On June 24, 2018, the local time in the United States, the 55th ACM/IEEE Design Automation Conference 2018 was held at Moscone Center, San Francisco. In the corresponding period of the conference, the award ceremony of the Design Contest FPGA Group was held. The undergraduate team with 6 students of Zeng Shulin, Chen Weicong, Huang Tianhao, Lin Yuyun, Zhu Zhenhua, et al. of our department of Grade 2014 led by their supervisor Wang Yu, won the first place in the contest by virtue of the hardware accelerator DPU IP, full-stack tool chain DNNDK, and deep compression technology at the algorithm level.

The System Design Contest was sponsored by DAC, aiming to provide a technical exchange platform for the global machine learning system designers and explore excellent solutions and talents in the field. This contest was based on the embedded system and conducted target detection on the image source provided by the UAV. The contest system stipulated to divide according to the two hardware architectures of FPGA and GPU for contest respectively. Among them, GPU used the NVIDIA Jetson TX2 platform while FPGA used the Xilinx PYNQ-Z1 board. Through the competition in the comprehensive performance in accuracy, timeliness and power consumption and so on, the top three groups with the best performance in each architecture were evaluated. A total of 114 teams participated in the contest in the world, including 53 teams in the GPU architecture group and 61 teams in the FPGA architecture group. The TGIIF team led by Professor Wang Yu participated in the contest of the FPGA architecture group. Swiss Federal Institute of Technology Zurich won the second place while the American University of Illinois won the third place.