8 comments

[ 3.2 ms ] story [ 28.6 ms ] thread
Since you are testing all beta versions here, why not use Chrome 7 beta for a new benchmark ?
To be fair, Chrome beta channel is still v6 series (6.0.472.59 beta). The dev channel is v7 series already though I believe.
Good question :) I didn't write the article, maybe ask @lucidchart on Twitter? According to the actual article title though, it seems the main interest was testing FF4 and IE9 against existing browsers, which would explain the absence of dev channel versions of Chrome.
Or more interestingly, they were looking at providing the best numbers to push their agenda the furthest.

As noted in a previous comment, a fair comparison would be to run a proper benchmark--which guarantees similar operating conditions.

At the very least they could've tested the current stable release of Opera - 10.62 - instead of 10.53.
Fx4b6 does not include the new JaegerMonkey javascript engine, so the lack of a speedup compared to Fx3.6 is not terribly surprising.

JaegerMonkey is in Fx nightly builds at the moment and is supposed to be featured in Fx4b7 (to be released later this month).

Note that it's an "user drags something in the UI and we measure something" benchmark, so it's potentially suspicious if the things measured are comparable at all in different runs and browsers.

For example, in games, frames per seconds are measured under basic assumption that the game has the mode in which there's no limit the increase of frames per second other than the speed of rendering and that that mode is always identically reproducible.

Are we sure about that here?

And as others noted, they are using the old version of Opera, 10.53 instead of 10.62 which is current.