GPU Implementation of Image Convolution Using Sparse Model with Efficient Storage Format

View Sample PDF

Author(s): Saira Banu Jamal Mohammed (VIT University, Vellore, India), M. Rajasekhara Babu (School of Computing Science and Engineering, VIT University, Vellore, India)and Sumithra Sriram (VIT University, Vellore, India)
Copyright: 2018
Volume: 10
Issue: 1
Pages: 17
Source title: International Journal of Grid and High Performance Computing (IJGHPC)
Editor(s)-in-Chief: Emmanuel Udoh (Sullivan University, USA)and Ching-Hsien Hsu (Asia University, Taiwan)
DOI: 10.4018/IJGHPC.2018010104

Keywords: Computer Science & IT / Grid & High Performance Computing / High Performance Computing / Information Science Reference

Purchase

View GPU Implementation of Image Convolution Using Sparse Model with Efficient Storage Format on the publisher's website for pricing and purchasing information.

Abstract

With the growth of data parallel computing, role of GPU computing in non-graphic applications such as image processing becomes a focus in research fields. Convolution is an integral operation in filtering, smoothing and edge detection. In this article, the process of convolution is realized as a sparse linear system and is solved using Sparse Matrix Vector Multiplication (SpMV). The Compressed Sparse Row (CSR) format of SPMV shows better CPU performance compared to normal convolution. To overcome the stalling of threads for short rows in the GPU implementation of CSR SpMV, a more efficient model is proposed, which uses the Adaptive-Compressed Row Storage (A-CSR) format to implement the same. Using CSR in the convolution process achieves a 1.45x and a 1.159x increase in speed compared to the normal convolution of image smoothing and edge detection operations, respectively. An average speedup of 2.05x is achieved for image smoothing technique and 1.58x for edge detection technique in GPU platform usig adaptive CSR format.

The IRMA Community

Research IRM

GPU Implementation of Image Convolution Using Sparse Model with Efficient Storage Format

Purchase

Abstract

Related Content

IRMA Sponsors