Low Voltage Is Key To Energy-Efficient Chip

An anonymous reader writes in with news from the International Solid State Circuits Conference in San Francisco of a new energy-efficient chip designed by researchers at MIT. It's said to be able to run on 1/10 the power of current chips. Texas Instruments worked with MIT on the design, which is maybe five years from production. "The key to the chip's improved energy efficiency lies in making it work at a reduced voltage level, according to... a member of the chip design project team. Most of the mobile processors today operate at about 1 volt. The requirement for MIT's new design, however, drops to 0.3 volts."

Comments

noshitposter

[Anonymous Coward]

I see someone tagged this "noshitsherlock". But this is a hard thing to do because the difference between "0" and "0.3" is smaller than "5" lowering the immunity to upsets like noise.

Architecture is far more important

[EmbeddedJanitor]

Less transistors switching per unit of work done means better power performance.

That's why your cell phone has an ARM CPU and not an x86.

You need to pervceive the right things...

[EmbeddedJanitor]

Sure some CISCs have a RISC under the hood, but that just means you need to have a "virtual machine" that emulates a CISC on top of the RISC. Those extra layers mean more internal operations which mean more switching.

Re:Physics

[Chris Burke]

The electrical characteristics of a CPU are somewhat more complicated than those of a resistor. True, but in fact a chip's power does scale with the square of the voltage. At a gross level you can approximate the chip as a certain constant resistance for static power, aka leakage, and as an RC circuit with a given constant for dynamic power, which scales linearly with frequency as well. Nobody actually does that, they just measure the power consumption and know that they the number is proportional to voltage squared and frequency.

Of course I just knew some jackass was going to use this fact to try to downplay the achievement. Okay, yeah, every computer engineer knows that to reduce power by four you drop the voltage by half, but the trick is actually making this work. That's why not every chip runs on 1E-20 Volts, Mr. Anonymous Idiot.

Re:All well and good

[unitron]

The amazing thing is that they were able to get the transistors to bias at that voltage. That was my first reflex thought (due to what I learned when), but I suspect that we're talking about Field-Effect Transistors where an electrostatic field affects the resistance of a unipolar channel and not Bipolar Junction Transistors where you need twice that much electrical pressure to get the base-emitter junction to conduct.

Re:Process Counts

[randyest]

The core voltage and the I/O voltage (which is where SSO is a concern) need not be the same, and rarely are in advanced processes. I'm sure the I/O's are not 0.3V. The rest of your comment was similarly confusing: using gate oxides aren't a "trick" (they're pretty much a requirement,) 65nm and under are more than "slightly" better then 90nm "in some ways," and I don't know what curve you're talking about.