Submitted by Mark D. Hill
http://www.cs.wisc.edu/~markhill/papers/isat2012_ACSWTP.pdf
Advancing Computer Systems without Technology Progress
Mark D. Hill and Christos Kozyrakis
ISAT Outbrief, April 17-18, 2012, of DARPA/ISAT Workshop, March 26-27, 2012

This outbrief–now released for public distribution–summarizes finding
from 48 researchers who gathered in Chicago earlier this year to discuss
the following:

For decades, computer systems designers have built better systems relying
in large part on dramatically better technology. However, scaling CMOS in
a power and cost effective manner is now difficult, while post-CMOS
technologies are not yet mature. Thus, for the foreseeable future,
we must either accept that computers are good enough or advance them
without (significant) technology progress. Since computer system
superiority has been central to U.S. security, government, education,
and commerce, this workshop seeks to catalyze the latter.

Hypothesis: CMOS transistors will soon stop getting “better,” especially
w.r.t. power, and it is out of scope to discuss post-CMOS technologies.

Charge: Given the hypothesis, how do we continue to make computers systems
“better”?