Abstract :
Summary form only given. As configurable computing matures there is continued interest in design productivity. While configurable computing machines (CCMs) are touted as re-usable, re-configurable platforms for accelerated computing, the design processes and tools employed to map applications to such platforms may have more in common with hardware ASIC design tools and processes than with conventional software development tools and processes. To address this there continues to be research in the community on higher-level languages, reusable core libraries, improved CAD tools, debug infrastructure, etc. but little focus on the big picture. That is, how do all these pieces fit together into a comprehensive approach for improving design productivity? Is a 10x increase in design productivity feasible in the next few years? How about 20x? What are the key research problems that must be solved? What are the most promising approaches to solving these problems? My talk will focus on these questions and will draw from the results of two events related to FPGA design productivity which were held recently. First, the day before FPT´2009, a workshop is being held at Taiwan National University to discuss these questions with all FPT attendees invited to participate. Details are found on the FPT conference website. Special guest speakers at the workshop will be asked to provide their viewpoints on FPGA design productivity and components of a research agenda to increase productivity over the next 5-10 year horizon. Second, during the past year in the United States, two multi-university research teams have focused on developing and outlining a research agenda to address FPGA design productivity, culminating in a workshop held in Salt Lake City during June 2008. My talk will discuss the current challenges in FPGA design productivity and address a range of solutions, based in part on the results of the above two workshops as well as other efforts going on in the research community.
Keywords :
application specific integrated circuits; field programmable gate arrays; logic CAD; reconfigurable architectures; FPGA design productivity; Taiwan National University; configurable computing machines; hardware ASIC design tools; Acceleration; Application software; Application specific integrated circuits; Design automation; Field programmable gate arrays; Hardware; Process design; Productivity; Programming; Software libraries;