Macworld Forums: Opera 12: A work in progress finally shows real improvements - Macworld Forums

Post your comments for Opera 12: A work in progress finally shows real improvements here

"This free, open-source browser’s designers"? Last I checked, Opera was closed-source

FYI, in the second paragraph below the comparison chart you use 'Chrome' a few times instead of 'Opera'. Still, great overview, thanks for the info.

You can set Chrome as your default browser, or any iOS browser for that matter.

Look up Cydia.

If you add Google's V8 benchmark and very extensive Dromaeo JavaScript benchmark to your browser tests, test results would better reflect real world usage.
Each of JavaScript benchmark has it's own strengths, but I believe SunSpider benchmark is pretty much outdated for modern JavaScript JIT driven runtimes.

You may also consider to test browsers with Peacekeeper (http://peacekeeper.futuremark.com/). It consists of a bunch of browser performance tests.

 bager, on 06 July 2012 - 06:24 AM, said:

I'm puzzled as to why you think wholly artificial benchmarks in any way represent real world usage. Particularly when the V8 benchmark was created specifically to run well in Chrome (it was designed with Chrome's architecture in mind).

 reformedgeek, on 06 July 2012 - 04:14 AM, said:

You're spot-on - Opera isn't open-source. Opera does run on several open-source OSes - the author's point is in error.

 makpe, on 06 July 2012 - 11:14 AM, said:

Benchmarking is a very difficult subject indeed. I do not believe there is a single benchmark representing all points of view, therefore I suggest MacWorld runs as many as possible, SunSpider, Dromaeo, Kraken, Peacemark, ... as each of them has their own strengths. SunSpider tests are so short living that a modern Javascript JIT compilers have trouble detecting and optimizing hot-spots. V8 tests last longer so JIT can optimize hot-spots. V8 tests are more representative of long running applications (e.g. webmail).

P.S: V8 benchmark was not create to run well in Chrome. It was created to improve Chrome JIT compiler.

Interesting... Opera only scored 385 +9 bonus points on my system, a 27-inch iMac running Snow Leopard. It was still higher than Safari and Firefox but below Chrome.

I wonder how much of HTML 5 is system dependent? Or is there something different in the implementation of these browsers on each platform?

 bager, on 08 July 2012 - 07:53 AM, said:

The point is that these artificial benchmarks are still artificial, and the V8 benchmark in particular is useless because it was designed with Chrome (V8) in mind. It's supposed to show off V8 and is optimized specifically for the way V8 does things (and leaves out things it's extremely slow at!), so other JS engines will always be at a disadvantage.