Analysing and improving the performance of software code for Real Time Embedded systems

Abstract

Real Time Embedded systems are characterized by the presence of a processor, on which application specific programs have to run. These application specific dedicated programs are called “Embedded Software or Firmware”. Since most of the activity of the processor is controlled by the software, its efficient design plays a key role in the system design. Efficient software or firmware programming also leads to high performance embedded systems. Also in the present scenario ARM based processors are most widely accepted for higher end applications. Hence in this paper we mainly focus on the analysis and optimization of the Embedded software code running on ARM processors. Various techniques like data dependency, loop optimization, etc., are investigated and studied to enhance the performance of the software in terms of code density, speed of operation, both at high level (Embedded-C) and at the assembly level. Comparing the obtained results with the works in the literature, shows an improvement of 40% with respect to the code density and by 30% in the speed of operation. Further efficient embedded firmware like linear convolution and LCD module are designed by using various optimization techniques. All the simulations are carried out using the IAR workbench.