NIce article although the author doesnt seem to know that there are two measures of a disks performance
1) Throughput - measured in MBytes/sec
2) Bandwidth - Measured in IOPS

These two measures are mutually linked ie 100MB/s can either be 1*IOPS*100MB, or 100*IOPS*1MB

The figures seem to also show a lot of cache effects, or have dubious provenance,  otherwise how does a USB connected drive achieve 2289.5 IOPS (how do you get .5 of an IOPS anyhow)

Hi Gerald,

I do understand there are several things to take into consideration when measuring the performance of any hard drive, however going into depth in so far as eliminating cache effects would go beyond the scope of what this article is attempting to achieve. You may have also noticed that I didn't mention or talk about the effect FSB speeds can have on a hard drive's performance either?

As I mentioned at the start of the article;

The reason you might want to skip the QD32 tests is that they're not particularly useful for real life scenarios and just take additional time to run. QD stands for "Queue Depth" so the software also measures "separate thread" transfer speeds. Information not of much use for our purposes. I only include the results that I got below for the sake of completeness.

I think the general flow of the article from that point forward clearly shows it is targeting beginners to hard disk performance measuring and not representing any type of advanced hard drive performance analysis.  

The figures seem to also show a lot of cache effects, or have dubious provenance, otherwise how does a USB connected drive achieve 2289.5 IOPS (how do you get .5 of an IOPS anyhow)

Indeed.  How do you get a .5 of an IOPS ?  

Or even a .7 or .8  

How about you tell us?  :-)

Thanks for your feedback.  Always great to hear back from ones readers.

Best regards,
Andrew