Title :
Three dimensional devices electromagnetic simulation using parallelism on GPUs
Author :
Castaneda, Zady S. ; Maicon, Faria ; Hernandez-Figueroa, Hugo E.
Author_Institution :
Microwave & Opt. Dept., Univ. of Campinas, Campinas, Brazil
Abstract :
In this work we report a several performance tests of a novel implementation of 3D FDTD algorithm in GPU devices using CUDA development. We use new capability 2.0 or more to simplify the process of attribution of threads to electromagnetic fields sites. For one CPU core versus a modern GPU we show speedups up to 100×, with turn to be a speedup of 12× for the CUDA code against a ideal 8 core CPU parallel code.
Keywords :
computational electromagnetics; finite difference time-domain analysis; graphics processing units; parallel architectures; 3D FDTD algorithm; CPU core; CPU parallel code; CUDA development; GPU; electromagnetic fields sites; electromagnetic simulation; parallelism; three dimensional devices; Electromagnetic fields; Finite difference methods; Graphics processing units; Indexes; Kernel; Three-dimensional displays; Time-domain analysis; 3D FDTD; Cuda™; Electromagnetism; GPUs; Nvidia®; Parallelism;
Conference_Titel :
Microwave & Optoelectronics Conference (IMOC), 2013 SBMO/IEEE MTT-S International
Conference_Location :
Rio de Janeiro
DOI :
10.1109/IMOC.2013.6646589
