Hands on: Boost system performance using Raid

Raid uses two or more hard disks working together as a single logical volume to deliver various combinations of greater reliability, performance or capacity. Raid is a popular option for the desktop thanks to the low cost of hard disks and the inclusion of Raid controllers on modern motherboards. Here we expand on an earlier article on Raid performance, with tests on four-disk arrays.

Test configuration
We built a system around an Asus N4L-VM DH motherboard with an Intel Core Duo T2600 and 2GB of Crucial Ballistix DDR2 memory. The Raid arrays were built and managed by a Promise Fasttrak TX4310 PCI host adapter. We used up to four identical Seagate 300GB 7200.9 Barracuda hard disks for each Raid array. The main system booted from a separate Seagate 400GB 7200.8 disk. We used a separate disk for the OS because HD Tach can only fully benchmark unpartitioned volumes.

Each Raid array was built using the Promise adapter’s Bios controls using the default 16KB stripe block size. For comparison with earlier results, we ran HD Tach’s Long Read mode. We tested three types of Raid array: 0, 1 and 5. There’s a full explanation of pros and cons in our earlier article on Raid.

Results
We first repeated our tests with arrays built from two and three disks. Starting with two disks configured as Raid 0, the array exploited the full 600GB capacity and delivered average read and write speeds of 98.6 and 74Mbytes/sec respectively, and a burst rate of 106.7Mbytes/sec. Rebuilding this as Raid 1 halved the available capacity to 300GB and delivered average read and write speeds of 59.9 and 44Mbytes/sec respectively and a burst rate of 1 02.5Mbytes/ sec. Raid 5 is not possible with two disks.

Raid 0 delivered faster performance and the maximum capacity at the cost of reliability. The Raid 1 array offered half the total capacity and delivered the average speed of a single disk, but sported 100 per cent redundancy.