Internet Explorer 10 CP vs. other beta/preview browsers

Previously we had benchmarked Internet Explorer 10, found in the Windows 8 Consumer Preview, against stable and current versions of browsers from other vendors. Now we have taken the same tests as last time, added some more, and benchmarked Internet Explorer 10 Consumer Preview against the beta versions of major browsers.

All tests conducted here were done one at a time on the same PC running Windows 8 Consumer Preview (build 8250) x64, with fresh installations of the browser and no other applications running or tabs open. Benchmarks were conducted more than twice and averaged.

To clarify: default settings were used across all browsers to ensure an even playing field. As some commenters have noticed, lower-than-expected scores were obtained in some tests from browsers that have hardware acceleration disabled by default. We did not enable this setting if it was disabled because an end-user would not know to enable this when trying the browser, and would luck out in terms of performance gains.

Benchmarked browsers: Internet Explorer 10 Consumer Preview Desktop and Metro (10.0.8250.0), Chrome 18 beta (18.0.1025.45), Firefox 11 beta 5 and Opera Next 12.00 alpha.

Sunspider (in ms, lower is better)

As most people know, Sunspider tests the JavaScript performance of a browser

 

Mozilla Kraken (in ms, lower is better)

Kraken is an alternative JavaScript test made by Mozilla. As this benchmark is designed by the maker of Firefox, the results may be swayed in Firefox's favor.

V8 Benchmark (higher is better)

V8 is another JavaScript benchmark, which is used to tune the V8 JS engine used in Chrome. As such, results may be swayed in Chrome's favor

Peacekeeper (higher is better)

This benchmark shows the all-round capabilities of the browser

HTML5 Test (out of 475)

The HTML5 test awards the browser a score out of 475 that reflects how much support for HTML5 the browser has

 

Particle Accelerator (frames per second, higher is better)

This is a HTML5 graphics test (which we did at 1920 x 1080) that is part of Microsoft's IE10 Test Drive suite. As such, results may favor Internet Explorer.

GUIMark2 HTML5 Tests (frames per second, higher is better)

This is a real-world HTML5 test that measures the performance of the browser performing various HTML5 tasks. There is also a Flash version of the same set of benchmarks.

Fishbowl (in FPS, higher is better)

This benchmark is made by Microsoft to show the HTML5 animation capabilities of Internet Explorer 10, and as such, results may favor Internet Explorer. We did this benchmark with 500 and 2000 fish at 1920x1080

View some of these benchmarks with stable browsers against Internet Explorer 10 here.

Report a problem with article
Previous Story

Contest: Win Heroes of Neweth's 100th Hero Kinesis

Next Story

Bang On: We Need a Third Party in Operating System Politics

40 Comments

Commenting is disabled on this article.

Firefox's D2D implementation has stability issues which DO NOT effect IE9/10. Firefox simply blocks old drivers - which work flawlessly with Direct2D. In other words, IE9 and IE10 will be using D2D in MANY more cases than Firefox. Mozilla hasn't improved their D2D support since... Firefox 4 (aside from bug-fixes and blocking D2D compatible drivers).


Many benchmarks test functionality that is almost nonexistent (and worthless from performance, security, and other perspectives) such as WebM and WebGL. These benchmarks MUST be marked as purposefully ANTI-IE9/10.

If some of say... the Peacekeeper and HTML5 benchmarks were not biased against IE10, IE10 would smoke them.

IE should have a new proper extensions/add-ons system.
I suppose users who rely on certain extensions/add-ons on the other browsers won't consider IE10 or 11 or 12 no matter how well it performs.

Glad to see these tests done. Fairer comparison to the last set.

Microsoft seems to be doing a pretty good job with IE10.

Thanks for this comparison!

Basically exactly like I commented last time for the IE 10 benchmark: Kraken + preview browsers against preview browsers.

Kraken shows interesting differences between the Metro and Desktop versions. I thought they used the same core. Kraken doesn't base any results on graphics, at least, so that's ruling out hardware acceleration differences.

Very interesting that the former "Fastest browser on Earth" (Opera) is falling behind IE 10 in such a modern benchmark. The tables have truly been turned there...

Northgrove said,
Thanks for this comparison!

Basically exactly like I commented last time for the IE 10 benchmark: Kraken + preview browsers against preview browsers.

Kraken shows interesting differences between the Metro and Desktop versions. I thought they used the same core. Kraken doesn't base any results on graphics, at least, so that's ruling out hardware acceleration differences.
...

IE10/Metro browser runs in 64bits.
IE10/desktop runs in 32bits (unless Enhanced Protected Mode is enabled by the user).

64bits apps in general are slower in non-multimedia tasks, such as those involving massive text/dom/pointers manipulations.

A couple of those tests use non-standard features such as WebGL and therefore not a good way of benchmarking browsers. And HTML5test.com is a joke, it only tests the existance of features, not if they actually perform as according to the W3C standard. Not to mention it even includes other stuff like WebGL which is not HTML5.

floopydoodle said,
A couple of those tests use non-standard features such as WebGL and therefore not a good way of benchmarking browsers. And HTML5test.com is a joke, it only tests the existance of features, not if they actually perform as according to the W3C standard. Not to mention it even includes other stuff like WebGL which is not HTML5.

browser tests are a joke.

That is just so Scorpus doesnt have to post it AGAIN....

Chrome is looking very good Beats everything on all tests (the ones that it suffered are "bais")

It only suffers in Javascript.

That being said, I still perfer to use IE.

htcz said,
That is just so Scorpus doesnt have to post it AGAIN....

Chrome is looking very good Beats everything on all tests (the ones that it suffered are "bais")

It only suffers in Javascript.

That being said, I still perfer to use IE.

Lawl? How can you like praise chrome and then say "that being said , i prefer IE" ???

bogas04 said,

Lawl? How can you like praise chrome and then say "that being said , i prefer IE" ???

Because Chrome is a dyno queen. I develop on Chrome/IE9/Firefox side by side and it's pretty startling, given how well it performs in synthetic benchmarks, how much slower Chrome feels than the other 2. Elements animated through javascript seem to run at 60fps on IE9 and Firefox but lag noticeably on Chrome.

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.

htcz said,
If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.

Well.... care to at least do some investigation BEFORE posting ?
If you have a modern GPU and drivers that are not ancient then it IS ENABLED by default in Opera 12.

Heartripper : Noticed that too with Opera 12 latest build.
Particle Accelerator I get 60FPS constant, in Fishbowl if my GPU is on 2D clocks I get ~40FPS with 2000 fishes, and with 3D clocks - I get 60FPS constant.
Something is fishy with those benchmarks.

PrEzi said,
Heartripper : Noticed that too with Opera 12 latest build.
Particle Accelerator I get 60FPS constant, in Fishbowl if my GPU is on 2D clocks I get ~40FPS with 2000 fishes, and with 3D clocks - I get 60FPS constant.
Something is fishy with those benchmarks.

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance. (yep, a repost)

Scorpus said,

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance. (yep, a repost)


Well.... care to at least do some investigation BEFORE posting ?
If you have a modern GPU and drivers that are not ancient then it IS ENABLED by default in Opera 12.

To be brutally honest, even though I use Opera 12, the hardware acceleration sucks big time.

I get that its beta and all that, but I'm sorry the browser window shouldn't lag when I'm trying to expand it. It's been that way since the very first alpha they released.

It still needs a ton of work. I love Opera and think its a great browser but they're quite far behind the other browsers when it comes to gpu acceleration. Good thing though is their software rendering is fast as hell sooo

Well, I was always a fan of internet explorer but after IE 7, it became a piece crap. But IE10 rocks now. It doesn't hang like before. It's fast I don't care for .002ms slow.
I'm happy now.

tuneslover said,
Well, I was always a fan of internet explorer but after IE 7, it became a piece crap. But IE10 rocks now. It doesn't hang like before. It's fast I don't care for .002ms slow.
I'm happy now.

Every version of IE were very stable. If you had hanging issue in previous versions, it was certainly caused by an unsable plugin, BHO, or antivirus integrating Inside IE in a non standard way (like Kaspersky)

do a clean install, and you'll see IE8/IE9 almost never crash or hang.

tuneslover said,
Well, I was always a fan of internet explorer but after IE 7, it became a piece crap.

You're not saying that you actually liked IE6, are you?
I actually only liked IE from ver 7 onwards.
IE 7 was a start. and IE8 was good. IE9 was excellent. IE10 is awesome.

FalseAgent said,

You're not saying that you actually liked IE6, are you?
I actually only liked IE from ver 7 onwards.
IE 7 was crap, and IE8 was good. IE9 was excellent. IE10 is awesome.

Fixed it for you

Scorpus said,

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.


honestly, neowin could add a script on this thread to autopost this comment, I struggle to grasp how hard it is for people to grasp this fact!

Scorpus said,

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.

Sorry, i though it was enabled by default

Scorpus said,

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.

I'm pretty damn sure Hardware Acceleration is enabled by default.

Scorpus said,

If it's not enabled by default, it wouldn't have been benchmarked with it enabled. An end-user wouldn't know to enable it and therefore would luck out on performance.


It is enabled by default as long as you have the correct drivers installed for the graphics card.

LostInNull said,
The most secure browser is chrome .
The fastest browser is chrome.

no It's not.

IE10 now has a better sandbox than Chrome called "Enhanced Protected Mode" (instead of "Protected Mode" for ie7/8/9).

It is 64 bits only, with enhanced support for ASLR (more entropy), and blocks read/write access to the filesystem and registry, blocks interprocess communications, including sockets connections to the same computer.

It is enabled by default on IE10/metro, and can be enabled manually on IE10/desktop (by default it still uses vista/IE7 protected mode sandbox). However plugins like flash player are NOT yet compatible with this sandbox (which is why it is not enabled by default on IE/desktop).
But IE has the ability to switch tabs to the old sandbox model when a page needs to use plugins like flash, as in this screenshot:

http://twitter.com/#!/jman.../175602322330755072/photo/1

tabs running in the new sandbox are shown as running in an AppContainer.
tabs running in the vista/IE7/chrome sandbox are shown as running as Low integrity

Firefox of course has no sandbox at all and is shown as running as Medium Integrity (same access level as the user account)

and concerning graphical performance, and real world web browsing chrome really sucks compared to IE10. Chrome leads only in some javascript benckmarks, not in overall performance.

Idk what's wrong with your system , FF11 should have HA enabled...

Strangely Metro vs Desktop version of IE have slight differences with either being faster in different tests...
Another big thing is that IE is now comparable to FF and Opera in HTML5 feature test , while Chrome is still the winner. Seems like all the modern browsers are good enough now.

bogas04 said,
Idk what's wrong with your system , FF11 should have HA enabled...

As I said in the other article: installed Windows 8 -> installed FF11. If it's not enabled by default then it wouldn't have been benchmarked with it enabled. If was supposed to be enabled but wasn't that's Mozilla's fault, because how would an end-user know to enable it and get better performance?

bogas04 said,

Strangely Metro vs Desktop version of IE have slight differences with either being faster in different tests...

IE10/Metro browser runs in 64bits.
IE10/desktop runs in 32bits (unless Enhanced Protected Mode is enabled by the user).

64bits apps in general are slower in non-multimedia tasks, such as those involving massive text/dom/pointers manipulations.

bogas04 said,

Another big thing is that IE is now comparable to FF and Opera in HTML5 feature test , while Chrome is still the winner. Seems like all the modern browsers are good enough now.

Actually, IE10 has a better standard support than the other browsers, as did IE9 when it was released ( http://www.zdnet.com/blog/hard...nformance-test-is-ie9/10213 )

HTML5test.com is NOT a reliable way to know the quality of the HTML5 implementation. It is just a test asking the browser whether it does implement a feature. I does NOT verify if it is well implemented, and compliant with the W3C standards. And it even test things that are NOT part of HTML5 and are not standard.

Chrome and Firefox are racing to implement tons of half baked HTML5 features that are NOT standard compliant, and will break websites rendering in future versions.

Look at the W3C's HTML5 test case to see the real quality of the html5 implementation in IE10 vs the other browsers.

http://samples.msdn.microsoft.com/ietestcenter/

what is the point of implementing tons of web standards is they are not reliable and have a 10% implementation fault rate? Do the webmasters miss the days of IE6 and want firefox/chrome to provide them more headaches?

Currently most self proclaimed "HTML5 based" websites only uses HTML4 features and a little CSS3.
Read this if you want to know what webmasters are going to face is they actually use advanced HTML5 features :

http://blog.millermedeiros.com/ipad-is-the-new-ie6/

Microsoft's biggest problem with IE is they still sync the releases with Windows. They need to decouple the IE release cycle from the Windows release cycle and do iterative releases far more often. Otherwise you'll only be current for a couple of weeks every two years .

Frazell Thomas said,
Microsoft's biggest problem with IE is they still sync the releases with Windows. They need to decouple the IE release cycle from the Windows release cycle and do iterative releases far more often. Otherwise you'll only be current for a couple of weeks every two years .

With each version of IE being supported at least 10 years (13 years for IE6), Microsoft certainly can't release a new version of IE every 3 months!

Even every year, that would make 10 versions of IE to support by 2022, with enterprises customers using any version of IE, not just the latest. It would be a headache for companies to certify business web apps for the ten supported versions of IE, when they don't know in advance what version of IE their customers will be running!

So I guess 2 years between each release of IE is very good.

Frazell Thomas said,
Microsoft's biggest problem with IE is they still sync the releases with Windows. They need to decouple the IE release cycle from the Windows release cycle and do iterative releases far more often. Otherwise you'll only be current for a couple of weeks every two years .

IE9 was not released with any specific OS version. Also the time between IE9 (March 2011) and IE10 (Fall 2012) is less than 2 years.