Cell architecture

摘要

On behalf of the organizing committee for the 2007 International Symposium on Physical Design (ISPD), we would like to extend a warm welcome. This is the 11th year of ISPD; from rather humble beginnings as the ACM/SIGDA Physical Design Workshops held intermittently during 1987-1996, the event has evolved into the most competitive forum for the research area. The research community for physical design is not big, but the problems addressed are large. Over the years, many groundbreaking ideas have appeared first at ISPD. The recent interest in benchmarking and open competitions also has roots in ISPD. >In earlier years, the perception outside our community was that physical design was simply the shuffling of rectangles to create masks. While the placement of rectangles is certainly a part of our work, modern physical design is much more complex and penetrating. Variability introduced by manufacturing, interconnect delay, and heat dissipation are all tied to the physical layout of a circuit. The papers included in these proceedings reveal many aspects of the physical design problem; as technology marches forwards, physical design tools will have to deal with more and more complex issues. >This year's call for papers attracted over 65 submissions of which 22 were accepted. The focus of ISPD attracts only papers that are on topic, so the accepted papers represent the best of the best. In addition to the technical presentations, we have keynote and invited talks that should be of broad interest. We are particularly pleased to have a keynote from IBM's Jim Kahle, Chief Architect of the Cell processor. The general consensus of the VLSI community is that further performance increases from clock scaling are unlikely; power delivery and heat removal are simply too expensive. Parallel computation appears to be the only economical way to improve performance, and the Cell processor is helping to lead the way. The symposium also features a panel discussion by leading experts on"Rules vs. Tools -- What's the right way to address IC manufacturing complexity?" for 45nm and below. It also features special sessions and invited talks on statistical/physical design for manufacturability and the future interconnects. >In the prior two years, ISPD has held a placement contest, featuring large benchmarks derived from industrial designs. There are at least a dozen groups around the world who are actively working on placement; this year, the newest results for these benchmarks will be posted. >A new contest for global routing is being introduced this year; like the placement contest, we expect a large turnout and fierce competition. The goal of both the placement and routing contests is to set clear objectives for research, and to be able to accurately compare different approaches. While perhaps not as widely known as World Cup Soccer, those involved are just as passionate. >The papers from the symposium are available on the ACM Digital Library; slides from most talks for this and prior years are available on the ISPD web site: http://www.ispd.cc.