SunSpider benchmark updated

The WebKit developers have announced SunSpider 0.9.1, an update to the popular JavaScript benchmark designed to improve the accuracy of its results. The major change in SunSpider is to allow for the combination of modern high speed JavaScript implementations and power management. When SunSpider was launched, typical results were around 10,000ms. As JavaScript implementations were improved this has gone down to as little as 300ms.

But, by design, SunSpider paused between each test. Where a system had some power management enabled, the combination of the pauses and the very quick benchmark made it appear to the system that it was nearly idle, or at least idle enough to reduce the CPU clock rate to save power. This in turn meant the subsequent tests would not run at full speed. The developers have reduced the delay between tests and reduced the amount of loading the benchmark does. The result is an "added bonus" as SunSpider now runs more quickly after losing the half second pauses.

The SunSpider update includes a number of other fixes to correct minor errors in some of the tests and the HTML code has been modified to be valid HTML5. The new SunSpider can be found at the WebKit.org SunSpider page. The developers recommend that SunSpider users update their benchmark scores with the new version as soon as possible.

(djwm)