Energy Estimation for Extensible Processors

Abstract

This paper presents an efficient methodology for estimating the energy consumption of application programs running on extensible processors. Extensible processors, which are increasingly popular in embedded system design, allow a designer to customize a base processor core through instruction set extensions. Existing processor energy macro-modeling techniques are not applicable to extensible processors, since they assume that the instruction set architecture as well as the underlying structural description of the micro-architecture remain fixed. Our solution to this problem is an energy macro-model suitably parameterized to estimate the energy consumption of a processor instance that incorporates any custom instruction extensions. Such a characterization is facilitated by careful selection of macro-model parameters/variables that can capture both the functional and structural aspects of the execution of a program on an extensible processor. Another feature of the proposed characterization flow is the use of regression analysis to build the macro-model. Regression analysis allows for in-situ characterization, thus allowing arbitrary test programs to be used during macro-model construction. We validate the proposed methodology by characterizing the energy consumption of a state-of-the-art extensible processor (Tensilicaýs Xtensa). We use the macro-model to analyze the energy consumption of several benchmark applications with custom instructions. The mean absolute error in the macro-model estimates is only 3.3 %, when compared to the energy values obtained by a commercial tool operating on the synthesized RTL description of the custom processor. Our approach achieves an average speedup of three orders of magnitude over the commercial RTL energy estimator. Our experiments show that the proposed methodology also achieves good relative accuracy, which is essential in energy optimization studies.