insideHPC » Computing Research

Researchers Deploy GPUs To Build World’s Largest Artificial Neural Network

Ralph — Tue, 18 Jun 2013 07:27:19 +0000

Today at ISC’13, Nvidia announced a collaboration with Stanford University to create the world’s largest artificial neural network built to model how the human brain learns. At 6.5 times bigger than the previous record-setting network developed by Google in 2012, the neural net will be capable of “learning” how to model the behavior of the brain — including recognizing objects, characters, voices and audio in the same way that humans do.

Creating large-scale neural networks is extremely computationally expensive. For example, Google used approximately 1,000 CPU-based servers, or 16,000 CPU cores, to develop its neural network, which taught itself to recognize cats in a series of YouTube videos. The network included 1.7 billion parameters, the virtual representation of connections between neurons. In contrast, the Stanford team, led by Andrew Ng, director of the university’s Artificial Intelligence Lab, created an equally large network with only three servers using Nvidia GPUs to accelerate the processing of the big data generated by the network. With 16 NVIDIA GPU-accelerated servers, the team then created an 11.2 billion-parameter neural network.

Delivering significantly higher levels of computational performance than CPUs, GPU accelerators bring large-scale neural network modeling to the masses,” said Sumit Gupta, general manager of the Tesla Accelerated Computing Business Unit at NVIDIA. “Any researcher or company can now use machine learning to solve all kinds of real-life problems with just a few GPU-accelerated servers.”

Read the Full Story.

The post Researchers Deploy GPUs To Build World’s Largest Artificial Neural Network appeared first on insideHPC.

Gordon Supercomputer Fosters New Fields of Research for HPC

Ralph — Sun, 16 Jun 2013 15:04:20 +0000

With over 300 TB of high performance Intel flash memory SSD, the Gordon supercomputer at SDSC is opening the door to new areas of research for HPC such as political science, mathematical anthropology, finance, and even the cinematic arts.

Gordon’s extraordinary speed makes it possible for researchers to tackle questions they couldn’t address before, simply because they didn’t have a system that was uniquely tailored to handle the challenges of data intensive computing,” said SDSC Michael Norman just prior to Gordon’s launch. “I view Gordon as a new kind of vessel, a ship that will take us on new voyages to makes new discoveries in new areas of science.”

Read the Full Story.

The post Gordon Supercomputer Fosters New Fields of Research for HPC appeared first on insideHPC.

Hungarian Datacenter Inaugurated

Staff — Sun, 16 Jun 2013 13:47:42 +0000

CERN and the Wigner Research Centre for Physics have inaugurated the Hungarian data centre in Budapest, marking the completion of the facility hosting the extension for CERN computing resources.

About 500 servers, 20,000 computing cores, and 5.5 Petabytes of storage are already operational at the site. The dedicated and redundant 100 Gbit/s circuits connecting the two sites are functional since February 2013 and are among the first transnational links at this distance. The capacity at Wigner will be remotely managed from CERN, substantially extending the capabilities of the Worldwide LHC Computing Grid (WLCG) Tier-0 activities and bolstering CERN’s infrastructure business continuity.

WLCG’s mission is to provide global computing resources to store, distribute and analyse more than 25 Petabytes of data annually generated by the Large Hadron Collider (LHC). It is a global system organised in tiers, with the central hub being the Tier-0 at CERN.

The experiments” computing resources needs will increase significantly when the LHC restarts in 2015. Hosting computing equipment at the Wigner Centre to extend CERN’s data centre Tier-0 capabilities is essential for dealing with this expected increase, and to the success of our physics programme. The remote capacity will also contribute to business continuity for the critical systems in case of a major issue on CERN’s site,’ said CERN director-general Rolf Heuer. “A number of sciences currently face exponential data growth. This innovative approach with Wigner could point the way for research centres to run their services in the future.”

This story appears here as part of a cross-publishing agreement with Scientific Computing World.

The post Hungarian Datacenter Inaugurated appeared first on insideHPC.

STFC to Develop Hybrid Supercomputer Applications

Ralph — Fri, 14 Jun 2013 10:30:04 +0000

The Science and Technology Facilities Council (STFC) at Daresbury in the U.K. has entered an agreement with Nvidia to help develop software for next-generation hybrid supercomputers.

This agreement combines NVIDIA’s leading-edge GPU accelerator technologies and HPC expertise with STFC’s software development expertise,” said David Corney, acting director of STFC’s Department of Scientific Computing. “This unique combination will enable the development of next-generation massively parallel applications, which will be used for exascale performance levels, or a thousand times more powerful than Blue Joule at STFC, the most powerful computer in the UK today.”

The collaboration will offer scientists access to one of the largest software development laboratories in the world, STFC’s Hartree Centre at Daresbury Laboratory, which is dedicated to modelling and simulation software, as well as to Nvidia’s expertise. Read the Full Story.

The post STFC to Develop Hybrid Supercomputer Applications appeared first on insideHPC.

Pedraforca Cluster to be First to combine ARM CPUs, GPUs, and InfiniBand

Ralph — Wed, 12 Jun 2013 14:47:14 +0000

Today the Barcelona Supercomputing Center (BSC) announced plans to deploy its next-gen ARM prototype cluster for HPC in July. Powered by ARM Cortex-A9, Nvidia Tesla K20 GPUs, and Mellanox QDR InfiniBand, the hybrid supercomputer will be named Pedraforca.

By using InfiniBand, Pedraforca enables direct GPU-to-GPU communication through RDMA on ARM. It features a low-power Nvidia Tegra® 3 (4-core Cortex-A9) to run the operating system and drive both the Tesla K20 accelerator and the QDR InfiniBand at the minimum power consumption.

Prototypes are critical to accelerate software development, both system software and applications. Pedraforca introduces multiple innovations to the ARM software stack, leading to a more energy-efficient platform for those GPU-centric applications that match the characteristics of the cluster.” says Alex Ramirez, leader of the Heterogeneous Architectures Research Group at BSC.

BSC deployed the first ARM-based multicore HPC cluster in October 2011 with a cluster called Tibidabo. In November 2012, BSC collaborated with NVIDIA and SECO in the development of the KAYLA development platform, the first hybrid ARM + CUDA GPU platform, which was field-tested in the second BSC cluster. Pedraforca represents another step forward in the BSC research roadmap on new technologies and innovative architectures towards energy-efficient HPC.

Read the Full Story.

The post Pedraforca Cluster to be First to combine ARM CPUs, GPUs, and InfiniBand appeared first on insideHPC.

Video: PRACE Supercomputing Spurs Astrophysics

Rich Brueckner — Tue, 11 Jun 2013 18:25:44 +0000

In this video, Dr Turlough Downes from the Dublin Institute for Advanced Studies discusses how advanced astrophysics is made possible by supercomputing resources from the European PRACE infrastructure.

The video is part of a series of case studies PRACE has published on its new Dare to Think the Impossible site, which looks like a great resource for promoting the value of supercomputing.

For more information, visit the PRACE booth #725 at ISC’13.

The post Video: PRACE Supercomputing Spurs Astrophysics appeared first on insideHPC.

Simulations Show How Comet Impacts Delivered Building Blocks of Life

Ralph — Sun, 09 Jun 2013 16:07:30 +0000

Over at Science World Report, Mark Hoffman writes that scientists using supercomputers have found that the materials and chemical reactions from comet impacts brought an abundance of life building organic compounds to Earth. Before these events, early Earth was not very hospitable when it came to jump-starting life.

The flux of organic matter to Earth via comets and asteroids during periods of heavy bombardment may have been as high as 10 trillion kilograms per year, delivering up to several orders of magnitude greater mass of organics than what likely pre-existed on the planet,” said Nir Goldman from Lawrence Livermore, who conducted the research with Isaac Tamblyn from the University of Ontario Institute of Technology.

Goldman’s earlier work is based on computationally intensive models, which, in the past, could only capture 10-30 picoseconds of a comet impact event. However, with new, more computationally efficient models, the team was able to capture hundreds of picoseconds of the impacts, which is much closer to chemical equilibrium.

Read the Full Story.

The post Simulations Show How Comet Impacts Delivered Building Blocks of Life appeared first on insideHPC.

World’s First Intercontinental 100 Gbps Link for Research Demos Today

Ralph — Mon, 03 Jun 2013 15:31:04 +0000

Today Internet2 announced the world’s first Transatlantic 100 gigabits-per-second transmission link for research and education between North America and Europe. As demonstrated during this week’s TERENA Networking Conference in the Netherlands, the Advanced North Atlantic 100G Pilot project will be used for testing applications, resources, monitoring techniques, and advanced technologies such as software-defined networking.

This achievement shows that research and education networks are at the forefront of innovation, thereby empowering the most advanced research by universities and research institutions worldwide,” said Erwin Bleumink, chief executive officer of SURFnet – the local organizer of TNC2013. “The impact of this development however will also be seen outside academia and help stimulate the global economy.”

Read the Full Story.

The post World’s First Intercontinental 100 Gbps Link for Research Demos Today appeared first on insideHPC.

TACC’s Hadoop Cluster Makes Big Data Research More Accessible

Ralph — Mon, 03 Jun 2013 11:00:31 +0000

Over at the Texas Advanced Computing Center, Aaron Dubrow writes that researchers are using a specialized cluster at TACC to do experimental Hadoop-style studies on a current production system.

This system offers researchers a total of 48, eight-processor nodes on TACC’s Longhorn cluster to run Hadoop in a coordinated way with accompanying large-memory processors. A user on the system can request all 48 nodes for a maximum of 96 terabytes (TB) of distributed storage. What’s special about the Longhorn cluster at TACC isn’t simply the beefed-up hardware for running Hadoop; rather it’s the ability for researchers to leverage the vast compute capabilities of the center, including powerful visualization and data analysis systems, to further their investigations. The end-to-end research workflow enabled by TACC could not be done anywhere else, and as a bonus, researchers get access to the full suite of tools available at the center to do computational research.

According to TACC Research Associate Weijia Xu, the best part is that Hadoop is easy to use without requiring users to be experts. It handles a lot of the low-level computing behavior, so people don’t need to have a lot of knowledge about I/O or memory structures to get started.

Read the Full Story.

The post TACC’s Hadoop Cluster Makes Big Data Research More Accessible appeared first on insideHPC.

XSEDE Industry Challenge Program Seeks Proposals

Rich Brueckner — Sun, 02 Jun 2013 15:55:30 +0000

The eXtreme Science and Engineering Discovery Environment (XSEDE) is accepting proposals for computationally intensive challenges facing industry that could be addressed with access to XSEDE expertise and resources. Faculty and researchers from accredited universities are eligible, as well as industry researchers, and researchers from government laboratories.

Proposals must describe a computationally challenging problem hindering industry today while demonstrating how the cooperative efforts of academia, industry and XSEDE, and potentially government research laboratories, will provide a framework for solving the problem.

Two-page letters of intent are due July 19, 2013 with full proposals are due Sept. 13, 2013. Read the Full Story.

The post XSEDE Industry Challenge Program Seeks Proposals appeared first on insideHPC.

CUG Best Paper Describes Trillion-Particle Simulation on Hopper

Rich Brueckner — Fri, 31 May 2013 11:30:05 +0000

The Cray User Group wrapped up in Napa Valley earlier this month with a team from NERSC winning best paper: Trillion Particles, 120,000 cores, and 350 TBs: Lessons Learned From a Hero I/O Run on Hopper.

An unprecedented trillion-particle simulation, which utilized more than 120,000 processors and generated approximately 350 terabytes of data, pushed the performace capability of the National Energy Research Scientific Computing Center’s (NERSC’s) Cray XE6 “Hopper” supercomputer to its limits. In addition to shedding new light on a long-standing astrophysics mystery, the successful run also allowed a team of computational researchers from the Lawrence Berkeley National Laboratory (Berkeley Lab) and Cray Inc. to glean valuable insights that will help thousands of scientists worldwide make the most of current petascale systems like Hopper, which are capable of computing quadrillions of calculations per second, and future exascale supercomputers, which will compute quintillions of calculations per second.

Read the Full Story or Download the paper (PDF).

The post CUG Best Paper Describes Trillion-Particle Simulation on Hopper appeared first on insideHPC.

Podcast: American Science Struggles through Budget Cuts

Rich Brueckner — Thu, 30 May 2013 17:37:35 +0000

In this NPR podcast, Tom Ashbrook looks at how the sequester and budget cuts are affecting American science and research.

American leadership in science has been a given for most of the last century. About a third of science research and development in this country has been supported by the federal government. Funding for about 60 percent of basic research in science comes from Washington. We’ve celebrated the results, from moon shots to the Internet. Now Washington’s cutting back, and it’s hitting American science — just when competitor nations are plowing more cash into the science frontier.

Alexandra Witze, correspondent for Nature
Paul Alivisatos, director of the Lawrence Berkeley National Laboratory. He’s also a professor of chemistry and materials science at the University of California Berkeley
Ann Bonham, chief scientific officer at the Association of American Medical Colleges and professor of pharmacology and internal medicine at the University of California Davis
Barney Keller, communications director of the Club for Growth, a 501(c)(4) organization that promotes reining in government spending

The post Podcast: American Science Struggles through Budget Cuts appeared first on insideHPC.

Podcast: Radio Free HPC Looks at Holographic Storage

Rich Brueckner — Tue, 28 May 2013 20:12:21 +0000

In this podcast, the Radio Free HPC team takes a look at pending holographic storage technology.

Henry Newman’s head is still spinning from the technology previews he saw at the recent IEEE Mass Data Storage Conference. Could this be a game-changer for the tape industry and archiving data? What are the specs, limitations, and most importantly, are these the droids we’ve been looking for?

View the Hitachi Slides * View the Akonia Holographics Slides * Download the MP3 * Download the mobile video * Download 1024p Video * Subscribe on iTunes * RSS Feed

The post Podcast: Radio Free HPC Looks at Holographic Storage appeared first on insideHPC.

Video: India’s Supercomputing Vision – Key Technical Challenges

Rich Brueckner — Tue, 28 May 2013 14:55:14 +0000

In this video, Dr. GV Ramaraju presents: India’s Supercomputing Vision – Key Technical Challenges. Ramaraju describes what is expected from research and academic institutes and vendors to provide the direction and thrust to build the necessary computing systems ecosystem.

The post Video: India’s Supercomputing Vision – Key Technical Challenges appeared first on insideHPC.

Video: Science as Voyage – Ian Foster at TEDxCERN

Rich Brueckner — Mon, 27 May 2013 11:30:29 +0000

In this video from the TEDxCERN event, Ian Foster takes us on a journey of Big Process for Big Data, with a call to action for the kind of collaborative processes we need.

The post Video: Science as Voyage – Ian Foster at TEDxCERN appeared first on insideHPC.

Video: HPC Drives Life Science Research at TACC

Rich Brueckner — Sun, 26 May 2013 16:24:35 +0000

In this video, from Joe Maglitta from Go Parallel sits down with Oscar Jiao from TACC. Jiao is the center’s life science computing specialist, and he describes how researchers are using the nine-petaflop Stampede supercomputer to pursue new discoveries in their fields.

The post Video: HPC Drives Life Science Research at TACC appeared first on insideHPC.

Whitepaper: Creating a Raspberry Pi-Based Beowulf Cluster

Rich Brueckner — Sat, 25 May 2013 14:51:11 +0000

The Raspberry Pi is a credit-card-sized single-board computer developed in the UK by the Raspberry Pi Foundation with the intention of promoting the teaching of basic computer science in schools. But could this ARM-based device be used to teach supercomputing as well? Joshua Kiepert, a doctoral student at Boise State’s Electrical and Computer Engineering department, published a white paper entitled: Creating a Raspberry Pi-Based Beowulf Cluster.

Although an inexpensive bill of materials looks great on paper, cheaper parts come with their own set ofdownsides. Perhaps the biggest downside is that an RPi is no where near as powerful as a current x86 PC. The RPi has a single-core ARM1176 (ARMv6) processor, running at 700MHz (though overclockingis supported). Additionally, since the RPi uses an ARM processor, it has a different architecture than PCs, i.e. ARM vs x86. Thus, any MPI program created originally on x86 must be recompiled when deployed to the RPiCluster. Fortunately, this issue is not present for java, python, or perl programs. Finally, because of the limited processing capability, the RPiCluster will not support multiple users simultaneously using the system very well. As such, it would be necessary to create some kind of timesharing system for access if it ever needed to be used in such a capacity.

Download the paper (PDF).

The post Whitepaper: Creating a Raspberry Pi-Based Beowulf Cluster appeared first on insideHPC.

Podcast: Pete Beckman on the Rise of the Super Smart Supercomputer

Rich Brueckner — Fri, 24 May 2013 21:08:03 +0000

In this podcast from WBEZ in Chicago, Pete Beckman from Argonne explains how math and supercomputers are accelerating scientific discovery and helping us predict the future. From discovering the secret inner workings of the universe to developing cars that can drive themselves, technology and science are fueling a new breed of massive, smart supercomputers that will improve our world.

The post Podcast: Pete Beckman on the Rise of the Super Smart Supercomputer appeared first on insideHPC.

Video: House Hearing on the Exascale Challenge

Rich Brueckner — Thu, 23 May 2013 00:06:57 +0000

In this video, the House Committee on Science, Space, and Technology’s Subcommittee on Energy holds a May 22 hearing to examine HPC research and development challenges and opportunities, specifically as they relate to exascale computing.

Testifying before the Subcommittee were Dr. Roscoe Giles, Chairman of the Advanced Scientific Computing Advisory Committee; Dr. Rick Stevens, Associate Laboratory Director for Computing, Environment and Life Sciences at Argonne National Laboratory; Ms. Dona Crawford, Associate Director for Computation at Lawrence Livermore National Laboratory; and Dr. Daniel Reed, Vice President for Research and Economic Development at the University of Iowa.

Exascale computing will be an important part of a larger effort to improve the U.S.’s overall high-end computing capability to address a broad range of academic, industrial, and national security needs. While research in next generation computing architecture and software continues to require strategic government investments, Members also explored the significant economic benefits that can arise from full utilization of existing high performance computing capabilities in ongoing scientific research.

Read the Full Story or check out the hearing charter.

The post Video: House Hearing on the Exascale Challenge appeared first on insideHPC.

Podcast: Is the D-Wave Machine Really a Quantum Computer?

Rich Brueckner — Wed, 22 May 2013 18:46:18 +0000

In this NPR Podcast, Geoff Brumfiel takes a look at D-Wave quantum computer. Some skeptics are saying that the machine does not really work with the quantum effects that D-Wave claims it does.

It’s not exactly science, what they’re doing,” says Christopher Monroe, a physicist with the Joint Quantum Institute at the University of Maryland. “It’s high-level engineering, and I think it’s high-level salesmanship, too.” But Monroe remains skeptical. He believes that the D-Wave team has never demonstrated that entanglement is happening on the chips in its machine. He believes that D-Wave’s supposedly quantum bits are actually working instead as tiny electromagnets. Those magnets, Monroe believes, could be interacting in ways to solve a certain problem very quickly without quantum mechanics. “There’s no evidence that what they’re doing has anything to do with quantum mechanics,” he says. If he’s right, then D-Wave’s machine may be far more narrow in its abilities than the company believes.

Read the Full Story * Download the MP3 * Download the transcript

The post Podcast: Is the D-Wave Machine Really a Quantum Computer? appeared first on insideHPC.