View Proposal

Proposer
Umar Farooq
Title
The Design, Implementation and Performance Evaluation of GPU based Computing for Intelligent Applications
Goal
Description
Graphical Processing Units (GPUs) are highly powerful parallel computing platforms, which are capable of executing computationally intensive applications such as computer vision, image/video processing, simulations, generative AI or large scale data processing, in time efficient manner. GPU’s architecture makes them particularly suitable for the tasks which are divisible into parallel tasks. This project aims to investigate the application of GPU-based parallel computing to a selected computationally intensive problem. It will explore potential application areas and select an appropriate problem based on its computational requirements, suitability for parallelization, availability of data and feasibility within the project timeframe. The selected domain will be analysed to identify computational bottlenecks and opportunities for GPU acceleration. GPU acceleration and performance evaluation remains the core technical contribution of this project. A prototype shall be implemented using appropriate GPU programming and computing technologies. A CPU-based implementation will be developed where appropriate to provide a baseline for comparison. The project will evaluate the effectiveness of GPU acceleration using measures such as execution time, speedup, throughput, resource utilization, scalability, and energy or computational efficiency where feasible. The project will demonstrate how parallel GPU architectures can be exploited to improve the performance of computationally demanding applications and will provide practical insight into the design, implementation, optimization, and evaluation of GPU-accelerated software.
Resources
Background
Url
Difficulty Level
Moderate
Ethical Approval
None
Number Of Students
1
Supervisor
Umar Farooq
Keywords
cpu, gpu, scalability, throughput
Degrees
Bachelor of Science in Computational Sciences and Software Engineering
Back to List