System Setup
Hardware
- Freecom FHD-2 60GB USB2.0
- Maxtor OneTouch 300GB FireWire400
- Maxtor Personal Storage 5000DV 160GB FireWire400
- Maxtor Personal Storage 3000DV 80GB FireWire400
- ASUS K8V Deluxe, Socket 754, VIA K8T800 Athlon 64, 1005.011 BIOS
- AMD Athlon 64 Model 3200+, 1MB L2, 10 x 200MHz
- ATI Radeon 9800XT (412/730)
Software
- Windows XP Professional w/SP1
- ATI CATALYST 4.1 and Control Panel
- VIA Hyperion 4.51v
- Simplisoft HDTach 2.61
- ATTOTech Disk Benchmark
- Xbit Labs File Copy Test
600GB of external storage to be benchmarked. You'll notice that I'm only benchmarking the OneTouch 300GB and the 5000DV in FireWire interface guise. The results barely differ from the USB2.0 results, bar CPU usage in HDTach, and the FireWire ends up slightly ahead in the battle of the interfaces on those disks. It lets me keep the graphs simple, use either interface if you end up purchasing either disk.
The USB2.0 driver from Hyperion 4.51v was used in the Freecom case, the standard Windows XP Professional FireWire driver was used for the VT6307 ASIC that the K8V carries. I'll comment on that actual ASIC as the benchmark results unfold, it has some peculiarities worth talking about.
In the case of the Xbit FC benchmark, it was only run once per disk, its duration in some of the tests hampering its use in our usual three run method. I'm confident that didn't affect testing. Being disk benchmarks, the system was rebooted between benchmarks to flush disk caches and OS disk caches that are resident in memory, to give each benchmark and disk a fair chance. Write caching was enabled for all four disks and that has an effect on performance, the operating system buffering write requests without actually sending them to the disk as soon as possible, instead sending them when the disk isn't doing much, so the performance impact of the write ops can be minimised.
I pondered turning it off for each disk and it's usually off by default for externally attached mass storage devices, but the performance impact was negligible in most cases, so it was enabled. The Freecom disk is most affected by write caching, losing most performance when it was turned off, but it wasn't huge.
Onto the graphs.