Intel Corp (INTC)

[wbmw]

Mas, Re: You only have to add the TDPs of 2 * 3 Ghz Prescott Celerons and you get 146W.

Let me ask you a question: did you misunderstand my earlier comment about chips having different leakage levels, or did you choose to ignore it? If understood and acknowledged it, you should not still be claiming that a dual core die made of adjacent chips would have 2x the power of what is listed in the spec as single core TDP.

Re: AMD already have 2 Ghz parts at 35W.

AMD reached those power levels by picking the lowest leakage parts from their manufacturing lines and lowering the voltage and frequency. Intel can do the same thing, but they haven't yet, since Pentium-M covers those markets. Remember that Northwood scaled to 35W at up to 2.6GHz, and Prescott is only 10-15% higher power at similar speeds (Northwood TDP 3.0-3.4GHz is 82-89W, Prescott 3.0-3.4GHz is 89-103W).

If Intel takes the lowest leakage parts from their assembly line, lowers the voltage, and keeps the frequency 3-5 bins from the top (AMD intends to do the same with their dual core), then the individual cores should dissipate <60W each and have frequencies >= 3GHz.

Re: Semiconductor and thermal physics may defeat you here

It might be best for you to familiarize yourself with these before using them in your argument.

I'll be sure to remind you next year about your prediction:

"it's the dual-core K8 clocking more than dual-core P4 that I am sure of"