ing. It is not unusual, for instance, for companies to replace their older computers with new, more energy-efficient ones in an effort to become more earth-friendly.

This practice might not always be the most environmental solution, says Teradata’s Wang. “What I propose is that we look at the entire life cycle of a computer, the whole picture, from manufacturing through day-to-day operation,” says Wang. “Every step consumes energy, and buying a new, more efficient computer may not always be the answer.”

Some computer manufacturers are retooling their products from a life-cycle point of view and making the decision to buy a new, energy-efficient computer much easier. Dell is accelerating its programs to reduce hazardous substances in its computers, and its new OptiPlex desktops are 50% more energy-efficient than similar systems manufactured in 2005, thanks to more energy-efficient processors, new power management features, and other factors.

Likewise, Hewlett-Packard recently unveiled what it calls “the greenest computer ever”—the rp5700 desktop PC. The rp5700 exceeds U.S. Energy Star 4.0 standards, has an expected life of at least five years, and 90% of its materials are recyclable. The computer is easy to disassemble and meets the European Union’s RoHS standards for the restriction of the use of certain hazardous substances in electrical and electronic equipment. Moreover, 25% of the rp5700’s packaging materials are made of recycled material.

for the future of the Planet

In an effort to ensure “computing can have a positive effect on our lives and the world,” Hopper and Andrew Rice, an assistant director of research at the University of Cambridge’s Computer Laboratory, have identified four principal goals in their paper “Computing for the Future of the Planet.” The first goal is an optimal digital infrastructure in which computing’s overall energy consumption is reduced and the efficient use of energy in the manufacture, operation, and disposal of computing devices is maximized.

The second goal is “to sense and optimize the world around us with reference to a global world model,” which would “inform us about the energy con-

sumption and other effects of our activities on the natural environment.”

The third goal is a new emphasis on predicting and responding to future events by modeling their behavior. According to Hopper and Rice, “The traditional role of computing as an execution platform for these models will continue to be important and must grow in performance to service both the increasing demands of high-er-fidelity models and also to accommodate any new overheads incurred by correctness checking.”

Lastly, Hopper and Rice are “ interested in the possible benefit of digital alternatives to our physical activities,” such as electronic versions of printed newspapers, music downloads rather than physical CDs, and online shopping as opposed to visiting stores and supermarkets. According to Hopper and Rice, “One might argue that a total shift from physical to digital seems unlikely in today’s world but for future generations this concept might seem as obvious as email is to us today.”

“People in the developing world,” Hopper and Rice note, “often live in resource-impoverished environments so a physical-to-digital paradigm shift has the potential to enable activities that were hitherto prohibitively expensive, and to support development whilst minimizing its impact. We seek to unlock methods of wealth creation in the virtual world.”

Hopper and Rice also suggest the development of a personal energy meter that would measure a person’s direct and indirect daily consumption, with individualized breakdowns of “the energy costs of travel, heating, water-usage and transportation of food [that] will help us target areas for reduction in our environmental footprint…. The data collected will not only provide useful information for analyzing consumption patterns but also has the potential to help individuals identify alternatives to their current activities.”

“I think we’ve only just started to address the issue” of green computing, says Hopper. “It’s just on the cusp of becoming important, and I think business, not academia, has led the way. They are driven by pragmatic concerns.”

 

Patrick Kurp is a freelance science writer in Bellevue, WA.

Artificial Intelligence
Super-
computer
Defeats
Human
Go Pro

The new Dutch supercomputer Huygens, armed with the MoGo Titan program, defeated a human professional Go player with a 9-stones handicap.

The victory appears to be the first-ever defeat of a high-level human Go player by a supercomputer in an official match.

Until recently, scientists were unable to create a computer program capable of beating even many amateur-level Go players. This state of affairs changed in 2006 when programmers Sylvain Gelly and Yizao Wang devised a revolutionary algorithm that has enabled the MoGo Titan program to attain new heights; since August 2006, MoGo Titan has been ranked number one on the 9x9 Computer Go Server.

Teamed up with the Huygens supercomputer, MoGo Titan achieved a noteworthy victory as its opponent, Kim Myungwan, is an 8 dan pro (the highest level is 9 dan) and a seasoned international competitor. In fact, the day before Myungwan’s official match with Huygens and MoGo Titan, he soundly defeated the duo in three blitz games played with varying handicaps.

“The current result forecasts that before 2020 a computer program will defeat the best human Go player on a 19x19 Go board in a regular match under normal tournament conditions,” says professor Jaap van den Herik of Maastricht University which, with INRIA France, co-developed MoGo Titan. “This is remarkable, since around 2000 it was generally believed that the game of Go was safe to any attack by a computer program. The 9-stones handicap victory casts severe doubts on this belief.”

The Korean-born Myungwan appears to have taken the defeat well. Two days after his loss to MoGo Titan, he won the 2008 U.S. Open.