RL-Sizer: VLSI Gate Sizing for Timing Optimization using Deep Reinforcement Learning
TimeWednesday, December 8th2:35pm - 2:57pm PST
Digital Design, Timing and Simulation
DescriptionGate sizing for timing optimization is performed extensively throughout Electronic Design Automation (EDA) flows. However, existing algorithms in commercial tools heavily rely on meta-heuristics which lead to sub-optimal sizing solutions. Reinforcement Learning (RL) is a disruptive paradigm that achieves high-quality optimization results beyond traditional algorithms. In this paper, we formulate gate sizing as an RL process, and propose RL-Sizer, an autonomous gate sizing agent, which performs timing optimization in an unsupervised manner. Experimental results on 6 commercial designs in advanced technologies (5-16nm) demonstrate that RL-Sizer consistently outperforms the sizing methodologies of an industry-leading EDA tool in total negative slack reduction.