Abstract | ||
---|---|---|
The success of Android is based on its unified Java programming model that allows to write platform-independent programs for a variety of different target platforms. In this paper we describe the first, to the best of our knowledge, offloading platform that enables Android devices with no GPU support to run Nvidia CUDA kernels by migrating their execution on high-end GPGPU servers. The framework is highly modular and exposes a rich Application Programming Interface (API) to the developers, making it highly transparent and hiding the complexity of the network layer. We present the first preliminary results, showing that not only GPGPU offloading is possible but it is also promising in terms of performance. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1007/978-3-319-49583-5_9 | ALGORITHMS AND ARCHITECTURES FOR PARALLEL PROCESSING, ICA3PP 2016 |
Keywords | Field | DocType |
Virtualization,GPGPU,CUDA,Cloud,Android,Offloading | Android (operating system),Programming paradigm,CUDA,Computer science,Server,General-purpose computing on graphics processing units,Application programming interface,Java,Operating system,Cloud computing,Distributed computing | Conference |
Volume | ISSN | Citations |
10048 | 0302-9743 | 4 |
PageRank | References | Authors |
0.42 | 7 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Raffaele Montella | 1 | 210 | 23.13 |
Carmine Ferraro | 2 | 18 | 1.00 |
Sokol Kosta | 3 | 276 | 24.92 |
Valentina Pelliccia | 4 | 24 | 1.43 |
Giulio Giunta | 5 | 162 | 13.35 |