In this video, Doug Miles from The Portland Group discusses how the PGI Accelerator, which is designed to help programmers make their code go faster on x64+GPU platforms. Recorded at SC11.
Using PGI Accelerator compilers, programmers can accelerate Linux, Mac OS X and Windows applications on x64+GPU platforms by adding OpenMP-like compiler directives to existing high-level standard-compliant Fortran and C programs and then recompiling with appropriate compiler options.
In this video, Bjorn Andersson from BlueArc discusses the company’s advanced storage solutions for Big Data and HPC. Recorded at SC11 in Seattle.
Joining forces with Hitachi Data Systems is a win-win situation for our employees, our partners, our customers and our shareholders,” said Mike Gustafson, CEO, BlueArc. “The strength of the long-standing partnership between the two companies spanning product integration, customer momentum and cross-functional collaboration is the foundation for this union. We are excited to accelerate our vision and innovation and to bring real solutions to customers, partners and the market. We view this as a critical time in information technology, where we will help customers assess and map their business priorities and deliver information services as we work collectively to continue to define next generation information centers amidst the expanding world of file and content.”
In this video Tim McIntire from StackIQ discusses how the company’s Big Infrastructure Management soluitons make it easy to deploy cloud computing for Big Data applications.
Analyzing Big Data takes a lot of computing power. We’ve spent the last 10 years building enterprise-class software that makes the task of designing, deploying, and managing compute clusters of any size a breeze. Now we’ve added support for the major Hadoop distributions so you can leverage Rocks+ to spin up production-ready Big Data solutions in minutes, not months.
In this video, Spectra Logic CMO Molly Rector describes the company’s latest tape storage technologies at SC11 in Seattle.
Spectra T-Finity sets a new benchmark in library design with an intelligent Enterprise library that offers a highly flexible hardware architecture and an evolutionary management interface, helping you achieve the highest scalability, availability and performance possible. Built upon Spectra’s history of leading edge library design and the comprehensive BlueScale software interface, T-Finity offers several advantages in a value-packed solution: component redundancy, high-performance dual robotics, industry leading storage density, extreme scalability, energy efficiency, ease of management, integrated encryption and enduring reliability.”
In this video, Joe Yaworski from QLogic discusses the company’s TOP500 momentum with their TrueScale InfiniBand technology. Recorded at SC11 in Seattle.
Building upon the advanced “host on-load” design is a very efficient protocol designed to deliver unsurpassed message rates to power even the most complex MPI applications and workloads. Measured using industry accepted benchmarks at nearly 30 million non-coalesced messages per second, no other InfiniBand solution comes close to matching QLogic TrueScale.
In this video, Bruce Toal from Convey Computer discusses how the company’s hybrd-core systems provide superior performance on graph computing applications. Recorded at SC11 in Seattle.
There is little doubt that memory systems are an ‘Achilles Heel’ of handling big data applications. Today’s commodity systems are optimized for sequential memory accesses, not the random accesses typically found in graph problems. This really hurts performance when processing large-scale analytics applications,” explained Bruce Toal, CEO and co-founder of Convey Computer. “Our hybrid-core solution combines a powerful memory subsystem, which is ideal for massive data analytics, and a graph-friendly architecture capable of managing multi-terabyte graphs with billions of nodes.”
In this video, Shai Fultheim from ScaleMP discusses the company’s new Foundation 4.0 software, which allows customers to aggregate servers into a single system image.
CPU speeds have advanced much more quickly than other server components. As a result, HPC systems today tend to be unbalanced – with very fast CPUs, but not enough memory to feed them efficiently,” said Dan Olds, principal analyst at Gabriel Consulting Group. “Customers end up having to purchase more systems to get the performance they need for memory-hungry applications, leading to higher costs and more management complexity. vSMP Foundation 4.0 gives customers the ability to truly optimize their systems to fit these applications – reducing costs and easing management at the same time.”
In this video, Gilad Shainer from Mellanox discusses the impact FDR InfiniBand is making on the HPC and TOP500 landscape. He also describes new virtualized GPU technology enabled by IB’s low latency as well as new MPI communication libraries including ScalableSHMEM 2.0 and ScalableUPC 2.0.
Mellanox InfiniBand interconnect solutions continue to demonstrate leading performance, efficiency, scalability and reliability for the world’s fastest supercomputer systems,” said Eyal Waldman, president, chairman and CEO of Mellanox Technologies. “With over half of the world’s Petaflop systems, as well as the top three most efficient systems on the list, Mellanox FDR 56Gb/s InfiniBand and 10/40GbE interconnect solutions with PCI Express 3.0 provide the best return-on-investment with leading system efficiency without sacrificing performance.”
In this video, Dr. Claudio Arlandini describes how the CILEA consortium in Italy uses PBS Pro software from Altair to manage its supercomputing workloads. Recorded at SC11 in Seattle.
PBS Professional is powerful and flexible when it comes to scheduling policies, and this has definitely improved in its latest versions,” says Arlandini. “It allows us to keep our policies working at maximum levels. All our users are very happy with their share of our compute resources, but without PBS Professional it would be difficult to keep them that way. Our user community is very heterogeneous. Some are expert, some are not, and they come from all kinds of scientific areas.”
In this video, Oz Rentas from Terascala demonstrates the company’s TISIS 3.0 software, a “wrapper” that makes Lustre easier to manage. Recorded at SC11 in Seattle.
The TISIS architecture is based on a unique approach to Fast Data storage optimization, which combines a number of typically separate data sources. With the ability to monitor parallel file system operation, underlying storage platform performance, and client access, TISIS is able to monitor, learn, and predict the characteristics of Fast Data compute jobs to optimize execution and resource allocation.
In this video, Solarflare President and CEO Russel Stern describes the company’s kernel bypass technology designed to deliver exremely low latency for extreme application performance. Solarflare has enjoyed great success in financial markets at places like the New York Stock Exchange and is now taking their solutions to the general HPC marketplace.
Solarflare offers two families of high-performance, low-latency 10GbE server adapters. The Enterprise server adapters provide unmatched performance and scalability. The Midrange server adapters combine excellent performance across a wide range of application environments with great value. The company’s OpenOnload application acceleration middleware offers a unique combination of ultra-low latency, high-message rates, and full application compatibility for performance improvement. Solarflare also offers SFP+ optical transceivers tested and certified with our 10GbE server adapters.
This video was recorded at SC11 in Seattle. Read the Full Story.
The core of the HPCC Systems is the Enterprise Control Language (ECL), which is a declarative, data-centric programming language optimized for large-scale data management and query processing, which automatically manages workload distribution across all nodes. The programming language manages all aspects of the massive data joins, sorts and builds.
In this video, Clemson’s Dr. Walt Ligon provides an overview of OrangeFS, an open source file system tailor-made for Big Data.
OrangeFS is an open source parallel file system developed from PVFS. OrangeFS provides good scalability for large HPC systems and can be used with emerging data intensive workloads. Perhaps more importantly, OrangeFS is very flexible and very easy to install, use, and maintain because it is a user level file system and does not involve any changes to the Linux kernel. Results from a new whitepaper are presented as well. We finish the presentation with a brief description of new directions the OrangeFS team is beginning to develop.
What makes OrangeFS interesting to me is that it runs in User Space, not the kernel. Recorded at SC11 in Seattle.