Called Kraken, Mozilla described the new benchmark as a “step in the right direction.” Engineer Rob Sayre said that it focuses on “real workloads” and reveals innovation progress. In terms of performance, Mozilla believes that benchmarks should reflect benchmark recommendations from Hennessy and Patterson’s “Computer Architecture” which lists the following criteria, exactly in that order: Real applications, modified applications, kernels, toy benchmarks and synthetic benchmarks. Kraken takes that order and prioritization into account.
So we ran our benchmarks and came to a surprising result (or not so surprising result, depending on your view). Firefox is faster than Chrome in JavaScript. Mozilla criticizes Sunspider as having too many toy benchmarks, but it is interesting to see that Mozilla has made huge progress in Sunspider nevertheless.
According to our test results, the current stable Chrome 6 finishes the test in 19,975 ms, while the latest Chrome 7 nightly build (524.0) arrives at 19,262 ms. Firefox 4.0 Beta 6 finishes at 22,970 ms, but Firefox JS with the new JaegerMonkey engines blazes through the test in only 10,096 ms.
You could blame Mozilla for some unfair fine tuning. That may be very well the case, but Firefox JS also posts a time of 414.6 ms in Sunspider, about 10% below IE9 PP4 and 20% below Firefox 4 Beta 6. Chrome still leads the field with 302.4 ms on our test system for the Chrome 7 nightly builds and 358.1 ms for the stable version of Chrome 6.
We would have loved to run IE9 Beta through the tests as well, but we were not quite important enough to receive an early look at IE9 Beta. We have been waiting for our IE9 Beta download to finish for the past hour and will run the benchmarks and update our article as soon as we can.
this post from: http://www.conceivablytech.com/2784/products/mozilla-firefox-4-is-twice-as-fast-chrome-7/
No comments:
Post a Comment