Ie9 Becomes the New King (wait, What?)

Yes, that was the sound of hell freezing over.

For anyone who isn’t familiar with Microsoft’s Internet Explorer, I’d like to tell you about this little thing called “The Internet”. I say this because as long as there’s been a World Wide Web, there’s been Internet Explorer (ok not quite, but pretty close).  For as long as it’s been around, the blue E has been a constant source of frustration to users around the globe.  Even when it was the only real option, it still seemed to be a headache.

When Internet Explorer beat Netscape in the browser wars of the late 90’s, we really didn’t know what to make of it all; Netscape had started off strong and was faster than IE by far, but it had started to stagnate towards the end of its life.  This meant that Microsoft had virtually no competition for many years.  Developers started making websites to look good in IE, which was a problem because IE didn’t render webpages using the correct web standards. Years pass and the Netscape engine was used to create a new browser called Mozilla.  Mozilla then created Firefox, and we all know what happened afterwards.

In the last year, however, Google’s Chrome browser has started taking large chunks of market share and hasn’t showed any signs of stopping.  Last week at South by Southwest, Microsoft officially launched Internet Explorer 9 which had been in public beta for some time.  Shortly after that, Ars Technica called it the most modern browser out there.  Five minutes after that, Elvis announced he was returning to Earth.

For the record, I’m officially eating crow after my review of the IE9 Beta.  Items that had been nice before (the tabbed browsing, plugin management, etc) are even more polished and work perfectly.  Items that didn’t do so well (speed and stability) are now through the roof!  A feature that has been added to Internet Explorer is called the Jumplist; by “pinning” a website to your taskbar, you keep it running in the background (handy for webmail sites) and also gain access to the most common areas of the site.  A site has to be coded to make use of this feature, but it’s so simple now that I foresee this becoming commonplace very quickly.  Still missing from the feature list, however, is the relatively new “User Syncing” that Google introduced with Chrome.  Users can sign into their Google account and then anywhere they are at a computer with Chrome, just sign in again in the Options Menu, and all your bookmarks/logins/extensions/history come with you.  This is an awesome feature that has since been copied into Opera and Firefox.  It shouldn’t be hard to add to IE as well, but knowing their development cycle, we might have to wait for IE10 (IE-X?).

Also released recently is the new Firefox 4.  This new release utilizes a new JavaScript engine that’s supposed to speed up Firefox considerably (sources quote a 60% increase in speed).  While this is nice, it doesn’t really help where Firefox is at its worst; it’s program overhead.  With the new releases hitting the scene, I thought this might be a good time to run all browsers through some benchmarks to see which ones are up to snuff!

All tests were done on the same machine: Windows 7 x64 Ultimate, 12GB DDR3-800, Core i7 920 (2.67GHz), Radeon HD3870 x2, 1TB HDD (4x WD VelociRaptors 10K RPM drives in RAID 0)

SunSpider Javascript Benchmark

SunSpider has been an old staple for benchmarking Javascript for awhile now.  It runs the browser through a load of java calculations to see  how long the engine takes to execute the command.  These calculations range from displaying forms to encrypting a 3D render, so while this is a good test to see how the engine performs, it’s not necessarily a good measure of how your browsing experience might be (as most of us don’t use websites that use these functions).  This test is CPU intensive and your score will be different than my scores.

Internet Explorer 9 takes the clear lead here as it completes all tasks much faster than any other browser.  This was the first test I ran, and it’s also the first test I’ve ever seen IE win in . . . so I ran it many times!  I ran it on my laptop, my desktop, and my friend’s computers.  Each time we ran it, it came back the same: Internet Explorer 9 consistently completed the test in ~220ms.

Peacekeeper by Futuremark

We’ve covered Peacekeeper here before when it was in beta and loved how it worked across all platforms.  It has left the beta stage and is ready to be unleashed upon the world!  Peacekeeper tests the browser across all areas one might see in day to day use (and many that you’d never see). Chrome is no longer at the top of its game here, slipping below Opera for the first time.  The version of Chrome I’m running happens to be an experimental build, however, and so may not be as polished as the official Opera 11 release.  For the sake of argument, and considering the narrow margin here, one should just assume the two browsers are equally matched. Not surprisingly, Internet Explorer falls behind the mighty Chrome.  VERY surprisingly is Firefox’s low score, finishing dead last and by quite a margin.  Thinking this must be an error, I compared my scores to others on Futuremark’s forum boards only to find that my scores were in line with theirs (give or take a couple hundred).  Annoyingly, the Firefox fanboys on the site thought this was a good score!

Note: Peacekeeper runs a system scan before running this test to gather information on what kind of system you’re running (CPU speed, RAM, OS, etc).  IE9 never moved past this point for some reason.  There is another option to run the benchmark without the scan, which is how these scores were obtained.

UPDATE: Evidently IE9’s new security features were actively blocking the site from scanning my system for information!  There’s also an option to permanently block all sites from gathering physical location data.  While this is a very nice feature (and adds some credibility to the idea that IE9 protects your data better than any other), it would be nice if it would tell me what it was doing and give me the option to share information.

Acid3

Acid3 is a very simple test that checks the performance of a browser’s rendering engine against web standard programming.  Long story short, it see’s how closely your browser follows the rules.  There are three stats to compare, which didn’t lend itself very well to a chart so we’ll have to go with a table.  On the chopping block is the speed at which the test was run (lower is better), how many times it had to retry the test to finish, and the percentage of tests passed.

Chome is the winner here, being able to complete the test in the shortest time possible and still scoring 100%.  While Internet Explorer got the lowest score, it did it in one run with a decent time.  Safari not just lost the foot race, it forgot to start it in the first place.  While it was able to get 100% in the end, it took it more tries and took far longer to finish than any of the competition.

FishIE Tank

This new test comes with a major grain of salt.  Called the FishIE Tank, it’s on Microsoft’s own site touting the speed of IE9.  In fact, there’s a whole battery of tests for you to compare its speed, HTML5 redering, security, and other features.  Since it’s on the ad site for IE9, it’s not hard to see why IE9 should do so well at them.  However, this isn’t to say we can simply disregard these tests either; each of these tests showcase a web technology that we either use everyday, or will be soon enough.

The FishIE test (results in Frames Per Second) shows off Internet Explorer’s powerful rendering capabilities by throwing fish (in this case, 20 of them) onto your screen and having them swim around.  IE9 makes use of your computer’s GPU (graphics processor) to render such elements, thereby easing the workload on the CPU and making clever use of part of your computer that’s usually sitting dormant.  No matter what resolution I set it to, minimized or fullscreen, and even with multiple windows open, the meter stayed pegged at 60FPS.  I’d dare say it would have gone higher, except that was the top of the gauge!

On the far far FAR other end was Safari that played like a Powerpoint presentation instead.  It claimed it was moving at 10FPS, but it was dead still the entire time.  Only when I clicked the screen did it ever change.  I ran nearly every test on that site, and Safari failed miserably at every single one.  I had been hoping to post the results of the Speed Read test, but where IE9 finished in 4 seconds and Chrome in 26, Safari hadn’t finished more than 10% after 15 minutes!

In closing, while I might not agree wholeheartedly with Ars Technica’s assesment of IE9, I can totally say it’s come a long ways from IE8.  While I may not be able to stop using my Google Chrome (thank’s to the ever-handy Sync feature), IE9 sure gives the other browsers a run for their money.  Most surprisingly about all this is how far down the ladder Firefox has fallen.  This might be due to its open source nature; since everyone CAN write a handy extension for it, everyone HAS.  As a result the official build just gets more and more bloated with everyone else’s code!  In the end, it’s the flexibility that everyone loved about Firefox that has ultimately killed it.  Until some major optimization can be done, it’s really not up where it once was in terms of performance.

And perhaps the most glaring result of all these tests is just how TERRIBLE Safari is on the PC.  I might be ok with this level of performance on my iPhone or iPad (and indeed, many of these scores are right in line with them), there is absolutely no excuse as to why it should just be so broken on a desktop PC!