In conjunction with the Supercomputing '10 conference under way in New Orleans several vendors have made product announcements or demonstrated their prowess in the supercomputing industry.
Intel powers TIanhe-1A
Intel announced that its Intel Xeon 5600 series processor is at the heart of the world's most powerful supercomputer, the Tianhe-1A. The supercomputer in Tianjin, China came in at 2.57 petaflops, powered by 14,396 Intel processors accompanied by NVIDIA GPU accelerator cards. “Our Xeon processor roadmap continues to deliver hugely powerful supercomputers that are helping solve mankind’s greatest challenges,” said Rajeeb Hazra, general manager of Intel’s High Performance Computing organization. “Securing the t op position on the Top 500 is a source of great pride for Intel, and demonstrates the tremendous leaps in performance and versatility that our processors are delivering across a range of compelling workloads.”
During the SC10 conference Intel has been conducting demonstrations of its recently announced Intel MIC (Many Integrated Core) architecutre, as well as showcasing compressed medical imaging developed with May Clinic on "Knights Ferry," the first Intel MIC design and development kits.
The Exascale Report interviews Justin Rattner, Intel Vice President and CTO on his perspectives for where the microprocessor industry is headed.
SGI introduces Jolt technology
At the SC10 Conference SGI announced the introduction of Jolt technology that will deliver vast improvements in speed, performance and scalability to their Altix UV. Jolt technology allows the Altix UV to scale from 2,048 to over 250,000 cores, from 16TB to 8PB of memory and from 4 to 512 cabinets. Stephen Hawking and Cambridge and the COSMOS Consortium selected the SGI Altix UV as their compute platform to unlock the mysteries of the origins of the universe.
"SGI's customers are among the largest and most innovative HPC facilities in the world, and they are embracing Altix UV and the breakthrough performance of shared-memory supercomputing. New doors are opening to Altix UV in commercial computing markets, further expanding the definition of supercomputing and applications it serves," said Tony Carrozza, senior vice president of worldwide sales at SGI.
SGI also announced Tuesday that it has received the HPCWire Editor's Choice Award for the "Best HPC Server or Technology." "The HPCwire Readers' and Editors' Choice Awards recognize companies that have made the most significant impact in the global HPC market," said Thomas Tabor, publisher of HPCwire. "I would like to congratulate SGI for its expertise and innovation in the HPC community, and for its technical computing leadership. These awards are a clear indication of how our readers and editors, among the most informed in the industry, view SGI's Altix UV."
Platform Computing debuts LSF8
At the SC10 conference Platform Computing announced the availability of the latest version of its flagship product family, Platform LSF, a workload scheduling solution for high performance computing (HPC). Version 8 of the software includes the ability to delegate administrative rights to line of businesss managers, live dynamic cluster reconfiguration, guaranteed resources to ensure service level agreements are met, flexible fairshare scheduling policies, and unparalleled scalability to support the large clusters in use today. Platform LSF will be available in January 2011.
“With Platform LSF we can intelligently schedule workload, ensuring that we get the maximum return on our significant investment in simulation tools and hardware,” said Nathan Sykes, Head of CFD, Red Bull Racing. “We partnered with Platform Computing to deploy a complete software management solution and, as a result, benefitted from a 20% improvement in the throughput of our engineering analysis.
Platform Computing also announced that it is making HPC in cloud environments a reality for HPC data centers through three flexible and powerful cloud bursting solutions. The first solution is an integrated cluster with the Cloud (Platofrm LSF) that pairs LSF's dynamic host capabilities with a cloud service such as the Amazon Virtual Private Cloud. The second solution is a multi-cluster to the Cloud (Platform LSF with Platform MultiCluster) and allows users to start up a new cluster in any cloud or hosting provider environment. The third solution is a dynamic cluster extension to the cloud.
Juniper powers 100G demonstration
Juniper Networks announced the industry's first demonstration of a standards-based 100G core-to-edge network, powered by the newly introduced Juniper Networks T4000 Core Router and the MX3D Series Universal Edge Router. "Internet2 is creating a high-capacity network that will enable advanced networking features for more than 200,000 essential community anchor institutions, and high-capacity bandwidth is a requirement for successfully building and running that network," said Rob Vietzke, Internet2 executive director of network services. "With next-generation 100G platforms, we will support advanced research and education applications that are not possible today, but will impact our economy tomorrow."
The T4000 will be available in the second half of 2011. The MX3D Series Universal Edge Routers are generally available today with 100Gigabit Ethernet interfaces available in 2011.
Force 10 demonstrates 40 Gigabit Ethernet
Force 10 Networks is at the SC10 conference and will exhibit 10/40 Gigabit Ethernet (GbE) solutions. Arpit Joshipura, chief marketing officer at Force 10 explained that “by demonstrating line-rate 40 GbE solutions in a collaborative environment, we anticipate that SC10 attendees will walk away clearly understanding that 40 GbE is market-ready to solve today’s performance needs while helping to curb growing data center complexity.” To show how 40 Gigabit Ethernet is already market-ready the Force 10 S-Series 4810 top-of-rack switch will be showcased as part of a comprehensive, multi-vendor multi-technology display in the Ethernet Alliance (EA) booth.
Extreme Networks 40 Gigabit Ethernet
Extreme Networks (EXTR) announced that it is working with NASA Goddard Space Flight Center (GSFC) at the SC10 conference Tuesday in evaluating high speed, standards-based 40 GbE connectivity in support of scientific research that studies the global climate. "Our High End Computer Networking (HECN) Team at NASA Goddard Space Flight Center is looking forward to using Extreme Networks' 40 GbE modules in our environment. We've been link aggregating multiple 10 GbEs for quite some time, and prefer to have 40 GbE, especially given Extreme Networks' announced price," said J. Patrick (Pat) Gary, project manager for GSFC's HECN Team. "The demonstrations at SC/10, involving the transfer of very large files, can help propel climate research further."