GPU Programlama tekniği ile yüksek performanslı araç takibi
Loading...
Date
2017
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Bilişim Teknolojileri Dergisi
Access Rights
info:eu-repo/semantics/openAccess
Abstract
Özet: Bu çalışmada, Grafik İşlemci Birimi (GPU) Programlaması kullanılarak yüksek performanslı bir araç takip uygulaması geliştirilmiştir. GPU programlama ortamı olarak Birleşik Hesap Cihazı Mimarisi (CUDA) kullanılmıştır. Uygulama, GeForce GT 630 ve GeForce GTX 550Ti isimli iki farklı ekran kartı üzerinde test edilmiş ve farklı GPU’lara sahip ekran kartlarının uygulamanın performansına olan etkisi incelenmiştir. Ayrıca uygulama MATLAB ortamında Merkezi İşlemci Birimi (CPU) üzerinde de gerçeklenerek, CPU-GPU karşılaştırılması yapılmış ve ayrıntılı sonuçlar sunulmuştur. Elde edilen sonuçlar, araç takibi işleminde GPU programlamanın kullanılmasının yüksek performans kazanımı getirdiğini göstermiştir
Abstract: In this study, a high performance vehicle tracking application was implemented using the Graphic Processing Unit (GPU) Programming. For the GPU programming environment, Compute Unified Device Architecture (CUDA) was used. The application was tested on two different graphics card, namely GeForce GT 630 and GeForce GTX 550Ti and the effect of the graphics cards with different GPUs on the performance of the application was inspected. Also, the application was implemented for the Central Processing Unit (CPU) in MATLAB environment, CPU vs GPU comparisons were done and detailed results are presented. The detailed results show that in vehicle tracking operation, using GPU programming achieves high performance.
Abstract: In this study, a high performance vehicle tracking application was implemented using the Graphic Processing Unit (GPU) Programming. For the GPU programming environment, Compute Unified Device Architecture (CUDA) was used. The application was tested on two different graphics card, namely GeForce GT 630 and GeForce GTX 550Ti and the effect of the graphics cards with different GPUs on the performance of the application was inspected. Also, the application was implemented for the Central Processing Unit (CPU) in MATLAB environment, CPU vs GPU comparisons were done and detailed results are presented. The detailed results show that in vehicle tracking operation, using GPU programming achieves high performance.
Description
Keywords
Grafik işlemci birimi, CUDA, araç takip, görüntü işleme, Graphic processing unit, CUDA, vehicle tracking, image processing
Journal or Series
Bilişim Teknolojileri Dergisi
