DARPA, SRC pony up $194 million to fund chip research
- — 17 January, 2013 20:34
The U.S. Defense Advanced Research Projects Agency and a consortium of top semiconductor companies are handing out US$194 million to universities for research that addresses the physical limitations of semiconductors and chips.
The funding is part of the Starnet program, which will support research conducted primarily at six universities -- the University of Illinois at Urbana-Champaign, University of Michigan, University of Minnesota, Notre Dame, University of California at Los Angeles and University of California at Berkeley -- over a five-year period, according to Semiconductor Research Corporation (SRC), a research consortium focused on university chip research. SRC is backed by companies such as IBM, Intel, Micron, Globalfoundries and Texas Instruments.
The research will focus on transistors, nanomaterials, quantum computing, scalable memory and circuits. A goal is for the industry to be ready to move into a new era of computing with smaller circuits that are energy efficient and practical to manufacture. Another goal is to create scalable computing architectures with new forms of chips, memory and interconnects.
The research is also intended to protect U.S. security interests, while making the country a leader in semiconductors, DARPA and SRC said in a statement. DARPA is a division of the U.S. Department of Defense, and has funded key technology research in the past.
As devices become smaller, chips are being scaled down in size while also becoming faster and more power efficient. Every two years, Intel reduces the size of its chips, and currently makes chips using the 22-nanometer process. But chips are approaching nanoscale, which could create challenges related to their manufacturing and safety. IBM, Intel and universities like the Massachusetts Institute of Technology are already conducting research to address those challenges.
As part of the Starnet program, the universities will have centers addressing different subject matters. The research covers a range of topics including interconnects, memory, processors and related topics including scalability and energy efficiency.
The University of Michigan will focus on circuit fabrics for 3D interconnects and memory. The University of Minnesota will take on spintronics, which is considered by IBM as the basis for cheaper memory and storage in the future. UCLA will focus on atomic scale materials for next-generation chips, Notre Dame will tackle integrated circuits for low-power devices, and the University of Illinois will focus on nanoscale fabrics. Berkeley will focus on technology that could be the backbone for distributed computing across smart cities.
Overall, 400 university students and 145 professors in 39 universities will contribute to the research as part of the Starnet program.