Surprising winner in JavaScript speed wars: Metro IE10

While it's impossible to objectively measure Web browser 'speed,' a few JavaScript benchmark tests produced surprising results

Late last week Google announced it had released an iOS version of the Chrome browser. Expectations for the iOS interface ran high, but nobody expected iOS Chrome to shake off the lethargy that's a defining characteristic of all Web browsers running on iOS.

To put it simply, Apple doesn't let Web browsers run their own JavaScript engines. As I explained a couple of months ago, non-Apple browsers are all forced to use a slow rendering class known as UIWebView, while Safari gets to run its Nitro version of JavaScript. It's anything but a level playing field.

JavaScript speed tests are hardly the alpha and omega of browser performance, but in this case, Chrome took a cruel hit. I pitted Chrome against Safari, using the SunSpider 0.9.1 JavaScript benchmark. SunSpider is hardly unbiased -- it's made by Apple's WebKit development team -- but it's reasonably well accepted as a valid test regimen that covers many real-world situations.

Running on my iPad 2 with iOS 5.1.1, Safari came in with a score of 1,722 milliseconds. The new iOS Chrome slogged along at 7,224 ms. Chrome took four times as long to complete the test regimen. Blech. That much was expected.

Here's what I didn't expect.

I ran SunSpider 0.9.1 on an older MacBook Pro with OS X 10.7.4 and Safari 5.1.5 -- standard stuff -- and clocked it at 185 ms. That's quick, and it makes the iPad look positively pokey. Safari on the iPad takes about 10 times as long to complete the SunSpider 0.9.1 regimen as Safari on the older MacBook Pro.

As you may recall, about two weeks ago Google released a beta version of Chrome that works on the Metro side of Windows 8. There's a lot of wailing and gnashing of teeth about that version of Chrome because it doesn't conform to Microsoft's Metro design guidelines. Metro Chrome takes advantage of a loophole Microsoft introduced that allows Windows 8 browsers running on the Metro side to draw on both the WinRT API and the Win32 API. That should, in theory, allow Google to build a browser for the Windows 8 platform that runs as fast as Internet Explorer. (Confusingly, Windows RT is a completely different kettle of fish.)

In for a penny, in for a pound, I decided to run the SunSpider 0.9.1 benchmark on the Metro version of Chrome, using a test desktop (not a tablet). Metro Chrome's still in beta, mind you, so the numbers are far from final. But I was stunned to see the Metro Chrome speed test come in at 196 ms -- within spitting distance of Safari running on full-blown OS X.

With no small amount of trepidation, I decided to run the same SunSpider 0.9.1 benchmark on the Metro version of Internet Explorer 10. Metro IE10 is also still in beta. The result: 155 ms -- the fastest of them all. With a beta browser. On a beta operating system.

It's true. Using this one test -- which is admittedly far from definitive and only tests JavaScript execution speeds -- Metro IE10 beat the pants off all the competitors.

Maybe there's a horse race underway after all.

This story, "Surprising winner in JavaScript speed tests: Metro IE10," was originally published at Get the first word on what the important tech news really means with the InfoWorld Tech Watch blog. For the latest developments in business technology news, follow on Twitter.