This quote is taken from Windows Secrets.
"But note: While benchmarks can point you in the right direction, they're not gospel. In fact, small performance differences really don't matter much at all because of human physiology. Most people — the mythical "average users" — are relatively insensitive to speed differences of less than about 15–20%.
I know that sounds hard to believe, but it's true. It's something I've seen time and again in the various PC testing labs with which I've been involved (Byte Magazine, Windows Magazine, NTSL, etc.).
People who live at their PCs all day, every day, are somewhat more sensitive than casual users and can often detect speed differences in the 10–15% range. PC testing professionals are the most speed-sensitive of all, and they often can detect differences in the 5–10% range. Lower than that, and almost everyone needs a stopwatch or other instrumentation to reliably detect PC speed differences. In other words, small speed differences fall below the threshold of human perception and, thus, don't matter a lot.
Plus, in most normal uses, PCs spend the majority of their time waiting for something to happen — a keystroke or mouse movement, a bit of data to arrive, or whatever. In between those actions, the PC is sitting there, mostly just twiddling its figurative thumbs. Whether it twiddles slightly faster or slower really doesn't matter much, and has no perceptible effect on the user experience.
All of which adds up to this: It's not worth agonizing over a few percentage points, one way or the other, in performance tests. Unless you're doing something very unusual with your PC, just about any of today's midrange and higher PCs will probably serve you well. Thus, you can focus on other important factors, such as the other components in the system, price, service, the software included with the system, and so on. Small performance differences among and between CPU types simply aren't something that will change your life for better or worse!"
Hope this helps.
