Benchmark Results: Audio And Video Encoding
Before we move on to the all-important game tests, we wanted to see what effect Alienware’s surprise lead in Sandra might have in CPU-constrained applications. We begin with Apple iTunes, which has proven itself to be poorly threaded and frequency-bound in most of our previous reviews.
The 3.2 GHz M17x should fall slightly behind the 3.33 GHz Leopard, unless the Leopard isn’t using its maximum Turbo multiplier effectively. We checked it, and found that while the system did occasionally reach 3.3 GHz, it more often ran at 3.2 GHz under light single-threaded loads.
Whether a result of cooling or programming, better engineering puts the lower-clocked M17x ahead of the faster X8100 in HandBrake. Though the difference is small, the victory of a less-expensive part is significant in this well-threaded application.
DivX results are atypical for the M17x, and that’s a problem we’ve seen before in our System Builder Marathons. We don’t change anything during the benchmark’s installation, but will point out that this particular benchmark only shows inconsistent results rarely. Xvid shows realistic results while running under the same TMPGEnc installation.
The super-expensive notebooks almost catch the desktop in MainConcept, and the X8100 again appears to throttle its CPU slightly.