The DoD's Defense Advanced Research Projects Agency (DARPA) will contribute $2m (£1.3m) toward IBM's Low Power Center, a ten-month-old initiative based at IBM's Texas research lab.
IBM's agreement with DARPA calls for it to develop technology for increasing the reliability and energy efficiency of high-end computing systems. Defence contractor BAE Systems plans to create prototypes in 2003 for testing some military applications of the research.
IBM also intends to use the power-conserving products internally and to develop design-analysis tools for estimating and analysing the power consumption and performance of systems utilising PowerPC processors.
As computing technology has advanced to allow systems of ever-increasing processing capacity, the power required to run and cool those systems has also ballooned.
A 1994 Department of Energy study found that 10% of the energy consumed in North America goes to supporting IT systems, according to IBM. Such power demands also eat up a significant portion of corporate datacentre budgets, the company said.
By studying low-power computing, with regard to hardware and software, IBM hopes to develop more energy-efficient products.
Research from the Low Power Center has already been used in designing IBM's 405LP PowerPC chip, planned for production later this year and intended for use in devices such as handhelds, and in a prototype of its Super Dense Server, an ongoing research project aimed at improving high-end enterprise server technology.
The new IBM initiative falls within DARPA's three-year-old Power Aware Computing/Communication (PACC) project, which aims to develop technology that minimises the power needed to complete a given task.
Power limitations constrain undertakings such as missile and satellite missions, according to DARPA. PACC's goal is to reduce that obstacle, facilitating cost reductions and new capabilities.
DARPA, best known as the agency that funded early Internet development, has worked with IBM on a variety of projects in the past, including research on parallel computing, security, cognitive computer systems and advanced computer architecture.