A paper by ECE graduate students, Jieru Zhao and Liang Feng, supervised by Prof. Wei Zhang won the prestigious IEEE/ACM William J. McCalla ICCAD Best Paper Award
A paper by ECE graduate students, Jieru Zhao and Liang Feng, supervised by Prof. Wei Zhang won the prestigious IEEE/ACM William J. McCalla ICCAD Best Paper Award. The award was announced at the 36th International Conference On Computer Aided Design (ICCAD) in Irvine CA on November 13th, 2017 where the following winning work was presented.
Paper title: “COMBA: A Comprehensive Model-Based Analysis Framework for High Level Synthesis of Real Applications”
Paper Summary: Field Programmable Gate Arrays (FPGAs) are integrated circuits that can be configured or reprogrammed by designers after manufacturing, which shortens the time-to-market. High Level Synthesis (HLS) tools make FPGAs easier to configure, but often result in inefficient hardware designs. To improve the performance of the generated designs, users need to carefully tune the parameters in HLS tools, which is not an easy work. This paper proposes a model-based framework to help users set the parameters, optimize their applications and generate efficient hardware circuits. The proposed tool, COMBA, can accurately analyze the effects of a multitude of optimization techniques related to functions, loops and arrays and quickly find an optimal configuration for complex applications, resulting in speed-ups of more than 100 times. The significant speed-up, small estimation error and fast design space exploration establish COMBA as a superior tool to optimize various complex applications.
