Navigate: Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home	.	.
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ Search Prices:

RAID Level0 Performance

As we showed on the last page, the burst performance of the HPT controller appears to be less than on par, at least in the particular algorithms used by HDTach. Since the sequential transfers by definition have to be lower than the burst transfers, the next exercise was to look at a Level0 performance of the MHT2060AH SA drives on the two PCI-based controllers. Keep in mind that the sequential transfer rates of the MHT2060AH SA was already about 1/2 of the burst transfer rate of the Highpoint controller, which led us to expect some serious speed matching conditions in a Level0 configuration with the RocketRaid controller whereas we did not expect the same degree of problem with the Fast Trak controller.

HighPoint RocketRaid 1650 RAID Level0 performance

Even though the burst rate has increased to some 72 MB/sec, it is still not enough to accomodate the sequential performance of two MHT2060AH SA drives. This speed matching condition is rather obvious from the alternating high-low-high-low zigzag of the graph.

Promise FastTrak S150 TX4 RAID Level0 performance

Moving to the Promise Fast Trak controller with its higher effective host transfer rate solves the problem in that the host transfer rate stays above the combined media transfer rate. The result is a smooth graph with a sequential performance close to a WD Raptor and more than 3 x the capacity of the latter. The random access performance is, of course not on par with the Raptors.

Again, we have to stress that HDTach results are not necessarily representative for real performance.

In any case, it is rather obvious that two MHT2060AH SA drives are already saturating the PCI interface of the controller cards, there is, therefore, no reason to move on to a 4 way RAID Level0 configuration, suffice it to say also that the failure probability of such an array would be extremely high. The next step, therefore, leads to what we always wanted to have, a RAID Level5 configuration, that is a block-striped array with distributed parity across all drives.

next page:    => RAID 5 Performance =>