SFF Roundup, Part I: Socket 478 and 754 Systems

Encoding and Network Benchmarks

For our encoding benchmark test, we used AutoGK to encode the high quality TV input from the Aopen into a 5 MB file using two passes and timed how long it took. We used both the DivX 5.2.1 and Xvid 1.0.2 with audio encoding enabled. While some might find it useful to test the video and audio encoding separately, with the assumption that SFF computers would often be used as HTPCs, we wanted to show how a real-world test runs. Video size was set to a fixed width of 640 pixels (i.e. no loss in video size) and audio was set to auto. We then timed the entire encoding process from start to finish. We have samples of the DivX and Xvid encoded video files, and when compared with the 28MB source file, you can see the benefit of re-encoding.

We also ran some rudimentary performance tests on the network connections, just to verify that gigabit Ethernet performance was reasonable. We copied a 4 GB file from one system to each computer - with both connected to a switch - and timed how long it took. Since the source computer was running a standard 7200 RPM SATA drive, we should be HDD limited, but even then GbE is still three to four times faster than 100Mb.