Cracking the energy saving code
As the world becomes ever more wired and the devices on which we rely become ever more complex and powerful, attention is focusing on how to reduce their energy use.
While hardware has been the common focus of these efforts, scientists at University of Washington have created a computer language; EnerJ, that reduces energy consumption in simulations by up to half. In time the team hope the energy savings can be as high as 90%.
Study author Luis Ceze, a UW assistant professor of computer science and engineering said: "With our system, mobile phone users would notice either a smaller phone, or a longer battery life, or both. Computing centres would notice a lower energy bill."
The new system relies on mistakes. Some processes can survive small errors and by introducing more the UW team say they can save energy.
The UW system separates the computer code into two: one that relies on precision, the other which can survive small errors.
Reducing emissions and saving money is one benefit of this approach, and it may also solve the so-called "dark silicon problem"; the limit on the number of transistors that can run on one chip.
"When I started thinking about this, it became more and more obvious that this could be applied, at least a little bit, to almost everything," said Adrian Sampson said. "It seemed like I was always finding new places where it could be applied, at least in a limited way."
The team are now working on hardware, which they believe will boost the energy savings from the 50% attributable to the Enerj software alone to 90%.
"Our long-term goal would be 10 times improvement in battery life," Ceze said. "I don't think it is totally out of the question to have an order of magnitude reduction if we continue squeezing unnecessary accuracy."
The team hopes to release the code as an open-source tool this summer.
Top Image Credit: © Andres Rodriguez