Performance Comparison of Parallel Execution Using GPU and CPU in SVM Training Session

Support Vector Machine (SVM) is a machine learning approach, which is used in a growing number of applications. SVM is a useful technique for data classification. This machine learning approach has been optimized using two (2) parallel computing approaches. This includes symmetric multiprocessor (SMP) approach and vector processor approach. The outcome performance of the implementation of symmetric multiprocessor approach and vector processor approach on SVM training session is the focus of this paper. We have carried out a performance analysis to benchmark between Central Processing Unit (CPU) and Graphics Processing Units (GPUs) optimization. The result shows the GPU optimization of SVM training duration achieves better performance than the CPU optimized program by 3.11 of speedup.