Computer Science Professor Jason Cong and his students Licheng Guo, Yuze Chi, Jie Wang, Jason Lau, and Weikang Qiao, in collaboration with Professor Zhiru Zhang and his student Ecenur Ustun at Cornell University, have received the Best Paper Award from the 29th ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA 2021) for their paper “AutoBridge: Coupling Coarse-Grained Floorplanning and Pipelining for High-Frequency HLS Design on Multi-Die FPGAs” in March 2021. AutoBridge addresses an important problem of how to raise the level of design abstraction while still achieving high-frequency designs. This paper provides important contributions to improving the timing quality of HLS compilation techniques and introduces an effective methodology by coupling coarse-grained floorplanning with high-level synthesis (HLS) to enable interconnected pipelining. AutoBridge improves the clock frequency by 2X on average when tested on a wide range of large-scale designs on multi-die FPGAs.
The ACM/SIGDA International Symposium on Field-Programmable Gate Arrays is the premier conference for the presentation of advances in all areas related to FPGA technology. Among the papers accepted this year, the program committee nominated three papers, which were reviewed by another confidential selection committee. The selection committee unanimously declared AutoBridge to be the final winner. A total of 135 papers were submitted to FPGA 2021, among those, 26 were accepted, and one was selected for the Best Paper Award.