Memory tests, Pifast, WAV encoding
A look at memory bandwidth and latency first.
A lack of PAT enhancement and looser memory timings account for the ~10% difference in reporting memory bandwidth scores.
The same can be applied for latency analysis. So on our clock-for-clock basis, I expect MV's Ixius to be a touch slower in 2D benchmarks.
That's pretty much the case for Pifast. The performance difference is only noticeable in graphs. I couldn't detect a difference in speed on a purely visual basis. Just remember that the Ixius 3.6 is being directly compared to the fastest non-Extreme Edition Intel Pentium 4 desktop setup that money can currently buy. It starts to look impressive in that light.
There's nothing to choose between them when encoding wave files into 192kb/s MP3.
