High performance computing systems for many-core processors

• Main Author: Soh, Alexius Matthias Sheng En
• Other Authors: School of Computer Engineering
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling
• Online Access: http://hdl.handle.net/10356/59201
• Institution: Nanyang Technological University

Computer simulations can be used in different research fields. One of them is to perform real world obstacle detection and avoidance for unmanned vehicles. In addition, very little research has been done on simulating several obstacle detection and avoidance for autonomous aircrafts. The objective of this research is to determine if it is better to perform simulations on the GPU than on the CPU when there are multiple simulations to be performed at any one time. Four to ten unique obstacle layouts were used for the experiment. The simulations were executed sequentially, in multi-threaded approach and using the GPU. The experiment reveals that executing the simulations on the GPU proves to be time efficient as compared to the CPU while allowing more simulations to run in parallel. The result of this study shows that implementing the simulations to run the GPU in comparison to the CPU does allow for greater time efficiency.