Page 1 of 1

Chip Saves Power by Fudging the Numbers

Posted: Fri Dec 20, 2013 5:16 am
by MSI
MIT Technology Review: Chip Saves Power by Fudging the Figures
  • Apparently in proving that algorithms control our Probably, Approximately Correct! lives, computer scientists are now looking at fudging the numbers to elicit more speed and power from smaller chips.
    • From the article:
      • Apparently scaling back the precision used to express certain values allow some circuit elements to remain idle.
      • Approximations also dials down the voltage required for some circuit elements to work on data
      • (hopefully?) The degree of noise acceptable for different parts of a program may make it possible to use approximation without overloading it with errors
Guess when you use your 'smart' phone to help you with a decision you best realize that it may not be a perfectly correct decision!
  • (NOTE: I take issue with calling them 'smart' phones! Aren't they really 'dumb' phones since our reliance on them decreases our ability to THINK and makes us 'dumber'??!
    Socrates said in Plato's Phaedrus that youth can be deluded by the seeming permanence of print to believe that they have found truth when they have only barely begun to probe for it.
    I'd add that today's youth are being deluded by their seeming infinite intelligence brought by their 'smart' phones which allow them to 'look anything up' rather than probe their minds to find or recall the correct answer!)
    These new chips indicate that the 'smart' phone is going to be further 'dumbing' us down!
but i digress...
Chip Saves Power by Fudging the Figures
Chips that save energy by approximating some calculations could allow mobile devices to be smarter at understanding the world.
By Tom Simonite on December 19, 2013
Which begins:
  • We owe our smartphones and supercomputers to the mathematicians and engineers who figured out in the 1940s and 1950s how to create machines that can crunch numbers at high speed with perfect accuracy. Some researchers are now going back on that principle by working on designs that sacrifice accuracy for power efficiency. The approach, known as approximate computing, could extend the battery life of mobile devices and enable advanced techniques such as computer vision.
See the full article:Chip Saves Power by Fudging the Figures