A Survey of Embedded Software Profiling Methodologies

Abstract: Embedded Systems combine one or more processor cores with dedicated logic running on an ASIC or FPGA to meet design goals at reasonable cost. It is achieved by profiling the application with variety of aspects like performance, memory usage, cache hit versus cache miss, energy consumption, etc. Out of these, performance estimation is more important than others. With ever increasing system complexities, it becomes quite necessary to carry out performance estimation of embedded software implemented in a particular processor for fast design space exploration. Such profiled data also guides the designer how to partition the system for Hardware (HW) and Software (SW) environments. In this paper, we propose a classification for currently available Embedded Software Profiling Tools, and we present different academic and industrial approaches in this context. Based on these observations, it will be easy to identify such common principles and needs which are required for a true Software Profiling Tool for a particular application.