INCREASING THE PERFORMANCE OF JPEG2000 ENCODING THROUGH HIGH-PERFORMANCE COMPUTING

摘要

This paper describes a methodology for parallelize a JPEG2000 encoder through the usage of a Beowulf Cluster and an already existing open-source JPEG2000 encoding software. It starts by giving an overview of the technical architecture solution chosen for the image cluster, justifying the choices that were followed. Next, the coding algorithm is presented, both on it's non-parallel as well as the parallel version, discussing which were the modifications performed in the original version to improve its performance on a high performance computing platform. Afterwards, the tests that were performed on the encoder are presented and the results achieved are discussed. Finally, some conclusions and future work are presented. The major focus of this paper is on High Performance Computing hardware aspects of the proposed solution, reinforcing the idea that high-performance solutions might be obtained through relatively low cost hardware architecture solutions.