Parallel execution of SVM using Symmetrical Multi-Processor (LIBSVM-OMP)

Abstract

Parallel computing is a simultaneous use of multiple compute resources such as processors to solve difficult computational problems. It has been used in high-end computing areas such as pattern recognition, defense, web search engine, and medical diagnosis. This paper focuses on the implementation of pattern classification technique, Support Vector Machine (SVM) using Symmetric Multi-Processor (SMP) approach. We have carried out a performance analysis to benchmark the sequential SVM program against the SMP approach. The result shows that the parallelization of SVM training achieves a better performance than the sequential code speed-ups by 15.9s.