Browser Speed Tests: Safari 5, Firefox 3.6, Chrome 6, and Opera 10.6 Beta
6/30/2010 01:18:00 AM
kenmouse
, Posted in
Firefox
,
Google Chrome
,
IE
,
IE9
,
Opera
,
Safari
,
Technology
,
0 Comments
Apple's stepped up with Safari 5, Firefox has brought forth a more crash-proof 3.6, and Opera's continuing to push forward in betas. Let's break out the timer and testing software to see how the latest browsers run on real hardware.
We're using much the same testing methods as we did with our last performance tests. If you're not familiar with our series, we use this editor's eyes and fingers (and Rob Keir's timer app) to time the loading of browsers from first boot-up ("cold"), after already running once ("warm"), and in opening multiple tabs at once—in this case, nine. For JavaScript and DOM/CSS scripting tests, we use Mozilla's Dromaeo suite, which incorporates tests from Google's V8 series, Apple's SunSpider tests, and other independent markers.
In measuring memory use, we've switched away from the Windows process manager to Google's about:memory page, which can measure memory use across browsers, and, more importantly, measures the resources that multi-process browsers actually use. Given that Chrome and Safari are using the same WebKit engine, and Firefox has moved to a Chrome-like sandboxing of plug-ins inside their own process, it made sense to use the total memory count from Chrome's measurement tool.
Finally, these tests are performed on a ThinkPad T61p laptop, running a 2.0 GHz Intel Centrino Duo processor and 2 GB of RAM, with a fairly fresh install of Windows 7. The start-up tests were run using a copy of Google's home page saved to the hard drive to negate network differences, and the nine-page loads were performed with an ethernet cable attached to a 50 Mb/s Time Warner cable connection. For every timing test, three results were gathered, and any obvious outliers were eliminated.
Now—onto the juicy number stuff. One important change to the charts this time around: we highlighted the winner, or winners, in green, for an easy visual orientation of whether you're looking at a "less is better" or "more is better" graph. Click any of the images below for a larger, wider view.
Boot-Up and Warm Loading—Winner: Opera 10.6 Beta!
Not a huge surprise, as Opera's 10.5 pre-alpha took the prize for cold start-ups last time. It's probably the most variable of the tests in this series, given the mysteries of hard drives, CPUs, and Windows app management—but Opera consistently starts up the fastest from a dead stop. The "warm" starts, after having run each browser once, are close enough together that Chrome's win there is notable, but not the decider.
Tab Loading—Winner: Internet—Wait, Seriously?—Explorer 8
Yes, I re-ran this test about three times. But, looking back two tests ago, I realize that Internet Explorer 8 was never a slouch in quickly pulling up content, even when hit with nine tabs at once. In this case, it was a home page for each browser tested, plus Google.com, Lifehacker and Gizmodo, and Hulu. Internet Explorer did quite well.
The other surprise that I couldn't shake after multiple tests? Safari 5 on Windows just would not load up all my tabs at once. It would load the top navigation bars from Lifehacker, Gizmodo, and Hulu, then freeze, with spinning dials on each tab for at least a minute. At least one other user saw similar multi-tab problems. It's not as though browser users are regularly loading up nine tabs at once, but it's still odd to see such an error.
JavaScript—Winner: Chrome (Dev Version)!
Last time, Opera jumped way, way out in the lead, but Chrome's development version picks up the top spot in the Dromaeo tests this time. The drop-off of Opera likely says something about the niggling details of JavaScript testing.
DOM/CSS—Winner: Opera 10.6 Beta!
Safari, Chrome, and Opera could be called evenly matched, given the variables of hardware and testing. Still, Opera's recent obsession with speed is starting to show returns.
Memory Use Without Extensions—Winner: Firefox 3.6!
It's a consistent result lately—with nothing installed, and having just been started, Firefox is really frugal with memory.
Memory Use with Extensions—Winner: Firefox 3.6!
Even though Firefox's extensions would seem to be more full-featured and memory-intensive than Chrome's, which seem to run like advanced user scripts, Firefox is better with a few standard extensions on memory than Chrome. Those extensions, by the way, are the same as in our last test: AdBlock Plus, Xmarks, LastPass, CoolIris, and a Gmail checker.
Here's a full comparison chart, showing each browser with and without extensions, with just one home page open and nine tabs loaded:
The Scores
As you can see, each browser has its strengths—even the most maligned browser in recent history. We took each of the five all-browser tests—start-up times, nine-tab loading, JavaScript, DOM/CSS, and memory use—and assigned a ranking to each contender. In the case of very close results, we assigned a "tie." In two cases—Internet Explorer on JavaScript and CSS, Safari on nine-tab loading—we had to assign zeros, though, obviously, each browser can perform those functions, but that's how it goes.
Here's how the browsers fared, in sum total:
- Chrome (stable): 22
- Chrome (dev): 20
- Firefox: 18
- Opera: 20
- Safari: 14
- Internet Explorer: 9
0 Response to "Browser Speed Tests: Safari 5, Firefox 3.6, Chrome 6, and Opera 10.6 Beta"
Post a Comment
Leave Your Thoughts & We Will Discuss Together