HDTach 2.61 Benchmark
The HDTach 2.61 benchmark specifically tests sequential and raw burst speeds for drives (irrespective of filing systems), together with associated CPU utilisation. In particular, the plotted graph shows the speed of transfer over the entire capacity of the drive.
HDTach scores for UATA-100 Drive
HDTach scores for RAID_0/ATA-100 config
From the above two graphs, which compares the standard ATA-100 vs ATA-100 RAID configurations, several pertinent conclusions may be drawn:
- The RAID setup only offers 20 - 25% increase in sequential read speed improvements over a normal single ATA-100 drive setup. And for burst speeds, the improvement becames even more marginal, at just 8% faster for a RAID setup. Typically, the ATA-100 RAID interface seems to produce speeds akin to an ATA-66 RAID setup, as seen in my previous review on the Iwill SIDE RAID 66.
- As with most RAID setups, a higher CPU utilisation penalty is incurred (17.5% vs 9.1%);
- Hopefully, better AMI Win2K drivers would improve arbitration performance for the RAID setup, whose inconsistent and "wavy" plot could indicate inefficient reads across striped disks. Yet, this may also be issues pertaining to hardware limitations (i.e. HDD or controller interfaces), where an expected double-fold improvement over blistering ATA-100 speeds is typically unachievable. Strangely, ATA-66 RAID controllers don't exhibit the same "jagged" phenomenon as seen in my previous findings.
Yet, even though the AMI MG80649 seemingly under-performs, the same problems seem to be exhibited by all other ATA-100 controllers as well. Hence, maybe this could just be an inherent issue with HDTach's handling of RAID ATA-100 setups. Specifically, my fellow editors - Kan and Wilfred, have had similar experiences with a Promise ATA-100 RAID and HighPoint 370 chipsets. Hence, regardless of whichever ATA-100 RAID chipset opted for, this inherent issue would inevitably surface. In that regard, Iwill shouldn't really be faulted.
Sisoft Sandra 2000
Sisoft Sandra 2000 also offers a benchmarking section for HDDs, which are represented below for ATA-100 RAID setups:
Sisoft Sandra scores using NTFS drives (TA-100 RAID)
Sisoft Sandra scores using FAT32 drives (ATA-100 RAID)
Here, we see more representative scores achieved from "RAID-ing" ATA-100 drives. In particular, the drive index indicates performance twice as fast as a single ATA-100 variant (refer to earlier section), and exceeding 2.5 times that of an ATA-66 drive. As a secondary observation, NTFS shows an incremental speed benefit over the FAT32 format.
Now, having quantified the value of a RAID_0 setup, let's see what RAID_1 offers...