Implementation of LTE Mini receiver on GPUs
Faculty of Electrical Engineering and Information Technology, Institute of Telecommunications, Vienna University of Technology
Vienna University of Technology, 2011
@article{vieira2011implementation,
title={Implementation of LTE Mini receiver on GPUs},
author={Vieira, Jose},
year={2011}
}
Long Term Evolution (LTE) is the latest standard for cellular mobile communication. To fully exploit the available spectrum, LTE utilizes feedback. Since the radio channel is varying in time, the feedback calculation is latency sensitive. In our upcoming LTE measurement with the Vienna Multiple Input Multiple Output (MIMO) Testbed, a low latency feedback calculation is necessary. The presented work consists of exploring the computational potential of Graphics processing unit (GPUs) NVIDIA Technology and Compute Unified Device Architecture (CUDA) programming model, with aim to optimize the execution time of a LTE Mini Receiver. Two implementations are presented, first is sequential and second is parallel using CUDA. Thus, I will present a brief study of some techniques for parallel programming and CUDA architecture. In the end, I will present and compare all obtained results between MATLAB, CUDA and C. The comparison is made through the processing time of these three implementations.
February 20, 2012 by hgpu
