As with my previous browser tests, I installed completely fresh copies of the three browsers on my Windows Vista laptop, with all settings left to defaults. With the second beta of Internet Explorer 8, I reset the browser to factory settings and chose whatever Microsoft suggested during the click-through setup.
My test system has the same specs as before: A 2 GHz Intel Core 2 Duo processor, 2GB of memory, and running Windows Vista Home Premium. For the time-based tests, I again used Rob Keir's ultra-lightweight timer app, simultaneously tapping the "\" key with "Enter" to launch a browser shortcut or folder full of bookmarks. I performed each test on each browser three times and averaged out the results, while eliminating obvious oddities. (With Vista's often empirical hard drive usage, there were definitely artificially long start-ups).
It's the same system I used to test Internet Explorer 7, Firefox 3 RC3, Safari for Windows, and Opera 9.5, so you can make fair comparisons between all the browsers. It's not scientific in the strict sense, but it's meant to measure browser performance as real humans experience it—load, click, and wait.
Test 1: Startup Time—Winner: Chrome!
Drawing inspiration again from Mark Wilton-Jones trend-setting tests, I timed each browser loading up "cold" load (straight off a system restart) and "warm" (having run twice already). I used a locally-saved copy of Google's minimalist home page to negate net connection variations, and, to compensate for Vista's start-up fickleness, timed each browser exactly two minutes after boot-up. Here are the first results:
Note the small scale of the time on the X-axis: Even though Chrome was (quite surprisingly) slower at startup than Firefox or even IE 8, it's less than a second of difference between them all. That's a bit more than an error from my twitchy fingers, but probably not enough to rate any one browser on. Let's check out the warm boots:
As you can see, Chrome's noticeably fast on reload, although all the results are so close it's hard to confidently crown a winner. Just like last time, IE 8 slightly edges out Firefox on warm boots, but lags just a bit behind when starting up.
You don't start your browser to look at clean, white, locally-saved pages, do you? No, you speed around your must-visit sites, and often keep a bushel of them open at once. For the next test, I led each browser page-by-page through the assortment of web sites pictured at right—some heavy with interactive elements, some just text and pictures—before jumping back to a blank page (entering
about:blank does this in any browser) and loading all the links at once. Each browser keeps a spinning icon on tabs as they load, so I measured from first click to the last tab settling in.
IE 8 and Chrome clock in too close to call, but Firefox fell behind. Based on the minuscule difference in cold-boot time and the two warm tests, I'd call Chrome the fastest, but definitely hand IE 8 a Most Improved Player trophy at the awards banquet.
Firefox bests Chrome in this test by a handy lead, while IE 8 takes nearly twice as long (in milliseconds, of course) to perform all the actions Sean runs it through. It's anybody's guess who's got the most objective test—CNET's testers show Chrome wrecking all comers, while Mozilla's own tests declare their orange scrapper the winner in tight races. I can only take away that IE 8 is definitely an improvement from IE 7's fall-behind pace, while Chrome and Firefox are pretty evenly matched...
...until I ran the CSS tests, that is. CSS determines the layout and appearance of a page, and nontropp's downloadable form makes a browser work like a page designer on an all-guarana-and-coffee diet.
Test 3: Memory Use—Winner: Firefox!
How far the great-great-nephew of Netscape has come in its respect for your system's resources. Measured by Vista's Task Manager from cold boots and then with eight tabs loaded, Firefox shows some serious savvy with megabytes:
Do note, however, that Chrome handles tabs differently than others—each tab loads as its own process, so that if it crashes or stalls, the rest of your reading doesn't go down with it. So if you've got solid-state chips to spare, it's not that much more of a hit to run Chrome in a busy session.
As with our last test, we'll note that browsing is much more than speed and bit usage—many of us can't imagine web life without our favourite extensions, or Windows integration, or, soon enough, Chrome's unique features.
What's been your experience with the newest competitors in the web field? Got your own criteria to compare? Share it all in the comments.
Kevin Purdy, associate editor at Lifehacker, wrote this feature in all three beta browsers. His feature Open Sourcery normally appears Fridays on Lifehacker.