Tiny ARM based devices are the backbone of the Internet of Things technologies, nevertheless the availability of high performance multicore lightweight CPUs pushed the High Performance Computing to hybrid architectures leveraging on diverse levels parallelism. In this paper we describe how to accelerate inexpensive ARM-based computing nodes with high-end CUDA enabled GPGPUs hosted on x86 64 machines using the GVirtuS general-purpose virtualization service. We draw the vision of a possible hierarchical remote workload distribution among different devices. Preliminary, but promising, performance evaluation data suggests that the developed technology is suitable for real world applications.
Virtualizing CUDA enabled GPGPUs on ARM clusters / Raffaele, Montella; Giulio, Giunta; Laccetti, Giuliano; Lapegna, Marco; Valentina, Pelliccia; Carmine, Ferraro; Carlo, Palmieri. - 9574:(2016), pp. 3-14. (Intervento presentato al convegno Parallel Processing and Applied Mathematics 2015 tenutosi a Krakow (Polonia) nel 6-9 / 9 / 2015) [10.1007/978-3-319-32152-3_1].
Virtualizing CUDA enabled GPGPUs on ARM clusters
LACCETTI, GIULIANO;LAPEGNA, MARCO;
2016
Abstract
Tiny ARM based devices are the backbone of the Internet of Things technologies, nevertheless the availability of high performance multicore lightweight CPUs pushed the High Performance Computing to hybrid architectures leveraging on diverse levels parallelism. In this paper we describe how to accelerate inexpensive ARM-based computing nodes with high-end CUDA enabled GPGPUs hosted on x86 64 machines using the GVirtuS general-purpose virtualization service. We draw the vision of a possible hierarchical remote workload distribution among different devices. Preliminary, but promising, performance evaluation data suggests that the developed technology is suitable for real world applications.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
