As I’ve alluded to in the past, I’m a bit of a browser tart – I’m forever flitting from one to the next when I start to get annoyed with a feature – or a lack of a feature within the current one. Lately I’ve been alternating between Safari and Chrome. My previous love affair with Chrome ended when it stopped displaying images in Flickr and the lack of bookmark sync through MobileMe started wearing thin too. Safari ticks most of the boxes but it does feel a little sluggish. Firefox 4 is still in beta, but it’s coming on quite nicely and whilst I’ve never spent much time with Opera, now seems as good a time as any to see what it can do.
I’ve decided to take a look at some HTML5 benchmarks to start with. HTML5 may still technically be in draft, but it’s never too early to ditch Flash. I’m testing on my Macbook Pro (MB990) with 4GB RAM and a 2.26ghz Core 2 Duo. The browsers I’m testing are Safari 5.0.3, Chrome 10.0.648.114, Firefox 4.0 11b, and Opera 11.01. First up, Clubcompy.
Clubcompy
Clubcompy claims to be a “real-world browser benchmark” and you can try it yourself here. It runs through 5 different tests and comes out with a score at the end. I ran it twice with each browser and took an average.
Chrome takes a strong lead here, outclassing the other browsers in the test. Safari and Opera take a close 2nd and 3rd with Firefox 4 bringing up the rear. I know Firefox is still in beta so I’ll reserve judgement for now. Beta 12 should hopefully see a bit of an improvement. It is interesting to see two browsers that use Webkit score so differently – Google’s obsession with performance and tweaking its V8 (cool name, mind) Javascript engine appears to pay dividends in this test.
Ok then, next test. This time the old favourite – Acid 3.
Acid 3
Acid 3 is more of a “standards” test than a performance benchmark, but it does provide an elapsed time for a completed test. Firefox 4 failed 3 of the tests and scored 97/100, whereas the other three passed all of the tests. I’ve included Firefox time in the graph for comparison, but it doesn’t really mean anything.
Acid 3 - Time - Lower is better (obviously)
Again, Chrome comes out on top here, completing the test in 2.29 seconds. Safari took 2nd place this time with 3.07 seconds and Opera in last place with 3.85 seconds (decimal time not actual time…evidently).
As I’ve said before, the Acid 3 test is more about standards compliance and the Webkit browsers excel here. Firefox is coming along nicely and not even the current build of Internet Explorer 9 comes out with 100%. Opera gets there, it just takes a little longer than its rivals.
Now, just to spice things up a little, I’m going to unleash Mozilla’s own benchmark called Kraken. The reason why I’m doing this is to demonstrate the murky world of dodgy benchmark claims by various developers. Firefox has been soundly beaten in each of the tests I’ve tried, but surely it won’t come out on top in Mozilla’s own test???
Lower is better...if a little dodgy
Well, fancy that – the slowest browser so far…wins. I’d disregard this test entirely and stick to the independent ones. Performance claims made using an in-house benchmark can safely be ignored. Similarly Microsoft’s claims about IE9′s performance and standards compliance is likely to be just as skewed.
Update
I downloaded the latest version of Firefox 4 (RC2 I believe) and re-ran Clubcompy just to see if any performance improvements had been made since I tested the beta. The score was about the same as before, maybe a little lower.
Related posts:
