Professor Purdy is investigating modifications to the area-time complexity model of computation, which will enable her to more accurately predict the performance of algorithms designed to be implemented in VLSI circuits. Initial research is in the CMOS technology, and development of tools for experimental research. Issues being addressed are: understanding the problems in deriving practical circuit models from theoretical models; more accurate inputs and outputs; formalization of "rules of thumb" for VLSI circuit design which will be useful in developing more powerful silicon compilers; and the development of a methodology for rapid initial design and simulation of large families of VLSI chips, based on different algorithms, in order to predict accurately trade-offs in efficiency and yield of fabricated chips. The proposed research is in the intersection between theory and practice of automated design of VLSI chips. The Principal Investigator's approach is novel. She has successfully begun a change in her research career from mathematics to IC design algorithms, and brings to bear useful mathematical methods and techniques on the problems. She has the potential to advance her career by doing quality research.
