ZDBOp has designed the disk tests in their WinBench 99 suite to stress a system's IDE channels in order to determine a disk drive's peak transfer rate. This number is not as important in games as it is in video editing or content creation.

Without a doubt, these theoretical bandwidth numbers prove that RAID makes a significant difference in hard drive performance, albeit at the cost of another hard drive. It should be noted though, that WinBench 99 is designed to push as much information across the bus as possible, making any weakness in a single drive solution painfully obvious. For a slightly better gauge of how IDE RAID affects real-world performance, we turn to Content Creation 2000.

Designed to test "hot spots," Content Creation 2000 has always been very sensitive to hard drive performance - so much so in fact, that we have purchased new drives for each system in both of the Sharky Extreme laboratories. That's a lot of storage space.

Interestingly enough, the single drive connected to the KT7's ATA/66 port performs better than the RAID array powered by the onboard Highpoint controller chip. Maybe RAID doesn't provide the real-world performance increase we had hoped? It sure seems that way.