[ The PC Guide | Systems and Components Reference Guide | Hard Disk Drives | Hard Disk
Performance, Quality and Reliability | Hard Disk Performance |
Hard Disk Performance Measurement ]
Putting Performance Measurement In Perspective
"There are three types of lies: lies, damned lies, and statistics."
-- Benjamin Disraeli
I seem to have a habit of taking somewhat controversial stands that fly in the face of
what most other hardware writers say on various subjects. This page will certainly be no
exception. :^) Here it is, blurted out in simple terms: hard disk performance is
important, but hard disk performance measurement is often overrated and
over-emphasized. The Internet is filled with discussions about various benchmarks;
arguments over which benchmarks are better and worse; people getting upset when their
hardware gets benchmarks scores that are "too low"; review sites recommending
one hard disk over another on the basis of it scoring a few percentage points over another
in a synthetic test; and so on. In my opinion, most of this has little relevance to the
way most people use their PCs, and the people who are so concerned about "which hard
disk to get" when they are looking at comparable models are worrying too much.
Sure, it's useful to see benchmark scores on hardware before you plan a purchase. But
you should keep them in perspective. An important rule of thumb in the PC industry is that
in real-world computer use, most PC users don't notice an increase (or decrease) in
performance of less than about 10%. That means that even if you could prove without a
shadow of a doubt that hard disk "A" was 7% faster than hard disk "B",
you probably wouldn't really notice any difference in your system. And in most cases,
competitive products put out at about the same time by high-end manufacturers often don't
differ in performance by more than 10%. The biggest mistake people make is to over-value
the numbers they read about various hardware benchmarks. Use them as a rough guideline
only. A month down the road you probably won't even remember what "your benchmark
scores" were, and they certainly won't matter much to you. I've been down this road,
believe me.
|
Here's an example of the type of benchmark chart you
will run into all too frequently (and one that would make
Disraeli shake his head). At first glance, drive "B" appears
to blow the doors off of drive "A", right? In reality, the
numbers differ by less than 4%--not even noticeable to most
users. The scale has had its zero point "chopped off" to magnify
the differences between the numbers. Note also the other
marketing tricks: the larger number has its column in bright red
to attract the eye; the smaller number is a plain light blue. The
larger figure is also in bold and a larger font has been used. |
Also remember that there are other attributes besides performance that are important
for selecting hardware. Lots of people try to get the very fastest hard disk but don't
consider other equally important issues: quality, reliability,
warranty, and data backup. People agonize
over which hard disk is a teeny bit faster than another--and then never defragment their
file system, or fill the hard disk up with junk so it runs less efficiently. Be sure to
keep the big picture view.
Finally, bear in mind that whatever is on the top of the hill in the hard disk world
doesn't stay there for long. Sure, it's a good feeling to think you are getting the
fastest disk around. But every few months, a new model comes out that is faster than
anything that preceded it. If you really want to always have the best hard disk, you have
to keep buying more hardware, which is an expensive proposition that few opt for. See here for more on this subject.
 Next: Objective Performance Measurement (Benchmarking)
Home - Search
- Topics - Up
| |
