Benchmark Results: Transcoding
Right off the bat we’re able to tell the difference between a processing architecture designed for the desktop and one built with low power consumption as the first priority. Nettop connoisseurs will say that the point of a platform like Pine Trail is its power draw. But when you make the leap to desktop computing, it’s important to take the tasks we take for granted, like a simple iTunes transcode, into consideration.
The Atom D510 shaves more than 30 seconds off of an Atom 330/Ion, but both are blown out of the water by the Pentium E2200. We’re reminded that, in order to use any of these super low-power setups for anything beyond Web browsing or word processing, you’ll need to be particularly patient.
Whereas iTunes isn’t threaded, MainConcept is. Thus, the dual-core Hyper-Threading-enabled Atoms stack up to the dual-core Pentium E2200 a little better. Even still, you’re looking at 12-13 minutes on either Atom setup or roughly six minutes on the Pentium.
It’s interesting that the Ion-based arrangement actually does better here, despite its less-power processor and slightly-lower memory bandwidth. This is a trend we’ll see continue in some of the other media-oriented tests, and a trend that will be bucked in the benchmarks more centered on storage.
New to the Tom’s Hardware 2010 benchmark suite (and a popular request from our readers) is the freely-available HandBrake 0.9.4. In this test, we’re using the High Profile to transcode the first .vob from The Last Samurai into .mp4.
It’s a process that takes roughly one hour on either of the Atoms and less than 30 minutes on the Pentium. And once again, the dual-core Atom 330 is able to slide past the new Atom D510 by finishing the transcode job four minutes faster.
Using the latest version of TMPGEnc, DivX, and Xivd, we again see the Atom 330 beating out its successor coupled with Nvidia’s Ion platform. Of course, in Xvid, which isn’t very well threaded, the Pentium E2200 boasts its most significant lead. Shifting over to DivX, where threading optimizations smile warmly on the inclusion of Hyper-Threading, the Atoms make up some of their performance deficit (though the D510 still takes almost three times as long to finish the job).
Finally, here again we see this-generation’s Atom D510 getting bested by the Atom 330 on Nvidia’s Ion platform.