[~Time Published March 04, 2014 at 05:52PM]
*/*/*/*/*/\*\*\*\*\*
When we reviewed Samsung's Galaxy Note 3 back in October, we noticed that it was doing something funny. Launching most benchmarking applications would kick the phone's CPU into overdrive, forcing all CPU cores to run at their maximum rated speeds for as long as the application was open. Maxing out the CPU frequencies threw a wrench in our standard suite of tests, inflating some benchmark scores by around 20 percent.
This wasn't the first time Samsung had been caught, um, "enhancing" the performance of its hardware—AnandTech had discovered similar CPU and GPU optimizations in the international version of the Galaxy S4 two months before. However, this was the first time we had seen it show up in the North American version of one of Samsung's phones. Even worse, as Samsung rolled the Android 4.3 update out to its older phones, other phones all began to exhibit the same behavior. Our AT&T Galaxy S III and Galaxy S4 didn't max out their CPUs under Android 4.1.2 and 4.2.2 (respectively), but they started to act just like the Note 3 after being updated. Not only were Samsung's benchmark boosting additions intentional, but they had become a "feature" baked into its branch of the Android tree.
We've changed our testing procedures to avoid this particular kind of benchmark boosting, and I was prepared to write it off as an annoying fact of life—after all, while benchmarks are useful as a way to determine the relative performance of a piece of hardware, they aren't the only factor (or even a particularly important factor) in most peoples' purchasing decisions. That was before I heard from John Poole, whose company Primate Labs develops the Geekbench CPU benchmark we use in most of our reviews. After working with Poole and the Primate Labs team for a few days, we can now say a lot more about just how Samsung was implementing the boosting "feature" in Android 4.3, and that it appears to be absent from the company's Android 4.4 update.
Read 21 remaining paragraphs | Comments
*/*/*/*/*/View@Source\*\*\*\*\* (http://ift.tt/1eRjP5c)
[IFTTTautoSHAREv1.015 | Shared with ifttt.com More shared news etc on: http://bit.ly/Schavuiten_blog | RSS source: http://ift.tt/1heth1N]
*/*/*/*/*/\*\*\*\*\*
Samsung's Android 4.4.2 update appears to correct the boosting behavior we encountered in 4.3.
Andrew Cunningham
When we reviewed Samsung's Galaxy Note 3 back in October, we noticed that it was doing something funny. Launching most benchmarking applications would kick the phone's CPU into overdrive, forcing all CPU cores to run at their maximum rated speeds for as long as the application was open. Maxing out the CPU frequencies threw a wrench in our standard suite of tests, inflating some benchmark scores by around 20 percent.
This wasn't the first time Samsung had been caught, um, "enhancing" the performance of its hardware—AnandTech had discovered similar CPU and GPU optimizations in the international version of the Galaxy S4 two months before. However, this was the first time we had seen it show up in the North American version of one of Samsung's phones. Even worse, as Samsung rolled the Android 4.3 update out to its older phones, other phones all began to exhibit the same behavior. Our AT&T Galaxy S III and Galaxy S4 didn't max out their CPUs under Android 4.1.2 and 4.2.2 (respectively), but they started to act just like the Note 3 after being updated. Not only were Samsung's benchmark boosting additions intentional, but they had become a "feature" baked into its branch of the Android tree.
We've changed our testing procedures to avoid this particular kind of benchmark boosting, and I was prepared to write it off as an annoying fact of life—after all, while benchmarks are useful as a way to determine the relative performance of a piece of hardware, they aren't the only factor (or even a particularly important factor) in most peoples' purchasing decisions. That was before I heard from John Poole, whose company Primate Labs develops the Geekbench CPU benchmark we use in most of our reviews. After working with Poole and the Primate Labs team for a few days, we can now say a lot more about just how Samsung was implementing the boosting "feature" in Android 4.3, and that it appears to be absent from the company's Android 4.4 update.
Read 21 remaining paragraphs | Comments
*/*/*/*/*/View@Source\*\*\*\*\* (http://ift.tt/1eRjP5c)
[IFTTTautoSHAREv1.015 | Shared with ifttt.com More shared news etc on: http://bit.ly/Schavuiten_blog | RSS source: http://ift.tt/1heth1N]
Geen opmerkingen:
Een reactie posten