Test Setup
Testing hard disks can be difficult, especially when using a hard disk intensive operating system like Windows Vista 64-bit - give it thirty seconds and it'll spin your hard drives up to defragment or index them, which isn't what we really want when we're trying to ascertain peak performance.
To get a decent idea of drive performance in a variety of real world circumstances we test using a selection of different tools. HDTach 3.040 gives us a good idea of theoretical drive performance, FC-Test’s intensive file transfer abilities let us see real world drive performance, and our GIMP image editing benchmark should tell us if the disk controller is unable to keep up with read/write instructions - something cheap SSDs are prone to, for example.
We've also now chosen to now include random read and write performance figures following requests from
our forum members. Older SSDs in particular suffer in the random write latency department, leading to the infamous "stuttering" effect as the disk lags. While we've been unable to retest every drive and SSD that we've covered in the last nine months, we've made an effort to include the major players and more popular drives.
Finally, we test hard disks in genuine real world circumstances, cloning an install of Windows Vista Home Premium 64-bit onto the drive and then timing the resulting clean boot time before also timing
Crysis level load times.
Each test is performed five times with the highest and lowest scores discarded and the remaining three results averaged, to help give an accurate finding. Below you can see the other hardware used in our testing rig.
Common Components
- Intel Core 2 Duo E6850 (operating at 3.00GHz – 9x333MHz)
- Gigabyte GA-X38-DS5 motherboard (Intel X38 Express with ICH9R southbridge)
- 2x 1GB OCZ FlexXLC PC-6400 memory (operating in dual-channel at DDR2-800 with 5-5-5-15-2T timings)
- ATI Radeon HD 4850 512MB
- PC Power & Cooling Silencer 750W PSU
- Windows Vista Home Premium x86-64
- Intel inf 8.3.0 WHQL
- ATI Catalyst 8.11 WHQL.
Want to comment? Please log in.