'Inexact' Chips Save Power By Fudging the Math 325
Posted
by
Soulskill
from the for-sufficiently-large-values-of-1 dept.
from the for-sufficiently-large-values-of-1 dept.
Barence writes "Computer scientists have unveiled a computer chip that turns traditional thinking about mathematical accuracy on its head by fudging calculations. The concept works by allowing processing components — such as hardware for adding and multiplying numbers — to make a few mistakes, which means they are not working as hard, and so use less power and get through tasks more quickly. The Rice University researchers say prototypes are 15 times more efficient and could be used in some applications without having a negative effect."
Re:Graphics cards (Score:4, Insightful)
Big difference between not dealing with full precision and encouraging erroneous behavior by trimming infrequently chunks of hardware.
Seems like nothing new (Score:4, Insightful)
Re:Turtles all the way down (Score:5, Insightful)
Furthermore, posting under the top post when your reply is nothing to do with the OP is considered a faux pas. Minus 50 DKP.
Re:AI Chip (Score:5, Insightful)
Re:Turtles all the way down (Score:2, Insightful)
Re:AI Chip (Score:1, Insightful)
Did you really want a long-winded answer about how the conditioning of the wetwired neural network in your skull responding to external visual stimuli works? It's not math, it's chemical potentials at the neuron level. At a higher level, it's an estimate of depth perception vs. how fast you think you can walk (or run) across that street. No math involved, just spatial guesstimates of that brain of yours.
Re:Turtles all the way down (Score:4, Insightful)
What about in a RT rendering (game/BD-Rom decode) situation, or a RT communication (Skype) situation?
Both of these do not need exact values, just close enough, and even if there was an error it will be transient and gone almost as fast as it was noticed?
-nB
Re:Whatcouldpossiblygowrong (Score:5, Insightful)
If I'm reading the article right, the chips are still deterministic, they just don't care about a few rare edge cases. So whether there is an error or not depends on the input, and in your case all four chips will make the same mistake. What you could try is modify the input a little for each rerun and try to interpolate the result from that, but that won't give you perfect accuracy.
Re:AI Chip (Score:5, Insightful)
Which is... math.
Just because something doesn't involve digits doesn't mean it's not math. I suggest you look up analogue computers, because that's what you just described - a neural net acting as an analogue computer.
--
BMO
how is that not math? (Score:3, Insightful)
It's still math, it's just in the hardware rather than the software.
Re:Prediction (Score:3, Insightful)
While everyone sits here making jokes all i can think of is...why? Dear sweet lord why would you want this? are you telling me those chips in the 50c calcs are so damned expensive you couldn't use one?
Everyone seems to be missing the most obvious answer which is thus: If your general purpose CPU sucks too much power doing math then DON'T USE THE CPU TO DO MATH and instead have a math processor...duh! I have NO doubt you could build a simple ARM chip that sucks almost no power and does all the basic math functions, hell it would probably do all that your average graphing calc could do and again the ARM arch is a power sipper so no problems there.
I just don't why we have to keep reionventing the wheel. back in the day the CPU sucked for certain functions so you had an ALU to do that job, so if your CPU still sucks too much then leave the CPU for other tasks and use an ultra low power ALU for the math. Isn't it funny how these things just seem to go round and round?