Mustapha, Muhammad Firdaus and Abd Khalid, Noor Elaiza and Manaf, Mazani and Ismail, Azlan (2017) Enhanced parallel SOM based on heterogeneous system platform. In: 6th International Conference on Computing & Informatics (ICOCI2017), 25 - 27 April 2017, Kuala Lumpur.

PDF Restricted to Registered users only Download (871kB) | Request a copy

Abstract

In this paper, we propose an enhanced parallel Self organizing Map (SOM) framework based on heterogeneous system platform, specifically Central Processing Unit (CPU) and Graphic Processing Unit (GPU) soldered together on a single chip.The framework is to improve speed of parallel SOM using GPU since processing parallel SOM on GPU burden by communication latency due to isolate device architecture with CPU.The parallel SOM has been extended to heterogeneous system platform and double kernel for calculation distance and find Best Matching Unit (BMU) are introduced.The results are tested using benchmark data on two different platforms: GPU and heterogeneous system. The proposed framework shows improvement compared to standard parallel SOM on GPU and heterogeneous system.

Item Type:	Conference or Workshop Item (Paper)
Additional Information:	eISSN 2289-7402 e-ISBN 978-967-0910-33-8 Organized by: School of Computing, Universiti Utara Malaysia Sintok.
Uncontrolled Keywords:	Parallel Self-organizing map, GPU computing, OpenCL, HSA
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions:	School of Computing
Depositing User:	Mrs. Norazmilah Yaakub
Date Deposited:	27 Jul 2017 01:08
Last Modified:	27 Jul 2017 01:08
URI:	https://repo.uum.edu.my/id/eprint/22835

Actions (login required)

View Item