Nathaniel Bullard <\/p>\n
Fifteen years ago, Japans Earth Simulatorwas the most powerful supercomputer on Earth. It had more than 5,000 processors. It consumed 6,400 kilowatts of electricity. It cost nearly US$400 million to build. <\/p>\n
Two weeks ago, a computer engineer built a deep learning box, using off-the-shelf processors and components, that handily exceeds the Earth Simulators capabilities. It uses a maximum of 1 kilowatt of power. It cost US$3,122 to build. <\/p>\n
For the first time in writing this, Im stumped for a chart. It is difficult perhaps impossible to show a 99.98 per cent reduction in energy use and a 99.99992 per cent reduction in cost in any meaningful way. It is enough to say that information technology has decreased in cost and increased in computational and energy efficiency to striking degrees. <\/p>\n
I would argue that this dramatic improvement has a flattening, or even depressing, economic influence on energy. Dramatically reduced inputs with dramatically increasing outputs is a boon for consumers and businesses, unless those businesses sell the energy that drives those inputs. Weve already seen this: In 2007, U.S. data centres consumed 67 terawatt-hours of electricity. Today, with millions of times more computing power, they consume 72 terawatt-hours, with less than 1per cent growth forecast by 2020. Not the greatest news if youre a power utility that has imagined that more and more information technology will mean more energy demand. <\/p>\n
Information technologys improvement over time has been largely a function of Moores Law (which is less a law than an observation). Now, with Moores Law potentially coming to its end, it would seem like the extraordinary improvements that got us from a room-sized US$400 million supercomputer to a US$3,000 desktop box in 15 years could be coming to an end, too. <\/p>\n
If technology companies are no longer able to jam more transistors into a chip, does that mean that improvements in energy consumption will also come to an end? If chip improvements plateau, and deployment increases, can information technology find a way to provide a boost to energy demand? <\/p>\n
I doubt it, for both hardware and software reasons. <\/p>\n
Even as Moores Law is tapping out for general-purpose chips, hardware is becoming increasingly optimized for specific tasks. That optimization for such things as graphics processing or neural network computations for machine learning leads to greater energy efficiency, too. Google now has its own application-specific integrated circuit called the Tensor Processing Unit for machine learning. The TPU delivered 15-30x higher performance and 30-80x higher performance-per-watt than central processing units and graphics processing units. <\/p>\n
Then there is the software that runs on that custom hardware, which has direct applications for electricity in particular. Last year, Google unleashed its DeepMind machine learning on its own data centres and managed to reduce the energy used for cooling those data centres by 40 per cent. <\/p>\n
So, new special-purpose chips are much more energy-efficient than older general-purpose chips and those efficient chips are now used to run algorithms that make their data centres much more energy-efficient, too. <\/p>\n
In a famous 1987 paper, the economist Robert Solow said you can see the computer age everywhere but in the productivity statistics. Today, we could say the same about the computer age and energy statistics. <\/p>\n
Nathaniel Bullard is an energy analyst, covering technology and business model innovation and system-wide resource transitions. <\/p>\n
Bloomberg View <\/p>\n
<\/p>\n
Go here to see the original:<\/p>\n
Tech improvements are becoming so dramatic that charts are ... - Financial Post<\/a><\/p>\n","protected":false},"excerpt":{"rendered":" Nathaniel Bullard Fifteen years ago, Japans Earth Simulatorwas the most powerful supercomputer on Earth. It had more than 5,000 processors. Continue reading