This year, computer science professor Jason Cong and former UCLA student Yuzheng Ding were recognized by the TCFPGA Hall of Fame for their 1994 paper, FlowMap: An Optimal Technology Mapping Algorithm for Delay Optimization in Lookup-Table Based FPGA Designs. The induction endorsement praised Cong and Ding for “pioneering work in the area of technology mapping for field-programmable gate arrays and presenting a theoretical breakthrough by solving the lookup table-based technology mapping problem for depth minimization optimally in polynomial time”. It concludes by noting: “This paper also serves as a great reference material and tutorial for students and industry professionals who are interested in learning about FPGA technology mapping and design implementation.”
The ACM SIGDA Technical Committee on FPGAs and Reconfigurable Computing is a SIGDA (Special Interest Group for Design Automation) Technical Committee formed to promote research within the realm of field-programmable gate arrays, as well as provide guidance for new researchers entering the field. This is the inaugural year of the TCFPGA Hall of Fame, which recognizes the most significant and influential papers which have shaped the field. Papers must at least 10 years old to be nominated, followed by detailed consideration by a panel of experts.
“I’m honored that the ACM/SIGDA Technical Committee on FPGAs selected our work on FlowMap back in the early 1990s as the first paper in the Hall of Fame FPGA and Reconfigurable Computing. It’s a recognition of the multi-decade research program on FPGA at UCLA,” Cong said. “I hope that this inspires more students to search for provably optimal algorithms for real-life problems instead of settling too quickly for simple heuristics.”
Cong began his work on the FlowMap algorithm in the early 1990s, shortly after he joined the UCLA computer science department. The FlowMap algorithm takes on the technology mapping problem, which is the problem of implementing a combinational circuit using the gates of a particular technology library, and plays an integral part in the VLSI (very-large-scale integration) design automation flow. During their time at UCLA, Cong and Ding were able to develop and present a theoretical breakthrough by developing a polynomial-time depth-optimal algorithm for the technology mapping problem, which would eventually become known as the FlowMap algorithm.
FPGAs have been widely used for communication infrastructures worldwide, and Cong’s group has been working on design automation tools to make it easy and more efficient to design FPGAs. In his more recent research, Cong proposed that the inefficiency of general purpose processors could be circumvented by instead using a totally customizable logic via FPGAs. His high-level synthesis solutions enable software programmers to design FPGA-based accelerators using C/C++ languages. Now, FPGAs are widely used as acceleration engines for many computationally intensive tasks (for example, in the Amazon public computing clouds), and Cong and Ding’s work serves as a cornerstone for performance optimization during the logic synthesis.
The TCFPGA Hall of Fame is an ongoing and sustainable mechanism to recognize the most significant contributions to the field of FPGAs and reconfigurable computing. Read more about TCFPGA and their efforts to support the FPGA community here.