StrongARM Computational Performance Benchmarks - An Analysis of FFT Algorithms

This document summarizes the results of experiments performed on a StrongARM-based embedded platform to estimate its computational performance under various operational conditions. Appendices provide detailed results and further analysis of the computational performance data.

Tests were performed using Fast Fourier Transform (FFT) algorithms, which are a staple of modern digital signal processing (DSP) systems. The algorithms were adapted from public domain sources and were implemented within "instrumented" software applications, so as to measure the computational speed of execution. Performance was measured in two operating system configurations: no operating system and with the Windows® CE operating system.

The experiments show that to calculate a 16384-point, fixed-point FFT on real time domain data requires approximately 62 milliseconds with no operating system and approximately 107 milliseconds with Windows CE. A 16384-point, fixed-point inverse FFT on complex frequency domain data requires approximately 127 milliseconds with no operating system and approximately 302 milliseconds with the Windows CE operating system.

View Entire Paper | Previous Page | White Papers Search

If you found this page useful, bookmark and share it on: