Re: AMD currently has better strain than Intel. AMD has SOI.
You can't judge a process based on the features. If it were capable of such ultra-low power, why can't AMD get Turion down to the power levels of Pentium M without using a specialized boutique design with only enough binsplits at the top to supply a limited number of 3rd tier OEMs? It looks like the strained silicon and SOI process enhancements weren't enough to offset a generally mediocre process.
I know... Pentium D is a power hog; therefore, Intel's 90nm process must be a joke. You know, I've heard that one before, it it doesn't get any funnier when other people tell it. Exclude Pentium 4/Pentium D when making process comparisons, because the power of these parts is entirely due to a power hungry design. Using the same process, Intel has shown that they can make very low power, high performance designs at 90nm, and 65nm has demonstratable improvements(according to public data, at any rate). The fact that it doesn't need exotic features to get there is even further testament to its superiority, IMO.
Re: Also, though it's not clear what Intel uses exactly there's a reasonable chance that AMD has an edge over Intel too with APM.
Ridiculous. APM is nothing more than AMD copying Intel's Copy Exactly methodology, except that Intel has been doing it for far more years and has far more experience with it across far more fabs.
Re: Not that important but just maybe current advantages are part of the reason why AMD single core cpu's outperform Intel single cores don't you think
No, I don't think that's true. Pentium M does a good job of competing in performance vs. Turion, in spite of having much less processor to memory bandwidth, and even a simpler micro-architecture (in some ways).
Re: remember rather 'hard' frequency ceilings for Banias and Dothan that prohibited scaling much past introduction frequencies
They are constrained by their power envelope. Experimentally, Dothan can overclock past 2.7GHz. Not that we'd ever see one of these in a 27W power envelope, however. Once the K8 core goes over 2.4GHz, the power envelope increases rather steeply due to leakage. Note the power dissipation of the Athlon FX. It's higher than even the dual core Athlon X2. Now, that is something to indicate a leaky process.
Re: Historically you're wrong as well. Remember copper (.18) shortly after the start of the Athlon production?
I remember Intel's .18u process. It enabled them to have quite a reasonable ramp from their .25u parts, though competitively, they couldn't reach the same frequencies as Athlon due to it being an older micro-architecture without the same clocking potential.
Re: If you really have no data or knowledge to base your above conclusion upon why bother to post it at all?
But I do have the correct knowledge to base my conclusions on. You just need to know the difference between design benefits and process benefits. Intel has the highest yielding, best performing process out there. AMD may be very close with theirs at 90nm (I'll admit that their process is no dud, but I will argue that it's not superior), but their design (K8 vs. Netburst) is currently superior to delivering the highest performance at the lowest power envelope.
I'll also concede that in mobile, AMD's low power transistors and boutique manufacturing have enabled a competitive part, both in performance and power, but they would never be able to ramp such a part in volumes that would compete for much more than a niche. Turion was designed to look good on paper, but it's a low volume product line meant to serve 10% (15% at most) of the mobile market. That's not a process achievement, it's a design tradeoff.
AI
