Mar 24 2008
Sun Microsystems, Inc. today announced that the Defense Advanced Research Projects Agency (DARPA) has awarded Sun $44.29 million funding for a five and a half-year research project focused on microchip interconnectivity via on-chip optical networks enabled by Silicon photonics and proximity communication. Part of DARPA's Ultraperformance Nanophotonic Intrachip Communication program, the project commences with an incremental delivery of $8.1 million to Sun Microsystems' Microelectronics and Laboratories divisions. For more information on research projects at Sun, visit https://www.oracle.com/index.html.
Building on research done under DARPA's High Productivity Computing Systems program, Sun's new project will accelerate the development of lower cost, high performance and high productivity systems. The project presents a unique opportunity to develop supercomputers through interconnecting an array of low-cost chips, with the potential to overcome the fundamental cost and performance limits of scaling up today's large computer systems. By providing unprecedented high bandwidth, low latency, and low power interconnections between the parallel computing chips in such an array, this research project will help enable a broad class of companies and organizations to utilize applications with high compute and communication requirements, such as energy exploration, biotechnology and weather modeling.
“Optical communications could be a truly game-changing technology — an elegant way to continue impressive performance gains while completely changing the economics of large-scale silicon production," said Greg Papadopoulos, chief technology officer and executive vice president of research and development for Sun. “Congratulations to Sun Labs and Microelectronics teams for their constructive creativity and for driving innovation into the semiconductor marketplace."
Sun's program combines optical signaling with Proximity Communication, its key chip-to-chip I/O technology, to construct arrays of low-cost chips in a single virtual “macrochip.” Such an aggregation of inexpensive chips looks and performs like a single chip of enormous size, thus extending Moore's Law; it also avoids soldered chip connections to enable lower total system cost. Long connections across the macrochip leverage the low latency, high bandwidth, and low power of silicon optics, and through this program Sun and DARPA will research technologies to dramatically further reduce the cost of these optical connections. The result is a virtual supercomputer.
“DARPA’s UNIC (Ultraperformance Nanophotonic Intrachip Communications) program will demonstrate high performance photonic technology for high bandwidth, on-chip, photonic communications networks for advanced (™ 10 trillion operations/second) microprocessors. By restoring the balance between computation and communications, the program will significantly enhance DoD’s capabilities for applications such as Image Processing, Autonomous Operations, Synthetic Aperture Radar, as well as supercomputing,” said Dr. Jag Shah, program manager in DARPA's Microsystems Technology Office.
Accelerating Innovation to Extend Moore's Law
The historic accuracy of Moore's Law, which predicts a periodic doubling of the number of transistors that can cost-effectively build on a single chip, is partly behind the impressive growth of microprocessor performance over the last 30 years. Today, though, continued improvements are slowing down, as power and size constraints limit the growth of chip clock frequencies. Boosting computer performance by accumulating hundreds or thousands of cores per chip allows users to exploit massively parallel execution, but it also requires large increases in the number of transistors on a chip, and hence an unconstrained continuation of Moore's Law. However, as Dr. Gordon Moore himself predicted long ago, economic limits on the global financial investment in semiconductors are now slowing down Moore's Law.