Today the InfiniBand Trade Association (IBTA) announced the completion of the first Plugfest for RDMA over Converged Ethernet (RoCE) solutions and the publication of the RoCE Interoperability List on the IBTA website. Fifteen member companies participated, bringing their RoCE adapters, cables and switches for testing to the event. Products that successfully passed the testing have been added to the RoCE Interoperability List.
Today Mellanox announced its EDR 100Gb/s InfiniBand solutions have been selected by the KTH Royal Institute of Technology for use in their PDC Center for High Performance Computing. Mellanox’s robust and flexible EDR InfiniBand solution offers higher interconnect speed, lower latency and smart accelerations to maximize efficiency and will enable the PDC Center to achieve world-leading data center performance across a variety of applications, including advanced modeling for climate changes, brain functions and protein-drug interactions.
UCX is a collaboration between industry, laboratories, and academia to create an open-source production grade communication framework for HPC applications. “The path to Exascale, in addition to many other challenges, requires programming models where communications and computations unfold together, collaborating instead of competing for the underlying resources. In such an environment, providing holistic access to the hardware is a major component of any programming model or communication library. With UCX, we have the opportunity to provide not only a vehicle for production quality software, but also a low-level research infrastructure for more flexible and portable support for the Exascale-ready programming models.”
“Learn about extensions that enable efficient use of Partitioned Global Address Space (PGAS) Models like OpenSHMEM and UPC on supercomputing clusters with NVIDIA GPUs. PGAS models are gaining attention for providing shared memory abstractions that make it easy to develop applications with dynamic and irregular communication patterns. However, the existing UPC and OpenSHMEM standards do not allow communication calls to be made directly on GPU device memory. This talk discusses simple extensions to the OpenSHMEM and UPC models to address this issue.”
“In this session we describe how GPUs can be used within virtual environments with near-native performance. We begin by showing GPU performance across four hypervisors: VMWare ESXi, KVM, Xen, and LXC. After showing that performance characteristics of each platform, we extend the results to the multi-node case with nodes interconnected by QDR InfiniBand. We demonstrate multi-node GPU performance using GPUDirect-enabled MPI, achieving efficiencies of 97-99% of a non-virtualized system.”