How Three Global Organizations Are Harnessing the Power of GPU Computing

By Nicole Hemsoth

November 3, 2009

The groundswell of enthusiasm for GPU computing was underscored recently by the Institute of Processing Engineering (IPE), Chinese Academy of Sciences (CAS).

Since IPE’s founding in 1958, the institute has made extensive use of high performance computing (HPC) for its investigations into chemical engineering, metallurgy, biochemistry, material science, energy and green technology. IPE notes that, “The semiconductor technology for integrated circuits, which has powered the dramatic development of HPC, is now approaching its physical limit, for the foreseeable future at least. Fine-grained multi-thread and many-core parallel computation is, most probably, the only get-around to maintain this development, and NVIDIA’s massively parallel CUDA™ architecture has established a new paradigm at this critical moment.

IPE is not alone. Companies, laboratories and research institutions around the world are turning to GPU computing to augment or even replace their CPU clusters. The power of the GPU model has become available to developers through CUDA, a software and hardware architecture that enables GPUs to be programmed using high level languages such as C. And now NVIDIA has introduced the next generation of CUDA GPU architecture, codenamed “Fermi”, that delivers breakthroughs in both graphics and GPU computing, plus support for C++ in addition to C, Fortran, Java, Python, OpenCL and Direct Compute.

GPU computing is an international phenomenon. Below are brief descriptions of how three organizations, the Institute of Processing Engineering at the prestigious Chinese Academy of Sciences, Bloomberg in the United States and the U.K.’s BAE Systems, are using NVIDIA GPU-based solutions to meet their HPC requirements.

IPE

IPE at CAS, located in Beijing, has been developing the energy-minimization multi-scale (EMMS) method for more than 25 years. Though the method was originally aimed at multi-phase flow, it gave birth to a highly scalable and general-purpose simulation methodology for different systems that took into account the structural similarities between the HPC hardware, software and the problem to be solved. An algorithm framework applicable to a wide range of physical systems and computer architectures was thereby proposed during 2000~2003, but put on hold due to the fact that the institute’s CPU-based systems could not handle the computational complexity involved.

In 2007, with the release of CUDA 1.0 from NVIDIA, the situation changed dramatically. Within a year IPE had deployed its first GPU computing system that embodied their EMMS principle. Based on 126 HP workstations with 200 NVIDIA Tesla C870 GPU computing processors, the system, Mole-9.7, delivered over 100 teraflops (TF) of peak performance in single precision, enough horsepower to tackle some of their most pressing simulation problems.

That was just the beginning. The IPE system later received another shot in the arm with the addition of over 600TF worth of NVIDIA GT200 GPUs, and was interconnected with other IPE computational resources over a 20 Gbp/s InfiniBand fabric. The whole GPU-based system, known as Mole -8.7, broke the petaflops (PF) barrier in March 2009 and was by far the largest GPU cluster in China and one of the largest in the world.

With its ability to realize the structural similarity between hardware, software and the computed system, IPE is in an advantageous position to solve industrial problems efficiently using such an HPC system.

Here are just a few of the CUDA applications at IPE:

  • Direct numerical simulation of multiphase flow – One of the most compute-intensive and intractable problems in classic physics, direct numerical simulation using Navier-Stokes equations should allow scientists to predict the flow field in engineering when other numerical approaches fail. Based on CUDA and MPI, the main application is aimed at fluid beds, stirred tanks, bubble columns, etc. So far a 20x boost in speed has been realized in practical applications, compared to one core of a mainstream CPU.
  • Quasi-realtime simulation of industrial reactors – Mole-8.7 is now capable of simulating meter-scale reactors on sub-millimeter scales in near real-time. A simulation using 480 GPUs and 60 million solid particles was recently completed. The goal is to improve the simulation methodology and speed so reactor designers can specify the configuration and operational conditions of the equipment and examine its performance in close to real time.
  • Simulations of oil recovery and carbon storage – The IPE HPC cluster is also being used to simulate the flow of oil and water in porous rocks, a capability important for oil exploration and recovery. The scientists are using multi-scale methods in order to simulate flow behavior on a geological scale (hundreds of kilometers) on one hand, and micro-scale effects (less than a millimeter) on the other. These simulations require an elaborate coupling between GPU and CPU computation, which is facilitated by CUDA.

In recognition of its extensive development and advancement of science through the use of CUDA and GPU computing, NVIDIA recently awarded CAS-IPE the distinction of a CUDA Center of Excellence. NVIDIA and CAS-IPE will continue to work together to further the value of GPU computing with next generation technologies.

Bloomberg  

Despite the balky economy, transactions involving thousands of mortgages are still a routine occurrence in the financial markets. For example, collateralized debt obligations (CDOs) and collateralized mortgage obligations (CMO) make up baskets of thousands of loans that are publicly traded financial instruments.

Bloomberg, headquartered in New York with offices around the world, is a leading financial services organization. One of its services is modeling the risks and determining the price of CMO/CDO baskets for its customers. The company uses powerful algorithms to calculate massive amounts of data and deliver pricing information based on large scale simulation. Using CPUs, these calculations would have taken 16 hours, an unacceptable solution for an overnight computational run.

Bloomberg’s solution was to implement an NVIDIA Tesla GPU computing system in its data center and port the CMO/CDO application to run on the CUDA architecture. The results were immediate and dramatic. Large calculations that dragged on for hours using the CPU-based cluster were now completed in minutes; smaller runs that once took 20 minutes take only seconds.

The speed-ups delivered by NVIDIA CUDA and GPU computing allowed Bloomberg to process the CMO/CDO prices up to 50 times faster than on a CPU. The company is now investigating the possibility of offering real-time pricing – a huge step forward for the industry. 

Bottom line, Bloomberg’s pricing calculations for securities backed by assets such as mortgages, home equity loans and auto loans, as well as mortgage-backed securities such as CMOs, have seen an 800% speed up in processing compared to its CPU solution. At the same time, the data center is now consuming three times less power and the NVIDIA hardware is taking up one quarter of the tile space. 

Bloomberg is also researching other uses for its GPU computing cluster such as valuations of certain types of derivative products, risk management and portfolio valuations. 

In the last analysis, Bloomberg’s customers are the biggest winners in this shift to GPU computing. They are now working with the most current pricing information using the most predictive models, a capability that provides them with a serious competitive advantage in a market where timing is everything. 

BAE Systems

As the premier global defense, security and aerospace company, BAE Systems designs, builds and supports a wide range of products, from unmanned aircraft to land vehicles, ships and submarines. Understanding the aerodynamic performance of its products, using the principles of computational fluid dynamics (CFD), is a vital part of the design process. However, as the requirements of their customers have become more complex and development timescales more challenging, the simulation tools currently available to BAE Systems’ engineers have begun to limit their ability to affordably explore design options with greater accuracy.

Physical model tests are extremely expensive, with a single wind tunnel test campaign costing more than $700,000, while virtual testing is limited by the power and cost of conventional computing systems. In its efforts to develop more affordable and comprehensive CFD simulations, BAE Systems has turned to NVIDIA Tesla GPUs.

One of the goals of the Mathematical Modeling team at the BAE Systems Advanced Technology Center is to make engineering design tools like CFD more affordable and capable so its simulations need no longer be limited to a small number of design conditions, such as cruise, but can be applied to the whole performance envelope that the products will experience in service. Assessing all the design points of an aircraft can require between ten thousand and twenty million individual simulations, which is an unfeasibly costly and time consuming process using conventional CFD techniques.

Jamil Appa’s team recognized that the highly parallel and computationally intensive nature of its CFD and visualization problems made them ideally suited to processing on the GPU. Using Tesla GPUs, BAE Systems has succeeded in creating a system which packs the power of 60 Nehalem cores into a workstation-sized unit capable of processing both visualization and CFD to generate high fidelity 3D simulations.

“Up until now, everyone thought real time visualization and CFD was too ambitious,” says Jamil Appa. “By running our visualization techniques and CFD codes on the GPU we’ve proved them wrong – we’re now very close to having a system which is not just faster but actually interactive. In tests with production test cases we used an NVIDIA Tesla S1070 solution and achieved 100x times to solution speed-ups, meaning tests which previously took hours now require a matter of minutes.

“GPU computing is extremely exciting because it can deliver both a cost saving, in terms of power, total cost of ownership and software licenses, and a significant increase in performance. We’re working with NVIDIA to continue developing the potential of our GPU systems – this technology is hugely significant for our industry and for high performance computing as a whole.”

GPU – A Technology That’s Time Has Come

The IPE at the Chinese Academy of Sciences, Bloomberg and BAE Systems are just three examples of the many organizations around the world that are benefitting from GPU computing. Now, with the recent introduction of the “Fermi” architecture in September, 2009, NVIDIA has taken a major step in ensuring that the GPU is a technology that’s time has come.

To find out how other organizations around the world are using NVIDIA’s GPU computing solutions to solve their HPC problems, visit http://www.nvidia.com/page/pg_20040317553375.html

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

ExxonMobil, NCSA, Cray Scale Reservoir Simulation to 700,000+ Processors

February 17, 2017

In a scaling breakthrough for oil and gas discovery, ExxonMobil geoscientists report they have harnessed the power of 717,000 processors – the equivalent of 22,000 32-processor computers – to run complex oil and gas reservoir simulation models. Read more…

By Doug Black

TSUBAME3.0 Points to Future HPE Pascal-NVLink-OPA Server

February 17, 2017

Since our initial coverage of the TSUBAME3.0 supercomputer yesterday, more details have come to light on this innovative project. Of particular interest is a new board design for NVLink-equipped Pascal P100 GPUs that will create another entrant to the space currently occupied by Nvidia's DGX-1 system, IBM's "Minsky" platform and the Supermicro SuperServer (1028GQ-TXR). Read more…

By Tiffany Trader

Tokyo Tech’s TSUBAME3.0 Will Be First HPE-SGI Super

February 16, 2017

In a press event Friday afternoon local time in Japan, Tokyo Institute of Technology (Tokyo Tech) announced its plans for the TSUBAME3.0 supercomputer, which will be Japan’s “fastest AI supercomputer,” Read more…

By Tiffany Trader

Drug Developers Use Google Cloud HPC in the Fight Against ALS

February 16, 2017

Within the haystack of a lethal disease such as ALS (amyotrophic lateral sclerosis / Lou Gehrig’s Disease) there exists, somewhere, the needle that will pierce this therapy-resistant affliction. Read more…

By Doug Black

HPE Extreme Performance Solutions

Object Storage is the Ideal Storage Method for CME Companies

The communications, media, and entertainment (CME) sector is experiencing a massive paradigm shift driven by rising data volumes and the demand for high-performance data analytics. Read more…

Weekly Twitter Roundup (Feb. 16, 2017)

February 16, 2017

Here at HPCwire, we aim to keep the HPC community apprised of the most relevant and interesting news items that get tweeted throughout the week. Read more…

By Thomas Ayres

Alexander Named Dep. Dir. of Brookhaven Computational Initiative

February 15, 2017

Francis Alexander, a physicist with extensive management and leadership experience in computational science research, has been named Deputy Director of the Computational Science Initiative at the U.S. Read more…

Here’s What a Neural Net Looks Like On the Inside

February 15, 2017

Ever wonder what the inside of a machine learning model looks like? Today Graphcore released fascinating images that show how the computational graph concept maps to a new graph processor and graph programming framework it’s creating. Read more…

By Alex Woodie

Azure Edges AWS in Linpack Benchmark Study

February 15, 2017

The “when will clouds be ready for HPC” question has ebbed and flowed for years. Read more…

By John Russell

TSUBAME3.0 Points to Future HPE Pascal-NVLink-OPA Server

February 17, 2017

Since our initial coverage of the TSUBAME3.0 supercomputer yesterday, more details have come to light on this innovative project. Of particular interest is a new board design for NVLink-equipped Pascal P100 GPUs that will create another entrant to the space currently occupied by Nvidia's DGX-1 system, IBM's "Minsky" platform and the Supermicro SuperServer (1028GQ-TXR). Read more…

By Tiffany Trader

Tokyo Tech’s TSUBAME3.0 Will Be First HPE-SGI Super

February 16, 2017

In a press event Friday afternoon local time in Japan, Tokyo Institute of Technology (Tokyo Tech) announced its plans for the TSUBAME3.0 supercomputer, which will be Japan’s “fastest AI supercomputer,” Read more…

By Tiffany Trader

Drug Developers Use Google Cloud HPC in the Fight Against ALS

February 16, 2017

Within the haystack of a lethal disease such as ALS (amyotrophic lateral sclerosis / Lou Gehrig’s Disease) there exists, somewhere, the needle that will pierce this therapy-resistant affliction. Read more…

By Doug Black

Azure Edges AWS in Linpack Benchmark Study

February 15, 2017

The “when will clouds be ready for HPC” question has ebbed and flowed for years. Read more…

By John Russell

Is Liquid Cooling Ready to Go Mainstream?

February 13, 2017

Lost in the frenzy of SC16 was a substantial rise in the number of vendors showing server oriented liquid cooling technologies. Three decades ago liquid cooling was pretty much the exclusive realm of the Cray-2 and IBM mainframe class products. That’s changing. We are now seeing an emergence of x86 class server products with exotic plumbing technology ranging from Direct-to-Chip to servers and storage completely immersed in a dielectric fluid. Read more…

By Steve Campbell

Cray Posts Best-Ever Quarter, Visibility Still Limited

February 10, 2017

On its Wednesday earnings call, Cray announced the largest revenue quarter in the company’s history and the second-highest revenue year. Read more…

By Tiffany Trader

HPC Cloud Startup Launches ‘App Store’ for HPC Workflows

February 9, 2017

“Civilization advances by extending the number of important operations which we can perform without thinking about them,” Read more…

By Tiffany Trader

Intel and Trump Announce $7B for Fab 42 Targeting 7nm

February 8, 2017

In what may be an attempt by President Trump to reset his turbulent relationship with the high tech industry, he and Intel CEO Brian Krzanich today announced plans to invest more than $7 billion to complete Fab 42. Read more…

By John Russell

For IBM/OpenPOWER: Success in 2017 = (Volume) Sales

January 11, 2017

To a large degree IBM and the OpenPOWER Foundation have done what they said they would – assembling a substantial and growing ecosystem and bringing Power-based products to market, all in about three years. Read more…

By John Russell

US, China Vie for Supercomputing Supremacy

November 14, 2016

The 48th edition of the TOP500 list is fresh off the presses and while there is no new number one system, as previously teased by China, there are a number of notable entrants from the US and around the world and significant trends to report on. Read more…

By Tiffany Trader

Lighting up Aurora: Behind the Scenes at the Creation of the DOE’s Upcoming 200 Petaflops Supercomputer

December 1, 2016

In April 2015, U.S. Department of Energy Undersecretary Franklin Orr announced that Intel would be the prime contractor for Aurora: Read more…

By Jan Rowell

D-Wave SC16 Update: What’s Bo Ewald Saying These Days

November 18, 2016

Tucked in a back section of the SC16 exhibit hall, quantum computing pioneer D-Wave has been talking up its new 2000-qubit processor announced in September. Forget for a moment the criticism sometimes aimed at D-Wave. This small Canadian company has sold several machines including, for example, ones to Lockheed and NASA, and has worked with Google on mapping machine learning problems to quantum computing. In July Los Alamos National Laboratory took possession of a 1000-quibit D-Wave 2X system that LANL ordered a year ago around the time of SC15. Read more…

By John Russell

Enlisting Deep Learning in the War on Cancer

December 7, 2016

Sometime in Q2 2017 the first ‘results’ of the Joint Design of Advanced Computing Solutions for Cancer (JDACS4C) will become publicly available according to Rick Stevens. He leads one of three JDACS4C pilot projects pressing deep learning (DL) into service in the War on Cancer. Read more…

By John Russell

HPC Startup Advances Auto-Parallelization’s Promise

January 23, 2017

The shift from single core to multicore hardware has made finding parallelism in codes more important than ever, but that hasn’t made the task of parallel programming any easier. Read more…

By Tiffany Trader

IBM Wants to be “Red Hat” of Deep Learning

January 26, 2017

IBM today announced the addition of TensorFlow and Chainer deep learning frameworks to its PowerAI suite of deep learning tools, which already includes popular offerings such as Caffe, Theano, and Torch. Read more…

By John Russell

CPU Benchmarking: Haswell Versus POWER8

June 2, 2015

With OpenPOWER activity ramping up and IBM’s prominent role in the upcoming DOE machines Summit and Sierra, it’s a good time to look at how the IBM POWER CPU stacks up against the x86 Xeon Haswell CPU from Intel. Read more…

By Tiffany Trader

Leading Solution Providers

Nvidia Sees Bright Future for AI Supercomputing

November 23, 2016

Graphics chipmaker Nvidia made a strong showing at SC16 in Salt Lake City last week. Read more…

By Tiffany Trader

BioTeam’s Berman Charts 2017 HPC Trends in Life Sciences

January 4, 2017

Twenty years ago high performance computing was nearly absent from life sciences. Today it’s used throughout life sciences and biomedical research. Genomics and the data deluge from modern lab instruments are the main drivers, but so is the longer-term desire to perform predictive simulation in support of Precision Medicine (PM). There’s even a specialized life sciences supercomputer, ‘Anton’ from D.E. Shaw Research, and the Pittsburgh Supercomputing Center is standing up its second Anton 2 and actively soliciting project proposals. There’s a lot going on. Read more…

By John Russell

Container App ‘Singularity’ Eases Scientific Computing

October 20, 2016

HPC container platform Singularity is just six months out from its 1.0 release but already is making inroads across the HPC research landscape. It's in use at Lawrence Berkeley National Laboratory (LBNL), where Singularity founder Gregory Kurtzer has worked in the High Performance Computing Services (HPCS) group for 16 years. Read more…

By Tiffany Trader

Dell Knights Landing Machine Sets New STAC Records

November 2, 2016

The Securities Technology Analysis Center, commonly known as STAC, has released a new report characterizing the performance of the Knight Landing-based Dell PowerEdge C6320p server on the STAC-A2 benchmarking suite, widely used by the financial services industry to test and evaluate computing platforms. The Dell machine has set new records for both the baseline Greeks benchmark and the large Greeks benchmark. Read more…

By Tiffany Trader

IDG to Be Bought by Chinese Investors; IDC to Spin Out HPC Group

January 19, 2017

US-based publishing and investment firm International Data Group, Inc. (IDG) will be acquired by a pair of Chinese investors, China Oceanwide Holdings Group Co., Ltd. Read more…

By Tiffany Trader

What Knights Landing Is Not

June 18, 2016

As we get ready to launch the newest member of the Intel Xeon Phi family, code named Knights Landing, it is natural that there be some questions and potentially some confusion. Read more…

By James Reinders, Intel

KNUPATH Hermosa-based Commercial Boards Expected in Q1 2017

December 15, 2016

Last June tech start-up KnuEdge emerged from stealth mode to begin spreading the word about its new processor and fabric technology that’s been roughly a decade in the making. Read more…

By John Russell

Intel and Trump Announce $7B for Fab 42 Targeting 7nm

February 8, 2017

In what may be an attempt by President Trump to reset his turbulent relationship with the high tech industry, he and Intel CEO Brian Krzanich today announced plans to invest more than $7 billion to complete Fab 42. Read more…

By John Russell

  • arrow
  • Click Here for More Headlines
  • arrow
Share This