Computer Architecture Research:
Is it Dead? Does it need revitalization?
Where do we go from here?
Abstract:
Computer Architecture is at the interface between the software that specifies what the computer is to do, and the hardware that actually carries out the work. Process technology continues to give us more and more resources. The first microprocessor had 2300 transistors, and ran at about 100 KHz. This year we are talking about 1.72 billion transistors on a chip, and clock frequencies in excess of 4 GHz. In a few years we will have 10 billion transisters on a chip, with clock frequencies greater than 10 GHz. However, too many people who should know better are speaking doom. Either they say there is nothing left to do in computer architecture, or they are trumpeting that computer architecture research needs revitalization. In this talk, I will examine both premises. I also hope to suggest some approaches that I expect will keep us moving forward, continuing to improve the performance of microprocessors over the next ten years, ... IF we do not lose sight of some fundamental principles.
Biography of Professor Yale N. Patt:
Yale Patt is the Ernest Cockrell, Jr. Centennial Chair in Engineering and Professor of ECE at Texas. He enjoys equally teaching both the large (400 students) freshman introductory course in computing and the advanced graduate courses in microarchitecture, directing the research of very bright PhD students (right now, 12 of them), and consulting extensively in the microprocessor industry. He earned the appropriate set of degrees from reputable universities and has received a substantial number of awards for both his research and teaching. Many of his research results have found their way into high performance microprocessor products. More detail can be found on his home page, http://www.ece.utexas.edu/~patt.
