My first impression was that this is definitely not production-ready code. The browser crashed several times on me during the process of capturing the benchmark. But when I did get the tests to run, it was fast. I mean visibly fast. Much faster than anything else I have on hand.
As before, the testing wasn’t extensive. I still haven’t tested Firefox 3 with TraceMonkey. The only test I performed this time was running the Safari nightly through Dromaeo. Here’s how it stacks up against the others:
Safari with SFX ran the test in 8184.6ms. The exact build of Safari used for this test is Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US) AppleWebKit/528+ (KHTML, like Gecko) Version/3.1.2 Safari/525.21. The nighly package that I downloaded was labeled WebKit-r36647.
Well, that’s the impression that I got from using it for a few days anyway. But yesterday I decided to go one step further and verify that feeling of greater performance. Admittedly, this isn’t the most scientific approach, but I have some numbers for you. I used one benchmarking tool, Dromaeo, to see how Chrome stacked up against my workhorse Firefox setup. John Resig has already run a similar comparison, but I wanted a firsthand look at the numbers. Here are those numbers:
Chrome ran the tests in 12658ms. The build of Chrome used for this test is Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US) AppleWebKit/525.13 (KHTML, like Gecko) Chrome/0.2.149.29 Safari/525.13. Firefox ran the tests in 13245ms. This was the 3.0.1 release build of Firefox, Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:18.104.22.168) Gecko/2008070208 Firefox/3.0.1. The tests were run on my Core Duo 1.8GHz 4GB Windows XP SP2 system.
As you can see, Chrome was indeed faster. I’m already using it more and more for Google apps, as I’m figuring that these will probably always run faster in Chrome. Firefox still is my main browser, though, because of the extensions. Chrome won’t be replacing that stuff any time soon.