Min-Cut Based Leakage Power Aware Scheduling in High-Level Synthesis

Nan Wang1,  Song Chen2,  Takeshi Yoshimura1
1Graduate School of IPS, Waseda University, 2University of Science and Technology of China


Abstract

In this paper, we address the problem of scheduling operations into control steps with dual threshold voltage technique under timing and resource constraints. We present a min-cut based algorithm for leakage power optimization. The proposed algorithm first initializes all the operations to high_Vth, then iteratively shorten the critical path delay by reassigning the set of operations covering all the critical paths to low_Vth until the timing constraints are met. A modified force-directed scheduling is implemented to schedule operations and to adjust operations' Vth assignments with consideration of resource constraints. During this procedure, mobility overlap graph (MOG) is constructed based on the mobilities of high_Vth operations. To guarantee the resource constraints are satisfied, operations' threshold voltages are adjusted by computing the min-cut of mobility overlap graph (MOG). Experimental results show the effectiveness of our proposed algorithm.