Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Exascale Computing Project – Driving a HUGE Change in a Changing World

“In this keynote, Al Geist will discuss the need for future Department of Energy supercomputers to solve emerging data science and machine learning problems in addition to running traditional modeling and simulation applications. The ECP goals are intended to enable the delivery of capable exascale computers in 2022 and one early exascale system in 2021, which will foster a rich exascale ecosystem and work toward ensuring continued U.S. leadership in HPC. He will also share how the ECP plans to achieve these goals and the potential positive impacts for OFA.”

OFA Workshop Posts Session Abstracts for Austin Meeting in March

Today the OpenFabrics Alliance (OFA) published the session abstracts for its 13th Annual OFA Workshop. Sponsored by Intel, the workshop takes place March 27-31 in Austin, Texas. “The workshop will include more than 50 sessions covering a variety of critical networking topics delivered by industry experts from around the world. Additionally, the OFA has announced that Al Geist of Oak Ridge National Laboratory (ORNL) will deliver a workshop keynote address on the impact of the Exascale Computing Project. The workshop program is designed to educate attendees and encourage lively exchanges among OFA members, developers, and users who share a vested interest in high performance networks.”

Video: Intel Omni-Path Fabric Management and Tool Features

In this video from the 2016 OpenFabrics Workshop, James Wright from Intel presents: Intel Omni-Path Fabric Management and Tools Features. “The Intel Omni-Path Fabric includes a number of hardware and software features to make fabric monitoring, management and diagnosis easier. This session will provide a brief overview of the management software architecture and key features.”

Video: UPC++ Parallel Programming Extension

In this video from the 2016 OpenFabrics Workshop, Zili Zheng from LBNL presents: UPC++. “UPC++ is a parallel programming extension for developing C++ applications with the partitioned global address space (PGAS) model. UPC++ has demonstrated excellent performance and scalability with applications and benchmarks such as global seismic tomography, Hartree-Fock, BoxLib AMR framework and more. In this talk, we will give an overview of UPC++ and discuss the opportunities and challenges of leveraging modern network features.”

Using High Performance Interconnects in Dynamic Environments

“Over the last years the OFA community has shown the potential of using high performance networks (InfiniBand) to boost the performance of virtualized cloud environments, however, the network reconfiguration challenges still continue to exist. In this session we present the work we have been doing on InfiniBand subnet management and routing, in the context of dynamic cloud environments. This work includes, but not limited to, techniques in order to provide better management scalability when virtual machines are live migrating, tenant network isolation in multi-tenant environments, and fast performance-driven network reconfiguration.”

Video: Exploiting HPC Technologies to Accelerate Big Data Processing

“This talk will present RDMA-based designs using OpenFabrics Verbs and heterogeneous storage architectures to accelerate multiple components of Hadoop (HDFS, MapReduce, RPC, and HBase), Spark and Memcached. An overview of the associated RDMAenabled software libraries being designed and publicly distributed as a part of the HiBD project.”

Video: Software-Defined Networking on InfiniBand Fabrics

A design for virtual Ethernet networks over IB is described. The virtual Ethernet networks are implemented as overlays on the IB network. They are managed in a flexible manner using software. Virtual networks can be created, removed, and assigned to servers dynamically using this software. A virtual network can exist entirely on the IB fabric, or it can have an uplink connecting it to physical Ethernet using a gateway. The virtual networks are represented on the servers by virtual network interfaces which can be used with para-virtualized I/O, SRIOV,and non-virtualized I/O. This technology has many uses: communication between applications which are not IB-aware, communication between IB-connected servers and Ethernet-connected servers, and multi-tenancy for cloud environments. It can be used in conjunction with OpenStack, such as for tenant networks.”

Video: Past, Present, and Future of OpenFabrics Interfaces

“The first release of OpenFabrics Interfaces (OFI) software, libfabric, occurred in January of 2015. Since then, the number of fabrics and applications supported by OFI has increased, with considerable industry momentum building behind it. This talk discusses the current state of OFI, then speaks to the application requirements driving it forward. It will detail the fabrics supported by libfabric, identify applications which have ported to it, and outline future enhancements to the libfabric interfaces and architecture.”

Gary Grider Presents: HPC Storage and IO Trends and Workflows

“Trends in computer memory/storage technology are in flux perhaps more so now than in the last two decades. Economic analysis of HPC storage hierarchies has led to new tiers of storage being added to the next fleet of supercomputers including Burst Buffers or In-System Solid State Storage and Campaign Storage. This talk will cover the background that brought us these new storage tiers and postulate what the economic crystal ball looks like for the coming decade. Further it will suggest methods of leveraging HPC workflow studies to inform the continued evolution of the HPC storage hierarchy.”

Gary Grider from LANL to Keynote OpenFabrics Workshop

Gary Grider from Los Alamos National Laboratory will keynote the 2016 OpenFabrics Workshop this year with a talk on HPC Storage and IO Trends and Workflows. The event takes place April 4-8, 2016 in Monterey, California.