Study: Internet Explorer 9 fastest browser on Windows

Who has the fastest web browser on Windows? Would you believe Internet Explorer 9? That's the claim from New Relic, a company that helps web sites boost their performance. They conducted a study on the speed of various web browser for Windows and Mac. New Relic states they conducted their study by observing 5 billion global page loads for a period of one week in March 2012.

On the Windows side, IE 9 was the fastest with an average page load time of just 3 seconds. Mozilla's Firefox 11 and 14, along with Google's Chrome 17, all tied for second with load times of 3.5 seconds each.

While IE 9 may be fastest on Windows, on the Mac it's a different story. As the chart above shows, the 10 fastest web browsers on the PC (combining results from both Windows and Mac) all belong on the Mac side, with the Mac port of Chrome 13 being the fastest to load a page at just 2.4 seconds.

In terms of web browser popularity, New Relic says that the various versions of Internet Explorer still rule on Windows with 40 percent of the market share. However, Google's Chrome is in second, according to the study, with 33 percent, followed by Firefox with 24 percent. Opera and Safari are well behind at two percent and one percent, respectively. On the Mac side of things, it's no surprise that Safari is the leader with 59 percent of the market share. Firefox is second with 21 percent and Chrome is not far behind at 19 percent.

Images via New Relic

Report a problem with article
Previous Story

More Windows 8 Consumer Preview secrets revealed

Next Story

Study: 25 percent of all households use WiFi

44 Comments

Commenting is disabled on this article.

BS. Chrome IS the fastest. Please refer to acid test. HTML 5 test and even Microsoft's own IE Testdrive website. in most of them Chrome beats IE 9.

S3P€hR said,
BS. Chrome IS the fastest. Please refer to acid test. HTML 5 test and even Microsoft's own IE Testdrive website. in most of them Chrome beats IE 9.

That is because your CPU is much faster than your GPU probably.

S3P€hR said,
BS. Chrome IS the fastest. Please refer to acid test. HTML 5 test and even Microsoft's own IE Testdrive website. in most of them Chrome beats IE 9.

And I see the exact opposite on my machine. The difference is IE9 takes advantage of the GPU. If your computer has a good GPU IE9 will blow Chrome away. On the other hand, Chrome is better optimized for CPU, so if you have a weak GPU, Chrome will handily beat IE9.

Beyond performance, there's the fact that Chrome tends to have strange rendering bugs, and partial "support" for things that it supposedly supports. IE9 doesn't support as many new CSS and HTML features, but what it does support it tends to support in a more consistent manner.

For example, Chrome doesn't even handle something as simple as borders correctly. See this jsFiddle if you don't believe me: http://jsfiddle.net/kGST9/6/

Edited by JonathanMarston, Apr 6 2012, 5:47am :

S3P€hR said,
BS. Chrome IS the fastest. Please refer to acid test. HTML 5 test and even Microsoft's own IE Testdrive website. in most of them Chrome beats IE 9.

Acid Test does not define speed son, it shows standards compliance. In addition, the HTML5 standards changed just after IE9 was released.

Not wanting to troll, but after trying to get CSS3 gradients to work in IE9 (tip: they don't) I would like to know what IE 9 is not rendering or the degree of accuracy in the final page output compared to IE10.

Latest builds of IE, Firefox, Opera, Safari here and IE 9 certainly is fastest in almost every test I throw at them... however I'll stick with Firefox till the others have the plugins I need.

I like it how when IE is on top it must be an MS sponsored survey etc.
Anyway here are my sunspider results on 12 core i7 3.2GHz with 32GB DDR3 RAM SSD HDD and a Radeon 6990 Windows 8 CP:
Latest Chrome: 218ms
Firefox Beta: 199ms
IE 10: 185.9ms

Is the test sponsored by Microsoft?
I tested in my computer and found that IE9 is much slower than Firefox11.

ray_bk said,
Is the test sponsored by Microsoft?
I tested in my computer and found that IE9 is much slower than Firefox11.

I tested in my computer and found the opposite...

Really?

thenetavenger said,

I tested in my computer and found the opposite...

Really?

Oh my god, its like software can perform differently on different computers!

Clearly the data have shown should be some kind of joke.

One Week ago, I published a new Web Browser comparative, and my data match with the last "Grand Prix Browser" from tomshardware:

http://www.tomshardware.com/re...fox-9-benchmark,3108-3.html
http://blog.theliel.es/2012/03...va-navegadores-2012-q2.html [its in Spanish, sorry, but you can view only the graphs]

In terms of load speed (conection to server, request and download data), IE9 is behind Firefox and Chrome. In terms of general rendering webpages (not only downloading the web) the results are even worse on warm boot.

In JS terms, IE9 stuck, only on SunSpider win but because a test cheat by Microsft, the proof are that on real web applications or any other JS test, the results are much much worse.

About MAC OS, Web browser performance on MAC OS, its a lot of worse than Windows. The best performance on MAC OS its with Safari, and its its a lot of far from the great performance on Windows with ANY other browser.

Anyway, IE9/10 its a good browser, but in general terms, the best browser with difference its Firefox.

Theliel said,
Clearly the data have shown should be some kind of joke.

One Week ago, I published a new Web Browser comparative, and my data match with the last "Grand Prix Browser" from tomshardware:

http://www.tomshardware.com/re...fox-9-benchmark,3108-3.html
http://blog.theliel.es/2012/03...va-navegadores-2012-q2.html [its in Spanish, sorry, but you can view only the graphs]

In terms of load speed (conection to server, request and download data), IE9 is behind Firefox and Chrome. In terms of general rendering webpages (not only downloading the web) the results are even worse on warm boot.

In JS terms, IE9 stuck, only on SunSpider win but because a test cheat by Microsft, the proof are that on real web applications or any other JS test, the results are much much worse.

About MAC OS, Web browser performance on MAC OS, its a lot of worse than Windows. The best performance on MAC OS its with Safari, and its its a lot of far from the great performance on Windows with ANY other browser.

Anyway, IE9/10 its a good browser, but in general terms, the best browser with difference its Firefox.

A very 'narrow' test versus a very 'wide' test will have different results. Narrow tests are easier to manipulate as the testers can pick and choose the websites/tests to use, that are often artificial.

A 'wide' test is often 'real world' applications or web sites in this example, that are ran a mind numbing amount of times.


Does Chrome ALWAYS beat IE9 on the same test/site in the Tom's example? If you answer yes, then rethink for a minute, and consider the tests run on an low powered Netbook with Atom 270 and then ran on various hardware configurations all the way up to a 12-94 cpu/core SMP Intel or AMD setup. Now rethink the quesion... Does Chrome 'ALWAYS' beat IE9 on the same test/site STILL? The answer is NO...

One thing that IE9 does well is the way it balances out GPU utilization and CPU cores, so that even the Atom 270 is properly using the HT aspect of the CPU to run the background Javascript JIT as a tiny example. IE9 also uses forms of GP-GPU functionality for things 'beyond' the rendering and it also directly sends content to the DWM.

With this in mind, on average, including low end systems, IE9 will provide a more consistent speed, especially in dynamic and graphical web sites because of how its engine balances out the CPU/GPU.

There are HTML5/CSS3 graphical tests out there that STILL run faster on IE9 on a Atom 270 netbook than Chrome on the fastest i7 desktop. This is where the GPU usage difference really start to hurt Chrome. Some of these test will even hit a higher FPS on a WP7 than on Chrome on a Desktop i7.

This leaves the 'unexpected' performance problem for developers when users come to the site with Chrome, as they can't just expect Chrome to handle the site in a certain way, it will matter what hardware Chrome is running on far more than it will for the IE9 visiting the site.

And since most of these circumstances are HTML5/CSS3 and pushing the limits of the new standards, web site developers are going back to using IE9 as their baseline, and hoping Chrome and Firefox will eventually catch up.

Tests are just tests, and there are thousands of factor that can 'influence' a test. I can take the Tom's tests, and give you a system that will show IE9 as 5 times faster than Chrome and Firefox in almost EVERY test, it is knowing what affects the test and shoving the favor to a particular outcome.

This is also important to realize when people that are doing the tests DO NOT know what changes or affects the outcome, and are not intentionally, but all the same shoving the results to an artificial result that will not be accurate.

One example is a test that has the Aero DWM turned on/off will change the numbers by often reducing the performance of IE9, and giving Chrome a slight boost because of the old methods Chrome renders content on Windows.


So how do you truly measure the metrics of this that get so freaking complicated so fast? You can't 100%, even this test that puts IE9 on top, would have to be ran on several architectures from the low end devices all the way up to the high end systems with various CPU/GPU combinations as they are both 'used' by IE9 and now Chrome.

Today all the browsers are very close in performance 99% of the time, so find the sites YOU personally use, and see what browser 'feels' faster to you.

If you are a developer are doing HTML5 work like the examples at Microsoft's ( http://ie.microsoft.com/testdrive/Views/SiteMap/ ) you will end up using IE9 or IE10, as Chrome and Firefox can't get close to the numbers. And in reverse, if you are math engineer and are running specific Javascript browser based computations, you might find Firefox, or Chrome to be your best choice.

thenetavenger said,

A very 'narrow' test versus a very 'wide' test will have different results. Narrow tests are easier to manipulate as the testers can pick and cho

i wish i had the patience to read this it looks interesting

technopercy said,
Well, My chrome 17 loads in 2 seconds in a windows machine even thought my pc is not powerful

They are talking web pages. Not load times.

Phouchg said,
Functionality > Speed.
When we're talking about fractions of a second, I 100% agree. If its like 2+ seconds, that would get annoying. I ditched early versions of firefox because it would take 3 seconds to open the browser after launching it. 3s doesn't sound like much, but it can sure be aggravating when it happens EVERY time.

And its weird to notice how Chrome is getting slower with each release , i am afraid after gaining decent market share they might take things lightly

I guess it depends on what system you are running it on. For me, its perfectly fine and I have no add-on's running. So personally, I have no reason to use any other browswer.

thekim said,
make a sunspider test. The IE9 ist 4 times slower than Chrome or Firefox.

sunspider just tests javascript, right? These guys tested page load which is probably a better indication of speed.

redavenger said,

This is an outright lie. IE9 is faster than both Chrome and Firefox...

Here are 3 runs i just tested with latest for all 3.


Interesting that, at the educational place here after launching IE 9 and waiting approximately 7 minutes for the page to 'stop responding', you can actually use it.
I don't quite know why IE9 is **** on every single PC there or what it's to do with but they all do it and it's very god-damn slow.

redavenger said,

This is an outright lie. IE9 is faster than both Chrome and Firefox...

Here are 3 runs i just tested with latest for all 3.


With my Hardware much different.

thekim said,
make a sunspider test. The IE9 ist 4 times slower than Chrome or Firefox.

You also realize the sunspider test was design and is for WebKit (ie Safari, Chrome).

A lot of people use it as a general 'javascript' benchmark tools, but it modeled to test the way WebKit javascript engines work.

Why this is important, even Firefox has to 'align' itself when running these tests.

IE9 also had to 'force' itself to run through meaningless iterations of code, that its new compiler technology 'predicts' around. So when IE9 was first being run on SunSpider it was having some tests return instantly, when the idea of the test was to loop, and IE9's Javascript compiler was able to see it was a meaningless loop, and process the result instantly. (Thus they had to adjust IE9 not to do this on the SunSpider test, as people were calling IE9 a fraud for doing exactly what it was designed to do, that the WebKit javascript engines are NOT.)

There is so much more to displaying a web page and even interacting with code on a web page than Javascript, that putting any weight in a pure Javascript test is meaningless when the variance between the products is microseconds.

n_K said,

Interesting that, at the educational place here after launching IE 9 and waiting approximately 7 minutes for the page to 'stop responding', you can actually use it.
I don't quite know why IE9 is **** on every single PC there or what it's to do with but they all do it and it's very god-damn slow.

This would be a good "benchmark" test for someone to conduct - which browser actually loads up the fastest! ...I agree whenever I open IE9 I have to wait an unreasonably long time for the interface to become "responsive" too!

GreatMarkO said,

This would be a good "benchmark" test for someone to conduct - which browser actually loads up the fastest! ...I agree whenever I open IE9 I have to wait an unreasonably long time for the interface to become "responsive" too!

IE9 has always been very responsive for me -- but I run with no plugins. Just like craplets can destroy Windows boot times, crappy plugins can destroy IE9 load times.

These tests are so pointless. How can i even test browser or Raid SSD setup with 100Mb/s Internet Connection. No difference!

The credibility of this survey is underminded by the fact it claims IE7 is faster than semi-modern builds of Firefox and Chrome

thealexweb said,
The credibility of this survey is underminded by the fact it claims IE7 is faster than semi-modern builds of Firefox and Chrome

Judging by my experience, anything is faster than Firefox at this point.

thealexweb said,
The credibility of this survey is underminded by the fact it claims IE7 is faster than semi-modern builds of Firefox and Chrome

They don't test the performance of dynamic content or javascript like most web tests do, (SunSpider, V8, Peacekeeper, etc), just raw page load speeds. The other browsers might be faster in actual use, but this is just a single case.

Dot Matrix said,

Judging by my experience, anything is faster than Firefox at this point.

The newest Firefox builds are quite nippy, much faster than IE7 and IE8 and probably faster than IE9 for most people despite what this article says.

Dot Matrix said,

Judging by my experience, anything is faster than Firefox at this point.

Firefox's UI is laggy, but I've never had issues with page loading speeds.

ViperAFK said,

Firefox's UI is laggy, but I've never had issues with page loading speeds.

I have Firefox Aurora, running over 20, 30 tabs at once (don't ask) and it still runs smooth.

~Johnny said,

They don't test the performance of dynamic content or javascript like most web tests do, (SunSpider, V8, Peacekeeper, etc), just raw page load speeds. The other browsers might be faster in actual use, but this is just a single case.

Actually i think real world tests are better than synthetic tests

Here is a link to the post: http://blog.newrelic.com/2012/04/05/fastest-browsers/