high performance computing on graphics processing units: hgpu.org

Recent years have seen an increasing need for computationally efficient implementation of software-defined radio (SDR) systems. Given the limitations of a typical SDR application running on a single machine, we present a distributed SDR system using high-performance techniques. To split a digital signal into multiple channels, we use an efficient digital signal processing technique: a channelizer. Distributed machines then process the channelized streams. To achieve this, we implement a load-balancer module and add new internet protocol communication capabilities to an existing SDR framework. Since the channelizer cannot be efficiently distributed, a single host machine has to process the entire wide-band signal. Furthermore, we optimize an existing implementation of a CPU channelizer and implement a new original GPU channelizer to push the system limitations further. The described techniques are applied to and tested with an existing implementation of a passive GSM receiver to add a full-band capability.