Summary
International Symposium on Nonlinear Theory and its Applications
2017
Session Number:C2L-E-2
Session:
Number:C2L-E-2-1
Parallel Computing of Neural Network Algorithm for Fixed Channel Assignment Problem in Cellular Radio Networks with CUDA
Sho Ikeda, Yoichi Tomioka, Junji Kitamichi,
pp.720-723
Publication Date:2017/12/4
Online ISSN:2188-5079
DOI:10.34385/proc.29.C2L-E-2-1
Summary:
In recent years, graphics processing units (GPUs) have been used for faster numerical calculation because they have many cores and can calculatevia parallel computing. In this paper, we propose a CUDA C program that aims to accelerate the extended maximum neural network algorithm for the fixed channel assignment problem (FCAP) in cellular radio networks using a general-purpose GPU (GPGPU). We evaluate the developed program using the existing benchmark problem in the FCAP. Results show that the processing speed of the developed program is 2.4 times to 15.1 times faster than in the case of using only a CPU.