Many accurate enough chips will use software for
correction to achieve up to 10,000 times higher
speed and lower power usage:
"Interested in solving tasks that benefit from floating point (“fp”),
but IEEE floating point unit takes over 500,000 transistors
• Could less accurate fp arith unit (eg, 1% error) be very small?
• Yes: at least 100x smaller - O(5000) transistors - will sketch
• If errors can be compensated in application software, can get
10,000x better speed, power than CPU (100x GPU)
DARPA funded the creation of Joseph Bates Singular Computing LLC’s
chip because fuzziness can be an asset when it comes to some of the
hardest problems for computers, such as making sense of video or
other messy real-world data. “Just because the hardware is sucky
doesn’t mean the software’s result has to be,” says Bates."
See:
http://nextbigfuture.com/2016/04/man...-will-use.html