With so much attention recently over the past few years being directed at the energy efficiency of electronic devices, I think it’s worth reflecting just how significant energy efficiency improvements are in conjunction with Moore’s Law. For those who are unaware, Moore’s Law specifically refers to the observed trend that the number of transistors that can be inexpensively placed on an integrated circuit doubles every every 2 years. In essence, Moore’s Law predicts the exponential increase in computational power over time.
More generally, however, Moore’s Law is used to describe the dual trend for electronics: they become ever faster while simultaneously becoming ever cheaper. A cheap cell phone today contains the same processing power of a supercomputer a decade ago. With the rise of cheap microprocessors for devices like laptop computers and especially cell phones, energy efficiency has become a core design template.
What makes this whole intersection of Moore’s Law and ever-increasing energy efficiency particularly interesting is the sheer computational ability that is becoming available at a significantly cheap energy price. Decoding High-Definition video can be done almost effortlessly on cell phones like the iPhone. The amazingly energy efficient Intel Atom processor can provide a comparable Desktop or Laptop computing experience for a fraction of the energy cost of a standard processor. What once consumed several hundred Watts of electricity can now be done for several dozen Watts. And more powerfully, too.
Perhaps I am overstating the effect in recent years. While improvements have certainly been made, the progress is better described as incremental rather than miraculous. But herein lies the importance of the slow-yet-inexorable march of Moore’s Law. So-called “bang-for-your-energy-buck” computing is improving at a substantial rate, and will continue to do so for the foreseeable future.
And not a moment too soon. Current energy consumption patterns in the U.S. are based largely on the availability of cheap energy sources. As those begin to dry up (at least until viable alternatives can be found/invented) the importance of conserving available energy will become very important. So while the future looks bleak in terms of the kind of cheap energy we enjoy today, we can probably anticipate a future where power consumption of electronics will evolve to match the shrinking supply of energy.
One final note of interest is just the sheer volume of computational power that is beginning to emerge in the larger technosphere – much of it largely unused or underutilized. Will the Folding@Home project eventually migrate from a small cadre of computer enthusiasts and Playstation 3 owners to anyone with a cell phone? Could weather prediction and climate modeling one day gather data from cell phones, laptops, and integrated sensors distributed worldwide? And with the potent combination of increasing energy efficiency and Moore’s Law, what today requires billions of research dollars and massive, power-hungry supercomputers might some day in the future be computed on billions of idle cell phones – all for a fraction of the infrastructure cost and energy consumption. That’s certainly a future to be excited about . . .