I bought this enclosure for the performance. With USB 2.0 maxing out at 480mb/s and Firewire 400 maxing out at, well, 400mb/s, both interfaces become bottlenecks for high performance external hard drives. Typically, USB 2.0 external hard drives can reach towards 25MB/s transfer rates, which is significantly less than what high performance desktop drives can achieve. Even many laptops drives (aside from the 4200RPM ones) can sustain transfer rates above the saturation point of the USB 2.0 connection. To unlock the full transfer rate potential, eSATA is the way to go, along with a PCMCIA/ExpressCard eSATA adapter for a laptop.
I will be testing the difference between using USB 2.0 and eSATA as well as using eSATA versus an internal SATA port, to see if there is any overhead associated with the external SATA implementation. A newly purchased Seagate 500GB 7200.10 SATA 3.0gb/s hard disk with 16MB cache will serve as the drive being tested in all cases. The SATA 1.5gb/s limiting jumper was removed from the drive to take advantage of the faster speeds of SATA 3.0gb/s where possible.
For the test system, the following was used (relevant details listed):
Intel Core 2 Duo E4300 1.8GHz 2MB L2
Biostar TForce 965PT (on board ICH8 SATA was used)
2 x 1GB DDR2
First up, I tested the X Craft 350 in USB 2.0 mode against an older Vantec Nexstar NST-350 with a 200GB Western Digital WD2000JB. The difference in hard drives shouldn’t make a difference as both the Seagate and WD can saturate the USB throughput long before they reach their own performance bottlenecks. This should isolate the quality of the USB controllers used.
As you can see, the X Craft 350 is slightly behind the Vantec in both average throughput and CPU usage. Not terribly inspiring so far.
Next up is the test of what I’m interested in: eSATA performance. It doesn’t disappoint.
As you can see, average throughput is more than doubled and burst rates are just about quadrupled. As well, CPU usage is minimal and the drive shows up as a native hard disk to Windows, so SMART monitoring can be enabled. In comparison, in USB mode, the X Craft will be detected as a USB removable storage device. Performance is actually on par with SATA 1.5gb/s standards. The reason for this may lie in Cooler Master’s recommendation to use the SATA 1.5gb/s limiting jumper on SATA 3.0gb/s drives. Although I removed the jumper, the mere fact that Cooler Master suggests this leads me to believe that there is some physical limitation relegating it to 1.5gb/s performance. This is not isolated to Cooler Master – Vantec also suggests using the limiting jumper with their Nexstar 3 eSATA enclosures.
Finally, I tested the drive’s performance connected directly to one of the SATA 3.0gb/s provided by the ICH8 southbrige. As you can see, this is where true SATA 3.0gb/s performance appears. Average transfer rates are actually pretty close, but burst rate is almost 50% faster, blasting through the 150MB/s theoretical maximum of SATA 1.5gb/s.
Benchmarks are nice, but real world performance is more important. Vista’s file transfer dialogue now includes a transfer rate measurement.
Cooler Master X Craft 350 USB 2.0 Vista Performance Dialogue
Cooler Master X Craft 350 eSATA Vista Performance Dialogue
As you can see, eSATA performance is significantly better than USB performance in real world testing as well. More than double the transfer rate is nothing to sneeze at.