Navigate: Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors Home PC Hardware from Made-in-China.com Search: Enter your search terms   Web www.lostcircuits.com Submit search form	.
Prices:

Theory and Practice: HDTach

The IBM 60 GXP datasheet shows sustained data rates of 40.8 to 20.9 MB/sec from the outer to the inner tracks, how do these numbers correlate with the measurements of HDTach? The Maxtor D740X-6L rates slightly higher on paper, that is 44.4 MB/sec to 24.2 MB Sec (outer vs inner diameter of the disks), in other words, the average sustained transfer is about 3.5 MB higher. Let's see whether this shows in HDTach

HDTach 2.61 performance of the IBM 60GXP. The values are almost identical with those given in the IBM datasheet. The inner track performance slightly lags behind specs, however, keep in mind that this drive is rather full and any data contained on the drive will result in lower test results For our purpose, the drive is right on target.

HDTach 2.61 performance of the Maxtor D740X-6L. Another very close match with the manufacturer's specs and not only that but also confirming our claim that the D740X-6L would score an average of roughly 3.5 MB/sec higher throughput than the 60GXP. Moreover, despite the fact that the drive itself is theoretically capable of delivering up to 133 MB/sec, in reality, this value can hardly be reached because of chipset bottleneck issues.

The next question we asked was whether the Maxtor/Promise IDE adapter would change the results:

HDTach 2.61 performance of the Maxtor D740X-6L running off the ATA133 controller. The performance is virtually indistinguishable from what we got with the IDE interface. The access time measurements of HDTach were showed too much variability between individual runs to be statistically significant. The same consistency between the IDE and the ATA133 interface was found with the IBM 60GXP which scored identical with or without the PCI adapter.

So far, we have seen what everybody else has seen and could just leave it like that. SiSoft Sandra 2002 File benchmark showed the same picture except that the Promise controller was actually somewhat slower than the original IDE interface. PCMark2002 showed the Maxtor drive ahead of the IBM with an average of 868 compared to 810 (Maxtor vs. IBM). The results are very clear, the Maxtor drive is faster and there is no benefit from an external ATA controller.

Fine for a foreplay but we were just warming up. The real issue is that there are strong differences between synthetic benchmarks and real life performance and we used one of these archaic monsters that ZD-Labs coughed up a few years ago and which still gave a detailed listing of what was going on at what point. Enter Winbench98. I can already hear the argument that the benchmark is outdated etc. It does not matter. If there are applications that show differences, these differences are valid beyond the read-out of synthetic benchmarks and not only on paper but in real life as well. Shall we?

next page:    => WinBench98 Business Performance =>