A new metric for basic-block level rough energy estimation for power-gated VLIW data-path model

Abstract

This paper proposes a new metric for basic-block level rough energy estimation for fine-grained power-gated VLIW (Very Long Instruction-set Word) data-path model. Power-gating (PG) brings a big benefit for leakage power reduction, but it makes an instruction scheduling difficult because applying PG usually takes dozens or hundreds of NOP cycles. To estimate the energy consumption of such power-gated VLIW data-path, an optimization of instruction scheduling for minimization of energy consumption is necessary. However, such instruction scheduling usually takes long time. This paper proposes a new metric for rough energy estimation taking into account power-gating effect. Proposed metric is constructed by application, architecture parameters, and power-gating parameters. Experimental results show that the correlation coefficient between proposed metric and energy consumption taking into account PG effect was more than 0.84, and the energy estimation by proposed metric is obtained less than one second.