You may find
this paper (PDF) interesting.
Thanks for posting that, even though it is completely over my head I still like to glance through a technical paper like that when I have the chance.
From the article:
said:
V. SUMMARY AND IMPLICATIONS
The 30-year-long trend in microelectronics has been to increase
both speed and density by scaling of device components
(e.g., CMOS switch). However, this trend will end as
we approach the energy barrier due to limits of heat removal
capacity. For nanoelectronics, this result implies that an increase
in device density will require a sacrifice, due to power
consideration, in operational speed, and vice versa. Thus, it
appears that we are entering a regime where tradeoffs are required
between speed and density, quite in contrast to the
traditional simultaneous benefits in speed and density from
conventional scaling.
Yep, consumers (like me) have been spoiled by both the simultaneous speed and density increase of the past node progressions.
The conclusion of this paper written in 2003 warns of the upcoming compromises that may bring that trend to an end. So far Intel has avoided this with 22nm FinFET. (offering a 50% reduction in die size along with a 50% reduction in power required essentially make the node heat density neutral.)
But where we will go from here? What will happen with our CPUs?
Will Intel be able to maintain or even increase single threaded performance at the 16nm or 10nm nodes with x86? If so, how will they do this? Maybe some "dark silicon" workarounds inside the CPU core to increase single threaded performance while reducing heat density? But if this happens, what will be the cost of the overhead? When does the bubble burst with respect to chasing the classic metrics with an old ISA?
I've really been wondering where this is headed? I've also been wondering how companies like Apple see this upcoming heat density situation?
Should be interesting times (over the next 2-8 years) in the CPU world that is for certain. I can't wait to read the future Anandtech CPU articles as this competition for the single threaded performance metric continues.