April 28, 2011
The Weekly Top Five features the five biggest HPC stories of the week, condensed for your reading pleasure. This week, we cover the TeraGrid effort to support the Japanese research community; NNSA's 'Supercomputing Week' coverage; Mellanox's new double-duty switch silicon; Platform's latest Symphony; and the Oracle Sun Server-based Sandia Red Sky/Red Mesa supercomputer upgrades.
TeraGrid Research Community Reaches Out to Japan
In response to the multiple disasters that have hit Japan, starting with March 11, 2011, earthquake and tsunami, the world has responded with an outpouring of support. An article from the National Science Foundation's (NSF) TeraGrid cites the importance of reaching out to Japan's science and research sector:
"In addition to the humanitarian crisis, Japan's industrial and research communities have been affected which has global impact. Many of the products we use in our daily lives come from factories that were destroyed. Japan's intellectual contribution to the global research community has been interrupted as the systems many relied on to do their work were demolished. Much collaboration between Japanese and U.S. research groups across all domains of science has ground to a halt."
Here are some of the ways that TeraGrid institutions are helping:
TeraGrid Program Director Barry Schneider commented on the research partnerships:
"This isn't the first time our TeraGrid family took the initiative to help in a crisis. Hopefully their efforts will help Japanese researchers return to some sense of normality, allow the world to gain a better understanding of earthquakes and tsunamis in general, and prevent future loss. It's a great example of how the U.S. investment in science contributes to global scientific, social, and economic progress."
NNSA Celebrates 'Supercomputing Week'
This week the National Nuclear Security Administration (NNSA), the agency tasked with maintaining the safety, security and effectiveness of the US nuclear stockpile without underground testing, kicked off its "Supercomputing Week" celebration with a week of features devoted to the science and technology work done by the Advanced Simulation and Computing (ASC) program. Day one ushered in a new and improved ASC webpage, and spotlighted the New Mexico Supercomputing Challenge, which was held at Los Alamos National Laboratory (LANL), April 25-26.
Coverage from other days included special video features. One video provides an overview of the ASC program complete with summaries and photos of supercomputers housed at NNSA's national labs. A second video examines the hardware, software, codes and data that go into the NNSA supercomputers, while a third looks at the role that NNSA plays in non-nuclear research, from space debris to medical science and climate change.
As part of its broader research mandate, a team of computational physics and engineering experts from NNSA and Lawrence Livermore National Laboratory have been using NNSA supercomputers to study space debris. The project relies on modeling and detection tools to help to prevent a space disaster. The same technology, which uses sensor operations and data anlysis techniques, also shows potential for strenthening nuclear security.
The NNSA is also doing its part to promote energy-efficient systems. Two computers in the top 25 of the Green500 list are housed at Los Alamos National Laboratory and six supercomputers housed at NNSA sites appear on the list. In addition, NNSA technology went into creating the number-one system, the Blue Gene supercomptuer at the IBM Thomas J. Watson Research Center in New York.
Don Cook, NNSA's Deputy Administrator for Defense Programs, stated:
"Energy efficiency is vital to our enterprise as we invest in the future by building a modern infrastructure that is smaller, safer, more cost effective and more programmatically effective. The work done by our supercomputer engineers and scientists is a reflection of their commitment in seeking ways to better do business while promoting energy awareness."
Mellanox Debuts Multi-Protocol Switch Silicon
In interconnect news, Mellanox has announced the availability of its multi-protocol ASIC, SwitchX. This new technology will support up to 36-ports of FDR 56Gb/s InfiniBand or 40 Gigabit Ethernet. SwitchX, as the name implies, can switch hit, performing either as an InfiniBand switch or Ethernet switch (including Data Center Bridging - DCB). It's also low-latency and low power and provides best-in-class switching capabilities in each category, according to the company.
The switch will likely be a match for a number of applications/markets, including high-performance computing, cloud computing, financial services, database, Web 2.0, Internet and virtualized datacenter. An official company statement expounds on this:
SwitchX is an ideal choice for OEMs that must address end-user requirements for faster and more robust applications, and for datacenters and cloud computing environments that need to adapt their infrastructure to accommodate a variety of applications. SwitchX's reduced power, footprint and fully integrated PHY capabilities provide network architects critical feature-sets to enable fabric-flexible server and storage systems to meet the increasing performance demands of their customers.
What to hear more about the SwitchX chip's multi-purpose design? Here's what John Monson, vice president of marketing at Mellanox, told Register reporter, Timothy Prickett Morgan:
"This is the first time someone has built a protocol-independent ASIC, as far as we know. The funny thing [is that] there isn't much difference between the two, Ethernet and InfiniBand, inside the chip. But don't get the wrong idea. When SwitchX is running as an Ethernet switch, it's not running in an emulated mode or in a wrapper. It is an Ethernet switch. And when it is running as InfiniBand, it is similarly not emulated or wrapped."
Platform Releases Symphony 5.1
Platform Computing has launched the latest version of its SOA management software, Platform Symphony, version 5.1. New features highlighted by the company include easier application on-boarding, GPU support, and higher scalability and performance limits. With the expanded scalability, the platform is capable of managing 10,000 services per application, 40,000 cores per cluster (5,000 hosts) and 300 applications per cluster.
According to company officials, Platform has increased its presence in the financial services market over the past year. The release notes several significant multi-million dollar deals in both the capital markets and insurance sectors, and cites competitive wins from a leading global bond fund manager, a private equity fund manager and a prominent pension fund. Platform has experienced a strong demand for its cloud management technology, Platform ISF, and sales have also increased for Platform's Symphony product line. Platform's customer base includes Wall Street's largest banks and trading firms, including 12 of the Bloomberg 20, and a number of marquee North American and European brand accounts, many of which rank in the Fortune 500. All in all, global demand for all Platform's offerings -- including cloud, cluster, grid and MapReduce solutions -- has doubled year over year.
Peter Nichol, general manager of Platform Computing, remarks:
"As organizations look to streamline operations and benefit from the latest datacenter and analytics tools, we're experiencing a healthy global demand for Platform's technologies across the board -- from MapReduce programming models to cloud management software to our flagship cluster and grid technologies. This increased interest is a direct result of the concern over data explosion, a lack of scalable environments, insufficient capacity to run complex analytics in real-time and enterprise-wide contention for computing resources, particularly among financial services firms. Platform is well-positioned to provide firms with the efficiency and dynamic compute power needed to operate in today's volatile market conditions, and our growth over the past year proves that the time for high performance datacenters has come."
Oracle Sun Server Lights Up Red Sky, Red Mesa Supercomputers
Oracle announced that its Sun Blade Servers were selected by Sandia National Laboratories to enhance the performance of the lab's Red Sky / Red Mesa supercomputers. With the new Sun blade servers, Sun storage and InfiniBand networking technologies in its supercomputer datacenter architecture, Sandia has been able to significantly decrease power, energy usage and costs, while boosting computational capacity. Specifically, the organization has seen a tenfold increase in computing power, while reducing energy consumption and cost by 77 percent.
With Oracle's Sun servers -- using Sun Blade Modular Systems, the Sun Cooling Door System and Oracle's network fabric -- the supercomputers now achieve a peak performance of more than 500 trillion mathematical operations per second. The Sun Cooling Door cooling systems removed 90 percent of the heat load on servers, while the overall design changes will save 5 million gallons of water per year.
Sandia National Laboratories uses the Red Sky / Red Mesa system for a variety of complex computational projects, including biofuels research and molecular dynamics modeling and simulation.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?