Tampere University of Technology

TUTCRIS Research Portal

Implementation of a low-complexity low-latency arbitrary resampler on GPUs

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Details

Original languageEnglish
Title of host publication2014 IEEE Dallas Circuits and Systems Conference: Enabling an Internet of Things - From Sensors to Servers, DCAS 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781479959235
DOIs
Publication statusPublished - 21 Nov 2014
Publication typeA4 Article in a conference publication
Event10th IEEE Dallas Circuits and Systems Conference, DCAS 2014 - Richardson, United States
Duration: 12 Oct 201413 Oct 2014

Conference

Conference10th IEEE Dallas Circuits and Systems Conference, DCAS 2014
CountryUnited States
CityRichardson
Period12/10/1413/10/14

Abstract

Modern communication systems have data rates and sampling rates that are tightly coupled. Resampling is necessary in order to convert to some desired sampling rate, which is usually a multiple of the data rate. The resampling process is an integral part of transceiver systems and must be designed accurately and carefully. In this paper, we present a low complexity and low latency arbitrary resampling method based on graphics processing units (GPUs). Our proposed flexible and all-software-based resampling method requires no precomputation of filters and yet yields high performance by taking advantage of unique features found in GPUs.

Keywords

  • Arbitrary resampling, DSP accelerator, GPU front-end receiver, GPU-based radio, sample rate conversion