albert, Re: "no matter how imperfect QS may be one always knows that 1800+ is slower than 2000+. You can not say say the same for P4's GHz. Hence there is a benefit in QS"
Not true. For the Pentium 4, 2.2GHz is always faster than 1.9GHz, which is always faster than 1.6GHz, which is always faster than 1.5GHz. But, you are talking about megahertz outside of the Pentium 4 micro-architecture, such as whether 1.5GHz for Pentium 4 is faster than 1.4GHz from Pentium III-S. It is not, but most informed buyers should know not to compare megahertz outside of a given micro-architecture. I knew that 10 years ago when I was comparing the 386 and 486. As for the uninformed buyer, that's the one that AMD is targeting with QuantiSpeed, but as I see it, QS does no better than megahertz in describing performance.
That's because model numbers and megahertz have similar faults, and I brought this problem up earlier. The example I used was when AMD saw fit to name a 1.3GHz Athlon 4 into a 1500+, even though they already used a 1500+ to name their 1.33GHz Athlon XP with a faster front side bus. You say that someone will always know that 1800+ is always slower than 2000+, but will they also realize that 1500+ is always slower than 1500+? See how it gets confusing? Now throw in the fact that a 1.2GHz Athlon 4 is still called 1.2GHz, and most people should agree that QS is no more beneficial in megahertz in describing performance.
wbmw |
