(Credit: CNET News)
Google began updating Chrome users with the new beta version, and my performance tests show the company has ratcheted the browser's speed up another notch. Google Chrome's latest version, 0.3.154.9, shows a 37 percent JavaScript performance improvement over the initial beta released two months ago.
JavaScript is a programming language used to add some pizazz to innumerable Web pages, but more importantly from Google's perspective, to power sophisticated Web applications such as Google Docs, Google Calendar, and Gmail. JavaScript is also up against Adobe Systems' Flash and Flex, Microsoft's Silverlight, and HTML 5, in the competition for what's the best foundation for Web applications.
(Credit: CNET News)
Using Google's JavaScript benchmark I pitted the newest Chrome beta, version 0.3.154.9, against both the initial beta from September and the more raw 0.3.154.3 developer release from mid-October. A higher number is better on this test, and the first beta scored 1,851, the 0.3.154.3 developer release 2,265, and the new 0.3.154.9 beta 2,546. Google's tests aren't the only game in town; many use the SunSpider test. Here, too, the new Chrome got a notch faster, getting the test done in 2,546 milliseconds compared with 2,904 milliseconds for 0.3.154.3. (We couldn't test the first version because the testing site was down at the time.)
The new Chrome score catches closer to the 2,250 millisecond score of Firefox 3.1 beta 1 with its new TraceMonkey JavaScript engine enabled. (Tech-Recipes has useful instructions on how to enable TraceMonkey.)
On blogger Matt Asay's advice, I tested Minefield, the cutting-edge version of Firefox that's updated daily. (Minefield is downloadable from Mozilla's FTP site for those willing to use very untested software).
It had the best SunSpider score so far on my machine, 2,147 milliseconds. However, Firefox still lags on Google's speed test. Chrome's latest score of 2,546 is miles ahead of the 215 score from Minefield.
(Credit: CNET News)
There could be something fishy going on here, though: Minefield, which has TraceMonkey turned on, actually is slower than Firefox 3.1 beta 1 with TraceMonkey turned off, which is hardly the result you'd expect for a JavaScript speed test. TraceMonkey-enabled Firefox 3.1 beta 1 couldn't run the test because of a bug, and though that bug was fixed in Minefield, there could be something else awry. And again with the caveats: These tests were run on a dual-core Windows XP machine, and your mileage will undoubtedly vary. They're synthetic benchmarks that may not accurately represent all the particular JavaScript you have to run. And JavaScript isn't the sole measure of a browser's speed.
No comments:
Post a Comment