When Intel during IDF last week demo-ed a 3.5GHz Pentium 4, AMD couldn’t stay too far behind and they released some details about their upcoming CPUs to the press. It is clear that AMD has no more cards up their sleeve to win the clockspeed battle, as Intel will be ramping up clockspeed pretty quickly. But as most of us know, clockspeed doesn’t equal performance, and in the case of AMD and Intel, a 1.4GHz Athlon more-or-less performs equal to a 2GHz Pentium 4. AMD has addressed this issue in the past with the PR-rating of their CPUs, which indicated what clockspeed Intel CPU their CPU is equivalent to.
In the light of AMD’s success in the enthusiast and home-user markets we’d actually advice AMD to not start using that PR rating again. The majority of these enthusiasts and home-users are well aware that clockspeed isn’t the only thing that determines a PC’s performance. However we can see the reasons for opting for such a rating when people buy off-the-shelf PCs at, for example, Best Buy
or Fry’s. I’m doubtful the sales guy will be able to properly explain why the AMD system of a lesser clockspeed can be just as fast or faster than the Intel.
What has AMD got planned then? They’ll actually be using a ‘model’ designation for their CPUs, for example their 1.4GHz part will be named Model 1600 to reflect that its performance is on par, or better than a 1.6GHz Pentium 4. But there’s more, by the looks of it they won’t be referring to any CPU in clockspeed anymore, just the model number. This will mean that even your motherboard will post with a ‘Model 1600 AMD Athlon’ instead of ‘AMD Athlon CPU at 1400 MHz’, confusing and actually a bad idea in my opinion.
By doing so AMD acknowledges that Intel actually controls the market and determines how performance is measured. I’d respect AMD for not sticking with that, but actually doing their own thing. It is not a shame that they don’t have a 2GHz part, on the contrary it is admirable that their 1.4GHz performs on par with the 2GHz Intel part, despite the 600MHz clockspeed advantage, a 400MHz FSB and RDRAM memory. If I were AMD I’d focus my marketing efforts on the fact that their CPUs are faster clock-for-clock, or to make a bolder statement, offer more performance per MHz.
I’m confident that once AMD gets their 0.13-micron problems sorted out and can actually start shipping faster Athlons, of the Thunderbird or Palomino variety, they have what it takes to take on the new 0.13-micron Pentium 4s, even if their CPUs run at a lower clockspeed. Therefore they’ll be much better off marketing their performance edge at a lower clockspeed, than to use a model designation for their CPUs, as they have nothing to be ashamed off. Their CPUs are faster clock-for-clock, not even a 600MHz clock difference enabled Intel to put the ‘Fastest x86 CPU’ sticker back on.
AMD should not try to play the clockspeed game, but actually rise above it and educate its prospective customers that clockspeed simply doesn’t equal performance. If they do a proper job they will have debunked one the biggest myths in the industry and have taken a number of cards from the hands of Intel’s PR and will force them to play a different game of numbers, performance rather than clockspeed.