In this video, Richard Graham from Oak Ridge National Laboratory presents: Advancing Research and Science through Supercomputing. Recorded at the HPC Advisory Council Israel Supercomputing Conference on Feb 7, 2012 in Tel Aviv.
Presentations will soon be available from the conference site.
In this slidecast, Mike Smith from Solarflare describes the company’s ApplicationOnload Engine (AOE), a new platform that moves application processing into the network adapter for applications that rely on real-time, high-performance network data.
Our new ApplicationOnload Engine is a new class of product that results directly from interaction with our end-user customers. Our engineers have worked closely with these customers to create a platform that leverages OpenOnload’s proven framework for creating a direct path from applications to the network, and incorporates on-the-fly processing of real-time network data,” said Russell Stern, CEO at Solarflare. “This solution provides not only the lowest latency and highest message rate network I/O performance, but achieves an unparalleled boost in application performance, all while maintaining a seamless, compatible interface with our existing server adapter products.”
Solarflare’s AOE combines a fully featured 10GbE server adapter with a state-of-the-art FPGA that provides a seamless, low-latency network interface to the host server and application processing. According to Smith, AOE is an open platform that utilizes applications developed by Solarflare, its customers, and third-party developers.
Read the Full Story * Download the MP3 * Subscribe on iTunes * If Dropbox is blocked, download from this Google page.
The XSEDE12 conference has issued its Call for Participation. The event will take place July 16-20 in Chicago.
The inaugural XSEDE12 conference will carry forward many of the successful elements of the TeraGrid conference series, while adding new features. XSEDE will showcase the discoveries, innovations, and achievements of those who use XSEDE resources and those who help build and support them. XSEDE12 also will create a forum for discussion of current needs and future plans among researchers, students, XSEDE staff, and NSF representatives.
Tutorial and Panel proposals are due April 13 and Paper and Poster submissions are due April 25.
The HPC Advisory Council has published the agenda for the Israel Supercomputing Conference coming up on February 7 in Tel Aviv. Featuring speakers from AMD, Intel, IBM, NetOptics, Mellanox, ORNL, ScaleMP, and Tel Aviv University, the one-day event will cover advanced HPC topics from around the world.
I’m looking forward to attending this event. The HPC Advisory Council recently reached a milestone of over 300 institutional members.
Thanks to its long history of high-tech breakthroughs by innovative small companies, Israel is often referred to as “Startup Nation.” What better place could there be to launch the 2012 HPC Event Season?
By Timothy Prickett Morgan • Get more from this author
The high-performance networking market just got a whole lot more interesting, with Intel shelling out $125m to acquire the InfiniBand switch and adapter product lines from upstart QLogic.
Intel has made no secret that it wants to bolster its Data Center and Connected Systems business by getting network equipment providers to use Xeon processors inside of their networking gear – that Intel division posted $10.1bn in revenues in 2011, and the company wants to break $20bn in the next five years.
The plan is to kill off mainframes and RISC machines, and to get Xeons inside of storage and network gear – but it also includes Intel being a major supplier of chips used in high speed switches.
Last July, Intel paid an undisclosed amount to get its hands on Fulcrum Microsystems, a maker of the FocalPoint family of ASICs for Ethernet switches and routers that run at 10GbE and 40GbE speeds. Fulcrum’s most famous customer was Arista Networks, the low-latency networking switch-maker founded by Sun Microsystems cofounder Andy Bechtolsheim. Intel never said what it paid for Fulcrum, but the company had raised $102m in venture capital since it was founded, and the price was very likely a multiple of that figure.
Despite the improvements in 10GbE and 40GbE switch chips over the past several years, InfiniBand still has important niches where even lower latency and still higher bandwidth are crucial – the supercomputing racket, for instance, or in database clustering. Just ask Oracle, which uses InfiniBand silicon from Mellanox Technologies in its Exadata database clusters and Exalogic web application server clusters, and which took a 10.2 per cent stake in the chip and switch-maker back in October 2010.
At the time, Mellanox assured Wall Street that Oracle had no intention of taking over the chipmaker, but with QLogic’s upstart InfiniBand biz snapped up by Intel, some systems or networking companies might now be tempted to take a run at Mellanox. But if Oracle or IBM or Cisco Systems are tempted to eat Mellanox, all that will do is eventually drive everyone into the loving arms of Intel, with its own Ethernet or InfiniBand ASICs. So, in a funny way, Intel is probably praying that someone does eat Mellanox.
And the funniest thing of all would be if AMD actually woke up and smelled the systems biz, and did it. By doing so, AMD would have the SwitchX two-timing Ethernet and InfiniBand ASICs and the ConnectX-3 switch-hitting server adapters, and could start integrating these deeper into its chipsets and eventually onto its chips.
InfiniBand has its roots in the Next Generation I/O project supported by Intel, Sun Microsystems, and Microsoft, along with the Future I/O alternative supported by IBM, Compaq, and HP. These specs were merged back in 1999, with Intel and IBM largely steering the process.
The idea was to provide a single switched fabric that would link computers and storage to each other from the desktop to the data center, and be an alternative to Ethernet networks for server-to-server and PC-to-server links, and to PCI-Express and Fibre Channel for linking peripherals.
Academically, InfiniBand was probably the right answer for a unified switch fabric – but markets don’t study in schools, they live on the mean streets and give and take hard knocks. And thus, InfiniBand has been relegated to a niche and, more importantly, the key technologies that made InfiniBand better, stronger, and faster than Ethernet have been borged onto Ethernet, closing the gap.
For now, Intel is saying that its acquisition of the InfiniBand chip, adapter, and switch business from QLogic is all about HPC, but it may be looking further down the road, when PCI-Express runs out of gas.
“At the International Supercomputing Conference 2011, Intel unveiled a bold vision to redefine HPC performance and break the exascale barrier by 2018,” said Kirk Skaugen, the outgoing general manager of Intel’s Data Center and Connected System Group, said in a statement. “The technology and expertise from QLogic provide important assets to provide the scalable system fabric needed to execute on this vision. Adding QLogic’s InfiniBand product line to our networking portfolio will bring increased options and exceptional value to our datacenter customers.”
Last week, Skaugen – who has been pushing Intel’s expansion into switching and storage chippery for the past several years – was tapped to run Chipzilla’s PC Client Group. Diane Bryant, who has worked for Skaugen in the past and who was most recently Intel’s CIO, has replaced Skaugen and will be driving Intel’s server, storage, and networking strategies.
By selling its InfiniBand biz to Intel, QLogic will be able to double down on its Fibre Channel and Ethernet switches and adapters. QLogic has had some success with its InfiniBand gear, landing the 2,000-node “Sierra” cluster with Dell at Lawrence Livermore National Labs and also being the switch supplier for the 20,000-node procurement awarded to Appro International last June by the US Department of Energy’s Tri-Labs: Lawrence Livermore, Los Alamos, and Sandia National Laboratories.
“The sale of these InfiniBand assets will benefit our shareholders by enabling us to provide better focus and greater investment in growth opportunities for the data center with our converged networking, enterprise Ethernet, and storage area networking products,” said QLogic’s president and CEO, Simon Biddiscombe, in his statement. “After the sale, our cash position will be further strengthened and we expect the impact on earnings per share to be neutral. In addition, the sale of these assets to a leading technology innovator and recognized HPC leader will provide a greater investment stream in high performance fabrics for InfiniBand partners and customers.”
Speaking to El Reg two weeks ago apropos of nothing about the InfiniBand racket, QLogic’s head of global alliances and solutions marketing for HPC Joe Yaworksi said that the reason why QLogic was winning more InfiniBand deals is that its TruScale chips offer better performance running at Quad Data Rate (QDR) 40Gb/sec speeds than do Mellanox’ SwitchX products running at Fourteen Data Rate (FDR) 56Gb/sec speeds.
The big reason for this, said Yaworksi, was that QLogic bought compiler-maker PathScale in early 2006, and it has a networking stack that was designed to handle millions of messages per second. (PathScale was sold to SciCortex in 2007, and when it went bust, Cray picked up the PathScale pieces in 2009 and an open source PathScale has emerged from the ashes with a license from Cray.) The combination of the TruScale InfiniBand ASICs and PathScale messaging stack and compilers is what gave QLogic the idea it could take on Mellanox and win.
Yaworksi told El Reg that QLogic was “taking a hard look at whether or not we will ship FDR InfiniBand,” although with Intel picking up the company, there will be more funds to do whatever might seem appropriate. The company was thinking that in the second half of 2013 or the first half of 2014 it might jump straight to Eight Data Rate (EDR) speeds, which runs the InfiniBand lanes at 25Gb/sec.
That would be a long time to wait between products and to live on QDR, and a gap that Intel is probably not likely to tolerate. But it all depends on what Intel’s plans are, and the company isn’t saying anything right now. If QLogic weren’t a public company, both would have probably said less.
Intel expects the QLogic InfiniBand deal to close by the end of March, and added that a “significant number” of the employees associated with the business were expected to accept job offers from Chipzilla. ®
This article originally appeared in The Register. It appears here in its entirety as part of a cross-publishing agreement.
Luxtera is looking to spur a new ecosystem by opening up its unique Silicon CMOS Process, which integrates photonics with transistor-based electronics. Today the company announced it will open its CMOS Photonics device library to Optoelectronic Systems Integration in Silicon (OpSIS), a foundry service that “provides access to optoelectronic integrated circuits to the community at large, at a modest cost.”
We are thrilled to be able to offer our community access to Luxtera’s unique process. It provides the opportunity to leverage the significant investment and maturation of the world’s first production proven CMOS Photonics design flow,” said Michael Hochberg, director of OpSIS and associate professor at the University of Delaware. “We believe that this will significantly accelerate the growth of the Silicon Photonics ecosystem. I’m particularly excited that this process will offer both academic and industrial users a chance to leverage a full electronics PDK as well as yield models for the key photonic components in order to accurately predict the performance and yield of complex systems-on-a-chip. I see this as a major step forward for the field as historically much of the innovation has been centered on process development. We’re now moving into an era where Silicon Photonics can enable a great deal of innovation at the system and architectural level.”
Luxtera CEO Greg Young said here is significantly more market opportunity than the company can service directly: “In working with OpSIS we are able to advance the wide scale impact of our Silicon Photonic offering as well as push the envelope for future commercialization.”
Read the Full Story.
In a move to prime their Ethernet products for the upcoming Intel Romley/Sandy Bridge chips with PCI Express 3.0, Mellanox today announced a new pricing structure and expanded channel for their end-to-end 10 Gigabit Ethernet solutions. With newly announced prices as low as $188 per port, Mellanox is looking to ride “significant movement towards 10GbE” that analysts are predicting for 2012.
The launch of the Intel Romley and Sandy Bridge platform in the coming months is going to give the industry a strong push towards faster interconnect speeds,” said Eyal Waldman, chairman, president and CEO of Mellanox Technologies. “For data center managers, 10GbE is the minimum data rate they should be considering for all deployments this year. Mellanox products and solutions are delivering 10GbE and higher data rates today – providing better application performance, lower power consumption and improved ROI.”
Read the Full Story or sign up for the Mellanox Product Webinar with Crehan Research on January 30 from 8:00am-9:00am PST.
A new whitepaper from Mellanox describes the HPC application performance advantages of InfiniBand vs. 10 Gig Ethernet. Written by Todd Wilde, Gilad Shainer, Pak Lui, and Tong Liu, the paper also makes the case that IB offers significant price/performance advantages as well.
Due to the superior end to end latency, bandwidth, scalability and advanced features of InfiniBand, it remains the clear choice for most high performance computing systems and applications. This paper explains why many applications will see advantages when using InfiniBand and shows real world results using a variety of applications including ANSYS FLUENT and Schlumberger ECLIPSE. Moreover, when performing a price/performance analysis of Mellanox QDR InfiniBand over Cisco Nexus 10GbE switching, the use of InfiniBand over 10GbE for HPC becomes even more compelling.
As you may recall, we did a post over the Holidays about a Cisco whitepaper that drew very difference conclusions. Your mileage may vary, but we always encourage comments and insights.
Download the whiitepaper (PDF). If your IT crowd blocks Dropbox, you can download from this Google page.
In this slidecast, Jay Botelho from Wildpackets describes how network forensics works and how the company’s provides enterprise-wide network monitoring and reporting.
Download the MP3 * Subscribe on iTunes * Subscribe on other podcast players
If your IT crowd blocks Dropbox, you can also Download the MP3 from The Google.
This week Mellanox announced the newest version of its VMA 6.0 Messaging Accelerator, which includes enhanced TCP and UDP acceleration support over ConnectX-3 VPI adapter cards. With ultra-low UDP latency under 1.4 microseconds and TCP socket latency under 1.7 microseconds, this messaging technology is reportedly more than two times faster than competitive offerings.
I caught up Mellanox’s Gilad Shainer to talk about how VMA 6.0 accelerates applications like high frequency trading.
insideHPC: How does VMA 6.0 accelerate performance in socket-based applications? Does it employ some form of kernel bypass for UDP?
Gilad Shainer: VMA bypasses the kernel by moving TCP and UDP to the user space. With that it can reduce the overheads and accelerate the latency. VMA also takes advantage of the world lowest latency NIC – ConnectX-3 to provide the lowest latency in the market.
insideHPC: How will socket-based applications like those used for high-frequency trading benefit from VMA 6.0?
Gilad Shainer: Low latency is the major item here, as well as the lowest jitter – VMA guarantees the same low latency regardless of how many processes run on the server
insideHPC: Is VMA 6.0 strictly a software solution for delivering low latency?
Gilad Shainer: VMA is a software solution that takes advantage of Mellanox adapters
insideHPC: VMA 6.0 works with ConnectX-3 adapter cards, which support both Infiniband and 10 GigE. Does the acceleration of VMA 6.0 only work
on the IB side?
Gilad Shainer: The acceleration works on both IB and Ethernet
insideHPC: How important is low-latency in this age of ever-increasing cores?
Gilad Shainer: Low latency is critical – more cores means more processes that need more from the interconnect side. The requirements are not only for low latency, but with more cores – to guarantee the same low latency regardless of how many cores/processes exist. This is a tough requirement, and other vendors are failing to deliver the desired solution.
This week QLogic announced that its InfiniBand technologies are powering the world’s fastest cluster powered by Intel’s “Sandy Bridge” Xeon E5 processors. Ranked at #15 on the TOP500, the “Zin” supercomputer at LLNL comprises 46,208 cores in 2916 nodes, producing up to 837 megaflops per watt.
The ever more powerful computing systems Lawrence Livermore requires to fulfill its national security missions must be balanced with increasing energy efficiency,” said Matt Leininger of LLNL’s Advanced Simulation and Computing program. “To meet its scientific computing demands, the Laboratory works with industry leaders to advance HPC.”
I talked to QLogic’s Joe Yaworski about what makes the Zin system so remarkable. He said that the fact that this Top15 cluster achieves nearly a Petaflop with less than 3000 nodes is a testament to the efficiency of the Sandy Bridge processor and QLogic’s TrueScale interconnect.
Yaworski also noted that the Scalable Unit architecture developed by the Tri-Labs enabled the cluster to be deployed very rapidly at the end of October. Using Qlogic’s TrueScale InfiniBand software, LLNL was able to install and configure the cluster in record time and complete the TOP500 benchmark.
Read the Full Story.
This week the HPC Advisory Council announced that will host the HPC Advisory Council Israel Supercomputing Conference in Tel Aviv on February 7, 2012. The conference is open open to the public and will focus on HPC usage models, benefits for small to large businesses, and the future of supercomputing in Israel.
Continuing our successful international conference programs, the HPC Advisory Council Israel Supercomputing Conference is expected to uphold our tradition of providing rich educational content, taught from some of the industry’s most recognized luminaries and scientists,” said Gilad Shainer, chairman of the HPC Advisory Council. “High-performance computing is critical for enhancing science, research and product development. Our conference attendees will gain the knowledge on how to best leverage high-performance computing for such activities and how to better leverage HPC system capabilities and improve productivity and efficiency.”
Read the Full Story.
What will the high speed networks of the future look like? This week RENCI at the University of North Carolina, Chapel Hill and Duke University announced a collaboration with IBM to build a nationwide test bed for next-gen networks. As part of the National Science Foundation’s GENI initiative, the project will deploy and operate 13 ExoGENI sites at research universities and labs across the U.S.
Future computer science and applied research must bring together computation, storage and network capabilities on a global scale to address emerging complex problems related to network science, large-scale distributed computations, large dataset mobility and future network architectures,” said Baldine. “With ExoGENI researchers will gain a global, elastic reconfigurable platform to conduct such research.”
Read the Full Story.
In this video, Sharon Besser of NetOptics presents: Leveraging Access Switching For HPC Monitoring. Recorded at the HPC Advisory Council Stanford Workshop. Download the Slides (PDF).
In this video, Todd Wilde from Mellanox presents: Workshop Tutorial - InfiniBand Technology. Recorded at the HPC Advisory Council Stanford Workshop.
insideHPC.com is a production of insideHPC, LLC. © 2006-2011 Sitemap
