Open AccessFeature PaperReview
Journal of Electrical Engineering and Computer Science (JEECS)

Equidistant execution of resemblance computation on GPU construction using CUDA

(This article belongs to Vol - 02, Issue - 01)
Download [ 3 ]

Abstract

Picture handling and example acknowledgment calculations set aside more effort for execution on a solitary center processor. Illustrations Processing Unit (GPU) is progressively mainstream now-a-days because of their speed, programmability, minimal effort and progressively inbuilt execution centers in it. A large portion of the analysts began work to utilize GPUs as a handling unit with a solitary center PC framework to speedup execution of calculations and in the field of Content based clinical picture recovery (CBMIR), Euclidean separation and Mahalanobis assumes a significant job in recovery of pictures. Separation recipe is significant on the grounds that it assumes a significant job in coordinating the pictures. Right now, we parallelized Euclidean separation calculation on CUDA. CPU with Intel® Dual-Core E5500 @ 2.80GHz and 2.0 GB of principle memory which run on Windows XP (SP2). The following stage was to change over this code in GPU position for example to run this program on GPU NVIDIA GeForce arrangement 9500GT model having 1023 MB of video memory of DDR2 type and transport width of 64bit. The realistic driver we utilized is of 270.81 arrangement of NVIDIA. Right now, the CPU and GPU form of calculation is being actualized on the MATLAB R2010. The CPU adaptation of the calculation is being dissected in straightforward MATLAB however the GPU variant is being executed with the assistance of moderate programming Jacket-win-1.3.0. For utilizing Jacket, we need to make a few changes in our source code so to make the CPU and GPU to work at the same time and accordingly lessening the by and large computational speeding up. Our work utilizes broad use of exceptionally multithreaded design of multicored GPU. An effective utilization of shared memory is required to enhance equal decrease in Compute UnifiedGadget Architecture (CUDA), Graphic Processing Units (GPUs) are developing as amazing equal frameworks at a modest expense of two or three thousand rupees.

Article Access Statistics

1
16

ISO CERTIFIED

Journal of Electrical Engineering and Computer Science (JEECS) ISSN 2287-106X Published by SDIP, London, United Kingdom.
Back to Top