Analyzing sensitivity of code execution efficiency to software and hardware parameters

Abstract

Groups of code basic blocks have been employed in the past as software-related input parameters of processor models. In this paper, we use different block group variations to study their relationship to the execution efficiency of a superscalar processor. For this purpose, we utilized frequencies of blocks to represent dynamic nature of 10 benchmark programs. We were able to get a reduced set of hardware and software parameters by employing correlation coefficients and principal component analysis.