Sign up for our newsletter and get the latest HPC news and analysis.

Video: The InfiniBand Subnet Manager from Fabriscale

fabriscale-logo4

In this video from the 2015 HPC Advisory Council Switzerland Conference, Tor Skeie from Simula Laboratory presents: The InfiniBand Subnet Manager from Fabriscale.

Video: Accelerating Big Data Processing with Hadoop, Spark and Memcached

DK Panda, Ohio State University

“Using the publicly available software packages in the High-Performance Big Data (HiBD) project, we will provide case studies of the new designs for several Hadoop/Spark/Memcached components and their associated benefits. Through these case studies, we will also examine the interplay between high performance interconnects, storage systems (HDD and SSD), and multi-core platforms to achieve the best solutions for these components.”

High Throughput Data Acquisition at the CMS experiment at CERN

imgres

“The CMS detector at the Large Hadron Collider at CERN underwent a replacement of its data acquisition network to be able to process the increased data rate expected in the coming years. We will present the architecture of the system and discuss the design of its layers which are based on Infiniband as well as 10 and 40 GBit/s Ethernet.”

HPC News Roundup for March 27, 2015

hatnews

I’m on my way home from a series of Springtime HPC conferences with boatload of new videos and interviews on the latest in high performance computing. Here are some notable items that may have not made it to the front page.

HPC Market Update from Intersect360 Research

addison2

“The drive toward exascale computing, a renewed emphasis on data-centric processing, energy efficiency concerns, and the limitations of memory and I/O performance are all working to reshape High Performance Computing platforms. Many-core accelerators, flash storage, 3D memory, integrated networking, and optical interconnects are just some of the technologies propelling these future architectures. In concert with those developments, the HPC vendor landscape has been churning in response to broader market forces, and these events are going to drive some interesting changes in the coming year.”

Video: Application Experiences on DDN’s IME Burst Buffer

ddn

“The DDN Infinite Memory Engine (IME) unleashes a new I/O provisioning paradigm. This breakthrough, software defined storage application introduces a whole new new tier of transparent, extendable, non-volatile memory (NVM), that provides game-changing latency reduction and greater bandwidth and IOPS performance for the next generation of performance hungry scientific, analytic and big data applications – all while offering significantly greater economic and operational efficiency than today’s traditional disk-based and all flash array storage approaches that are currently used to scale performance.”

Video: Increasing Cluster Throughput while Reducing Energy Consumption for GPU Workloads

rcuda

“The use of GPUs to accelerate applications is mainstream nowadays, but their adoption in cur- rent clusters presents several drawbacks. In this talk we present the last developments of the rCUDA remote GPU virtualization framework, which is the only one supporting the most recent CUDA version, in addition to leverage the InfiniBand fabric for the sake of performance.”

HPC and Software Modernization

Dr. Herbert Cornelius, Intel

“As we see Moore’s Law alive and well, more and more parallelism is introduced into all computing platforms and on all levels of integration and programming to achieve higher performance and energy efficiency. We will discuss Multi- and Many-Core solutions for highly parallel workloads with general purpose and energy efficient technologies. We will also touch on the challenges and opportunities for parallel programming models, methodologies and software tools to achieve highly efficient and highly productive parallel applications. At the end we will take a brief look towards Exascale computing.”

Video: Slim Fly – A Cost Effective Low-Diameter Network Topology

torsten

“We introduce a high-performance cost-effective network topology called Slim Fly that approaches the theoretically optimal network diameter. Slim Fly is based on graphs that approximate the solution to the degree-diameter problem. We analyze Slim Fly and compare it to both traditional and state-of-the-art networks. Our analysis shows that Slim Fly has significant advantages over other topologies in latency, bandwidth, resiliency, cost, and power consumption.”

Video: High-Performance and Scalable Designs of Programming Models for Exascale Systems

DK Panda, Ohio State University

“This talk will focus on programming models and their designs for upcoming exascale systems with millions of processors and accelerators. Current status and future trends of MPI and PGAS (UPC and OpenSHMEM) programming models will be presented. We will discuss challenges in designing runtime environments for these programming models by taking into account support for multi-core, high-performance networks, GPGPUs, Intel MIC, scalable collectives (multi-core-aware, topology-aware, and power-aware), non-blocking collectives using Offload framework, one-sided RMA operations, schemes and architectures for fault-tolerance/fault-resilience.”