Krogoth wrote:Power consumption is driven both by "clockspeed" and active transistor count.
I said the power consumption
"problem", in which I was alluding to the same phenomenon that JBI's article did: in the early 2000s gate leakage suddenly become a huge concern in ways it never had been before. You might even remember some of this; I think you're probably old enough to remember Prescott and Intel's 90nm process. It was around that time that Dennard's Scaling broke, not Moore's Law.
That's what changed: power consumption was no longer driven in the same way by those factors.
That's what I was saying. The fact that more transistors consume more power is obvious, entirely non-controversial and clearly
not in contention.
krogoth wrote:Adding more cores/ICs into the design will still run into the problem of power consumption and thermal output if they are all being utilized.
This, again, is obvious. You've added nothing to the discussion by adding it. Absent violating several bedrock laws of physics, more switching transistors = more power used. That's just as true in 1965 as it is today.
krogoth wrote:The current power saving schemes on modern chips involved reducing voltage, clockspeed and putting idles parts of the silicon into a "sleep state".
As opposed to the
non-current power-saving schemes?
Those techniques are't ultra-modern, it's just that we need to do them today in ways we didn't have to do them 15 years ago. Why? Not because Moore's law broke but because Dennard's scale did.
i.e., "the power consumption problem is primarily related to how fast the transistors are driven, not really how many transistors there actually are."
Krogoth wrote:Moore's observation has been invalided by scale of economics on the production end.
Demonstrate how.
Krogoth wrote:It is becoming more and more expensive to move to the next process node while the returns are diminishing.
Perhaps, and I don't dispute that perhaps Moore's observation won't hold true very soon, but you said it stopping holding true 10 years ago.
I ask you again, please demonstrate how.
Because that's not what the IEEE article states, or is even really about, and because you've shown that you don't properly understand Moore's law in the first place.