The integration of graphics processing units (GPUs) on high-end compute nodes has established a new accelerator-based heterogeneous computing model, which now permeates high performance computing. The same paradigm nevertheless has limited adoption in cloud computing or other large-scale distributed computing paradigms. Heterogeneous computing with GPUs can benefit the Cloud by reducing operational costs and improving resource and energy efficiency. However, such a paradigm shift would require effective methods for virtualizing GPUs, as well as other accelerators. In this survey paper, we present an extensive and in-depth survey of GPU virtualization techniques and their scheduling methods. We review a wide range of virtualization techniques implemented at the GPU library, driver, and hardware levels. Furthermore, we review GPU scheduling methods that address performance and fairness issues between multiple virtual machines sharing GPUs. We believe that our survey delivers a perspective on the challenges and opportunities for virtualization of heterogeneous computing environments.

Abstract. Some basics of combinatorial block design are combined with certain constraint satisfaction problems of interest to the satisfiability community. The paper shows how such combinations lead to satisfiability problems, and shows empirically that these are some of the smallest very hard satisfiability problems ever constructed. Partially balanced´¼ ½µ designs (PB01Ds) are introduced as an extension of balanced incomplete block designs (BIBDs) and´¼ ½µ designs. Also,´¼ ½µ difference sets are introduced as an extension of certain cyclical difference sets. Constructions based on´¼ ½µ difference sets enable generation of PB01Ds over a much wider range of parameters than is possible for BIBDs. Building upon previous work of Spence, it is shown how PB01Ds lead to small, very hard, unsatisfiable formulas. A new three-dimensional form of combinatorial block design is introduced, and leads to small, very hard, satisfiable formulas. The methods are validated on solvers that performed well in the SAT 2009 and earlier competitions.

The astonishing development of diverse and different hardware platforms is twofold: on one side, the challenge for the exascale performance for big data processing and management; on the other side, the mobile and embedded devices for data collection and human machine interaction. This drove to a highly hierarchical evolution of programming models. GVirtuS is the general virtualization system developed in 2009 and firstly introduced in 2010 enabling a completely transparent layer among GPUs and VMs. This paper shows the latest achievements and developments of GVirtuS, now supporting CUDA 6.5, memory management and scheduling. Thanks to the new and improved remoting capabilities, GVirtus now enables GPU sharing among physical and virtual machines based on x86 and ARM CPUs on local workstations, computing clusters and distributed cloud appliances.

No abstract

Abstract