Saturday, April 9, 2011

TenFourFox Final vs. Beta 12

On the TenFourFox development blog, the 4.0 final release came with a caveat that a last minute bug fix caused a 10% to 15% slowdown in javascript performance (stability wins over speed). This bug wasn't present in the betas, just the release candidate. The developer then released another quickie update that he said largely corrected the speed hit, so I wanted to check out which was faster, the last beta (12) or the updated final release (4.0s).

I tested both browsers with Sunspider and here were the results (lower number is faster):

Beta 12 - 7392.0 ms
4.0s - 7724.3 ms

Yeah, I know, but I'm on a Sawtooth. Cut me some slack.

Then I went over to Dromaeo, and the results were (higher number is faster):

Beta 12 - 18.89 runs/sec.
4.0s - 16.78 runs/sec.

So in Sunspider, the beta's javascript performance remains about 4% faster than the final, and slightly faster still in Dromaeo, though these results aren't always absolutely consistent when repeated. However, the beta was consistently very slightly faster.

So should you use the beta instead of the final? Well, the final had security fixes, but I have a need for speed so I'll stick with the beta.

Danger is my middle name.

Addendum: Lest you think the final release is some kind of slouch, it still beats the pants off Firefox 3.6, whose Sunspider score was 18786.2 ms, a full two and a half times slower. So if you're using Firefox 3.6 and not TenFourFox, then you probably haven't read this far yet 'cause your renderer is so slooooooooow.