A statistical model of input glitch propagation and its application in power macromodeling

Power macromodeling technique can result in huge underestimation without the consideration of input glitches. In this paper, we propose a statistical model for the propagation of input glitches and their effects on circuit power consumption. Based on this model, we develop an analytical power macromodeling approach incorporating input glitches. Specifically, we divide the macromodel parameter space into three regions and characterize each region separately. We have evaluated the proposed technique on ISCAS85 and LGSynth93 benchmark circuits. Compared with switch level simulation results, the average power estimation errors are 1.8% and 3.1% for combinational and sequential circuits, respectively. Our model also provides useful insight for glitchy circuit identification and input glitch power reduction.