a few less pixels, but for the end user is usable, and faster than Chrome running on a AMD X2 4600 with an NVidia GTS 250 GPU.

(As for my testing, I ensured the Window sizes were the same size for each browser on each machine. *Exception Netbook is lower screen size, and WP7 is lower screen size.
Chrome both levels of GPU acceleration enabled, latest release, and newer test build were tested with neither having an advantage over the other. Laptop )


So if WP7, an old Netbook, and two mid level systems run the HTML5 tests I picked at random from IETestDrive run at 60fps or acceptable levels, yet Chrome is all over the place being so CPU bound, that it is worthless for graphical and higher end HTML5 content.

With this...
Is Chrome really the future of HTML5?
How will web developers deal with the disparity of performance?

My guess, IE9/IE10 will take over the HTML5 RIA world for web developers if Google and Chrome doesn't do something, they and other browsers will once again start losing to IE, as developers will need the IE performance in handling the higher end RIA content and HTML5 graphics.


*So how good is this test? Worthless....

+1

I find everything on Internet Explorer 9/10 loads really fast on my mid range laptop.

thenetavenger said,
Wow, not sure where to begin...

1) The tests dock your score if you don't support WebM and Theora, which are not a standard any more than any of the other codecs. -

......

Disproves 3rd party test using Microsoft.com test, nice.

PotatoJ said,

Disproves 3rd party test using Microsoft.com test, nice.


You should read the whole comment. I am not experienced in web development, but I am impressed with his detailed explanation, rather than simply "oh no no no, <insert_browser_here> is always faster blah blah blah so the test is invalid".

Anyway, besides client-side scripting engine, I think the site speed affects a lot too.

So basically your saying Chromes utilization of more than 1 core needs some work as does their GPU acceleration for anything other than WebGL.
Another informative read, good stuff

With regards to WebGL being vulnerable, i certainly don`t like the idea that a webpage containing malicious code could access the grahics driver API completely trusted!
The potential to have screen shots captured and dos is there, but until it starts getting actvely exploited nothing will probably be done...

I'm with the development team of Peacekeeper, and would like to clarify some points,

thenetavenger said,
Wow, not sure where to begin...

1) The tests dock your score if you don't support WebM and Theora, which are not a standard any more than any of the other codecs. - Fail

Actually, the HTML5 tests are not calculated into the score. They are simply reported at the end, but the overall score is not affected by them. Specifics on score calculation are available here,

http://peacekeeper.futuremark.com/faq.action


2) The Iconia W500 tests are running IE10, but the numbers are either in non-accelerated mode or from Windows 8, that are running with reduced performance GPU drivers. I couldn't gather any more information, as it isn't provided.
*(Why this is weird, is that a 2007 Acer Atom based Netbook, running IE9 on Windows 7 gets a higher score, yet it is a MUCH slower computer than the Iconia W500.)
-Fail

The W500 is indeed running Windows 8.

4) WebGL - Holy crap batman, really? This is NOT a standard and is HIGHLY dangerous, as it allows the website to directly access and run code on your GPU.

WebGL by itself is of course a standard. It's not a required section in HTML5, but as most browsers support it is makes sense to have a test for it. Again, as with all the other HTML5 test, it does not affect the overall score.


5) With the actual tests, on just my test laptop here, Chome flat out pegs one core of the CPU, leaves the other core alone, and yet returns numbers lower than IE9. On the other hand, IE9 threads to both cores on the test with neither Core going over 10%.
*Yet Chrome gets nearly double the score because it supports non-standards like WebM and WebGL, which makes this test 'worthless'.

-EPIC Fail

I'd be quite interested in seeing how this could happen, as again the HTML5 tests do not affect the score in any way.

Also note that all the tests are single threaded, with the exception of the Web Worker test(s).

GraphiteCube said,

You should read the whole comment. I am not experienced in web development, but I am impressed with his detailed explanation, rather than simply "oh no no no, <insert_browser_here> is always faster blah blah blah so the test is invalid".

Anyway, besides client-side scripting engine, I think the site speed affects a lot too.

You should not be impressed by thenetavenger comments.

He hides demagogy behind technical semi-truths.

This guy is a MS loyalist (nothing bad about it) and considering the time he loses writing comments here to support MS i would not be surprised he is actually paid for it.

He turns everything in favor of Microsoft.

xan K said,
I use them all, but I prefer Firefox greatly.

Really even Safari ?

Do you use add-ons for them all too ?

xan K said,
I use them all, but I prefer Firefox greatly.

Same I've used it all but I prefer Internet Explorer cause it works and suites my needs without all those useless extensions which hogs down my system.

Lachlan said,
I am looking forward to IE10 since fiddling with the Windows 8 preview.. it seemed super responsive and smooth

I totally agree with that when I tried Windows 8. Didn't keep it though because I wasn't sure if Ad Muncher, Firefox, or Battlefield 3 would work.

Lachlan said,
I am looking forward to IE10 since fiddling with the Windows 8 preview.. it seemed super responsive and smooth

Internet Explorer 10 is really fast for product yet in beta.

Estienne Taylor said,

Internet Explorer 10 is really fast for product yet in beta.

Too bad the IE reputation will hold it back.

De.Bug said,

Too bad the IE reputation will hold it back.

People should try before judging a product based on experience. Internet Explorer 9/10 is nothing like Internet Explorer 6 back in the days. Those days are long gone and it's time to move forward.

Estienne Taylor said,

Internet Explorer 10 is really fast for product yet in beta.


Even IE9 is much faster than Chrome in daily use. Though Chrome isn't slow and on pair with Opera. FF is slowest and lags by fart behind of those 4, still i'm on it.

I wish there were a (from an objective) souce a list of the HTML5 features all the browsers support in a catigorized, itemized tabularized comparison chart with checks and X marks.

Speed is releative. Should only be gauged by the content that is being made. Soon I imagine we will see benchmarks for browsers that are more mainstream... showing realword situations and realistic content passing through the browser engines. Then we can see which is better.

Chrome gets updated like everyday so I imagine they can continue to tweak and optimize. Good for them, I use each browser for their strengths since not one of them runs everthing perfectly. I feel sorry for web developers and designers... they have to deal with the uneven implementations and having to figure out the limitations for each browser and work around it... that is the worst.

mranderson1st said,
I wish there were a (from an objective) souce a list of the HTML5 features all the browsers support in a catigorized, itemized tabularized comparison chart with checks and X marks.

Have a look at http://caniuse.com/#index .

smooth3006 said,
I find firefox very responsive. Never been a fan of chrome.

Agreed. the overall feel of Firefox makes it the best if you ask me (plus with RAM use reduction in Firefox7/8 it's just pushing it that much more ahead). plus extensions are a nice bonus.

p.s. but i would like to thank Chrome for forcing Firefox to do more performance tweaks as i don't think we would have seen all these performance tweaks in Firefox lately if it where not for Chrome as Chrome is keeping Firefox devs from getting lazy and just going threw the motions.

ThaCrip said,

Agreed. the overall feel of Firefox makes it the best if you ask me (plus with RAM use reduction in Firefox7/8 it's just pushing it that much more ahead). plus extensions are a nice bonus.

p.s. but i would like to thank Chrome for forcing Firefox to do more performance tweaks as i don't think we would have seen all these performance tweaks in Firefox lately if it where not for Chrome as Chrome is keeping Firefox devs from getting lazy and just going threw the motions.

Yep exactly. It helped push the Firefox Devs and I do thank it for that! Even Opera but they really don't have to but they still do even so.

smooth3006 said,
I find firefox very responsive. Never been a fan of chrome.

agreed... but i found chrome faster than FF... its just personal preference....

smooth3006 said,
I find firefox very responsive. Never been a fan of chrome.

Chrome is quite fast on slower computers. Opera alway hits a homerun out of the park on faster computers. But most benchmarks are biased against Opera so I'm used to it. Firefox? oh heck FF 8 is worthless. If you can get it to run on older computers. It'll crash on older computers

PatrynXX said,

Chrome is quite fast on slower computers. Opera alway hits a homerun out of the park on faster computers. But most benchmarks are biased against Opera so I'm used to it. Firefox? oh heck FF 8 is worthless. If you can get it to run on older computers. It'll crash on older computers


Opera has always been slow for me and a only a small amount of my friends while it runs fast for our other friends.

PatrynXX said,

Chrome is quite fast on slower computers. Opera alway hits a homerun out of the park on faster computers. But most benchmarks are biased against Opera so I'm used to it. Firefox? oh heck FF 8 is worthless. If you can get it to run on older computers. It'll crash on older computers

older computers as in?

my motherboard is from 2006 (A8N32-Sli Deluxe , AMD Athlon X2 dual core & Radeon HD 5670 512MB and 2GB of RAM) and Firefox 8 is pretty rock solid for me.

so while my PC is not ancient, you could say it's pretty much as slow as you can get for a dual core or better PC.

PatrynXX said,

Chrome is quite fast on slower computers. Opera alway hits a homerun out of the park on faster computers. But most benchmarks are biased against Opera so I'm used to it. Firefox? oh heck FF 8 is worthless. If you can get it to run on older computers. It'll crash on older computers

Yep. I have a dual core 1.8ghz processor with only 1gb ram and chrome is unbelievable on how fast it runs. Where every other browser takes ages to load themselves and load the pages!