Speed
Browser developers are making a big fuss over page-loading speed right now--after all, everyone likes to be the fastest.
Though we saw some very slight speed differences in our page-loading tests--see the chart here for a summary--our findings indicate that most browser speed claims are overblown. (For more, see "Does Speed Still Matter?")
JavaScript Performance
According to the SunSpider browser benchmark, Google Chrome has the best JavaScript performance, as it completed the tests in 301.1 milliseconds on average. Opera and Safari followed in second and third, completing the benchmark in 575.4 and 597.7 milliseconds, respectively. Firefox came in fourth at 902.5 milliseconds. Internet Explorer brought up the rear, finishing the test in 3903.2 milliseconds (no, that isn't a typo; IE 8 really did lag that far behind).
As we were wrapping up this review, Opera Software released version 10.6 of the Opera browser. It came out too late to be included in our live-site testing, but I did have a chance to run it through the SunSpider Benchmark. Although Opera claims that the new release offers a 50 percent increase in performance over version 10.50, it finished about even with Opera 10.53 in the SunSpider tests.
Since JavaScript is becoming so prominent in Web apps and even on regular Websites, it's a good metric to test. SunSpider is a useful way to gauge JavaScript handling, too. It isn't, however, a great indicator of overall browser speed, as plenty of other aspects contribute to browser performance.
BEST JAVASCRIPT HANDLING: Chrome
Google's browser surpassed its four major competitors by a fair margin on this particular test.
Page-Loading Tests
Though Chrome was the fastest of the group in our page-loading tests, the speed differences among the browsers were negligible.
Internet Explorer 8 put up a respectable showing overall and was the fastest browser in five of our tests, but Chrome's average page-load time of 1.75 seconds was the speediest of the five browsers we looked at. Safari 5 came in second overall at 1.89 seconds.
In many instances page-loading speeds were close. For example, in our ebay.com test, four of the five browsers loaded the page within eight- or nine-tenths of a second; only Opera took over 1 second to load the page, and even then, at 1.09 seconds, its page-load time wasn't horrible.
That said, Opera was the slowest browser in three of our eight live-site tests. In two of those three tests (pcworld.com and en.wikipedia.org), Opera lagged significantly behind the rest of the pack, loading pages over a full second slower than the next-slowest browser. In real time, a second isn't that big of a difference, but in terms of percentage, it's significant. On the other hand, in our internal-network tests, Opera came out ahead in two of the three tests we performed. We weren't quite able to explain why this difference exists--our live-site page-loading tests put up repeatable results, by and large--but it's possible that other browsers handle network latency better than Opera does.
In real-life use, browser speed claims are probably overstated. Though your results will vary depending on your PC, its operating system, and which sites you visit, among other factors, you likely won't notice the difference between browsers in regular Web surfing unless you perform tests similar to the ones we did. Ultimately, any browser you select will be fast enough.
BEST PAGE LOADING: Chrome
In this speed test, Chrome enjoyed only a slight lead over the other four browsers we examined.
Which Browser Should You Choose?
All five of the browsers we looked at are free, so it never hurts to try a different one. But jumping from browser to browser has one hidden cost: your time. It can take a while to set up a browser to look and behave just the way you like it. So with that in mind, if we had to pick only one browser, Chrome would be our top recommendation. We like Chrome for its clean and friendly interface, good performance, and strong security. It covers all the basics for most users, and offers plenty of customization for power users.