So this benchmark measured "how the population actually experiences web browser performance", and the article pronounces Google Chrome as the obvious winner, with the slight sidenote that FF 5 has a faster "perceived render time"?<p>It just seems slightly disingenuous.
"The second metric, perceived render time (green), refers to the amount of time it takes for the visible portion of the page to load in the browser. Again, Chrome did well here (2.374 seconds), but in this case, Firefox 5 did better (2.18 seconds)."<p>I bet this gap gets larger the more tabs you add. Chrome starts choking after about 40-45 tabs... Firefox can handle 200 easy.<p>The problem with Firefox is that "tab groups" function is horribly broken(as in, should not have been included yet) and that Tree Style Tabs just encourages opening hundreds of tabs.
Chrome is very fast and I do love it, but it still doesn't have a hardware accelerated HTML5 canvas. So if you open up a HTML5 game, IE and FF5 skip along nicely, but Chrome gets very choppy.<p>I hope they fix this soon, because if your "real world" involves "playing HTML5 games", Chrome is one of the slowest!<p>Edit: here's a little HTML5 "game", try it in IE9, FF5 and Chrome: <a href="http://www.scirra.com/labs/ghostshooterfullscreen/" rel="nofollow">http://www.scirra.com/labs/ghostshooterfullscreen/</a>
I find Chrome for Os X as being very "aggressive" with caching.
Sometimes it's misleading as what I am seeing is not what I should see.
That forces me to clear the cache more often than other browsers.
I love the speed but start to wonder wether the price to pay is worth.
I'm not even sure what I am saying makes any sense, does it?
It would be interesting to see some additional data - for instance, the load & perceived render times for each browser under different operating systems - how is Safari on OS X vs Safari on Windows? Chrome & Firefox across Linux, Mac, and Windows?<p>I imagine that Safari is predominantly on OS X, and IE on Windows. Does that affect the load and render times? I guess since the two fastest (Firefox & Chrome) are also on all platforms, and probably more equally distributed, that would suggest that the OS isn't as important - but without more information, that is just a guess.
I only know one thing for sure around this discussion:<p>FF runs netflix without hitching on my 2008 macbook. Chrome drops frames about once a second and I get an un-watchable movie.<p>This very well could be because of Silverlight, but I don't know enough of about the underlying connections between Silverlight and the browser to judge this. As an end-user the experience is still the same.<p>I do, however, still use chrome for everything else. The unified search/url bar hooked me and I can't go back.
I'm not sure how big a difference it really makes, but you can tweak the "perceived" render speed in FF by changing how long it waits between starting to get data and rendering the partial page. Go to "about:config" and add a new integer called "nglayout.initialpaint.delay" to something lower than the default of 250 (milliseconds).
And how much faster will page load times be with thorough blocking of ads and the kind of third-party scripts needed to gather this data? I'd bet even a halfway-decent ad blocker will make far more difference than choice of browser.
I'm curious about how much of this could be caused by the confounding influence of real world Chrome users being more technically savvy and therefore more likely to have good broadband connections.