Tenchusatsu,
But ever wonder why I use the line "It all depends on what your definition of the word 'is' is" so often? It's because if you can redefine the definition of any word to your advantage, you can get away with anything because there will be no standard by which to judge you.
I agree that it would better if there was an objective standard for measuring performance (agreed to by the industry). But there is nothing comprehensive out there that would give you an overall number, which would reflect performance. You may be completely right in that AMD's initiative is not ideal. But it's better than nothing. It's better than not measuring performance at all, and measuring clock speed instead.
As long as clockspeeds served as a good proxy for measure of performance of mainstream PC processors, there was no urgent need to come up with real performance measurement. But as the architectures of Intel processors moved in one direction (sharply lower IPC) and AMD moved in the opposite direction (slightly higher IPC), the need for the performance measurement became urgent.
I would have preferred an industry approach to this, rather than unilateral move by AMD, but in absence of an industry standard, AMD had no choice but to move unilaterally.
If Intel were to release a new Pentium 4 tomorrow which competed against the Athlon XP clock-for-clock (and I'm talking about real clocks, not "quanticlocks"), do you honestly think Jerry "We'll have a model 10,000 by the time Intel gets to 10 GHz" Sanders would give up the quantispeed nonsense?
If (when) Intel regained performance lead, do you think Intel will keep it a secret, and the world would depend on Jerry Sanders to announce it, or reflect it with quantispeed (or whatever)?
AMD marketing has hard enough time to get the truth out. The last thing you need to worry about is AMD spreading a falsehood, and Intel (industry, consumers, you name it) being somehow victimized by it.
Joe |
