Inside a NASA Production Supercomputing Center

By Nicole Hemsoth

June 18, 2012

It takes a super workload management tool to power grid, cluster and on-demand computing environments for computational modeling and simulation applications at NASA.Photo Courtesy of NASA Ames Research Center

In the past, scientific and engineering advancements relied primarily on theoretical studies and physical experiments. Today, however, computational modeling and simulation are equally valuable in such endeavors, especially for an agency such as the National Aeronautics and Space Administration (NASA). With a mission “to pioneer the future in space exploration, scientific discovery and aeronautics research,” the use of high-end computing (HEC) for high-fidelity modeling and simulation has become integral to all three of NASA’s mission directorates: aeronautics research, human exploration and operations, and science.

These HEC resources are provided at NASA’s Advanced Supercomputing (NAS) Division at Ames Research Center, Moffett Field, Calif. NAS offers production and development systems to U.S. scientists in government, industry and at universities, with users currently numbering over 1,500. Projects such as designing safe and efficient space exploration vehicles, projecting the impact of human activity on weather patterns and simulating space shuttle launches are studied using the facility’s supercomputers. “We provide world-class HEC and associated services to enable NASA scientists and engineers in all mission directorates to broadly and productively employ large-scale modeling, simulation and analysis for mission impact. We pursue a future where these services empower ever greater NASA mission successes,” says William Thigpen, deputy project manager for the High End Computing Capability (HECC) Project at the NAS Division.

The facility’s current HEC systems include two supercomputers, a 115-petabyte mass storage system for long-term data storage, two secure front-end systems requiring two-factor authentication and two secure unattended proxy systems for remote operations. Key system resources at NAS include: Pleiades, a 126,720-core, 1.75 petaFLOPS (Pflop/s) SGI® Altix® ICE cluster; Columbia, a 4,608-processor SGI Altix® (Itanium 2); and the hyperwall-2, a 1,024-core, 128-node GPU cluster.

Since 300 to 400 jobs are typically running 24 hours a day, seven days a week, the NAS staff works nonstop to meet the demands for time on these machines. “Our mission is to accelerate and enhance NASA’s mission of space exploration, scientific discovery and aeronautics research by continually creating and ensuring optimal use of the most productive HEC environment in the world,” says Thigpen. “Our viewpoint is that we spend a lot of money getting hardware in here, but it really makes sense that it is effectively exploited by our users because the bottom line is we’re not about big hardware, we’re about big science and engineering.”

Building a Supercomputer – Pleiades

Originally installed in the fall of 2008, the Pleiades supercomputer is an SGI Altix ICE 8200/8400 InfiniBand® cluster with Intel® Xeon® quad, hex, and eight-core processors. Considered one of the most powerful general-purpose supercomputers ever built, each of the Pleiades 182 racks (11,776 nodes) has 16 InfiniBand switches to provide the 12D dual-plane hypercube that provides the interconnect for the cluster. The InfiniBand fabric interconnecting Pleiades’  nodes requires more than 65 miles of cabling. Pleiades is the largest (measured by number of nodes) InfiniBand cluster in the world.

Ranked as one of the world’s most powerful computers, the Pleiades supercomputer was built to augment NASA’s current and future high-end computing requirements. “Pleiades is a general-purpose machine and provides for all three components of supercomputers – [capability, capacity and time critical],” says Thigpen. “We have users running jobs using over 18,000 cores, providing new insights into the formation of the universe. There are numerous users running parameter studies (often thousands) using from one to a few thousand cores. Pleiades is also being utilized to answer time-critical questions concerning the shuttle.”

Choosing Components and Software

Pleiades was built to meet as many of the emerging NASA science and engineering mission requirements as possible while remaining within the HEC budget. “The Pleiades architecture was chosen because it provided the best performance/cost ratio of the systems we looked at. Since its original installation in 2008, it has undergone eight expansions. We will continue to build it out as long as the fundamental economics of the system remain sound, and the science and engineering returns remain high,” states Thigpen.

To build Pleiades, NAS engineers began with the components recommended by the vendor and those being used on other systems. The result has been an easy transition to the new environment for NASA users. “We want an environment where the components complement each other, are an easy natural transition for our users and provide a reliable environment,” says Thigpen. For example, the SGI ICE 8200 and 8400 are standard products that have been taken to an extreme size at the NAS facility. Additionally, the InfiniBand network was expanded to incorporate both the data analysis and visualization cluster, as well as the storage system.

“Another consideration is outlining and selecting a scalable architecture,” explains Alan Powers, HPC architect with Computer Science Corp., which holds the primary support contract at the NAS Division,. “We chose SGI because it had a certain architecture that allowed us to build and grow it. [It also had] the best price/performance based on our workload. Where we are today, we’re near a petaflop capability, and it’s been built over a couple of years; we’ve been adding to it slowly. The other vendors’ price/performance wasn’t even close to this platform.”

Managing the Workload

When providing supercomputing resources to 1,500 users, 24/7, workload management is a top priority. Originally developed at NAS in the 1990s and then commercialized, PBS Professional® workload management software has been used since its inception. Commercially developed by Altair Engineering, Inc., Troy, Mich., the PBS platform is designed to power grid, cluster and on-demand computing environments. PBS Professional is used to manage all HEC resources at NAS, including Pleiades.

PBS Professional is a resource allocation tool that makes it possible to create intelligent policies to manage distributed, mixed-vendor computing assets as a single, unified system. Based on a policy-driven architecture, it continually optimizes how technical HEC resources are used, ensuring that they are used effectively and efficiently. Simply put, the software looks at the jobs that want to run, looks at the resources available for them to run on and makes the best match based on a number of criteria. “Those criteria can include the user that’s running and how many jobs that user currently has running, or how many cores his job is currently using. It can also be the queue that a user submits their job in, and those queues can have things limiting them, like how many jobs are running or how many cores all of the jobs together are using. It also can be the mission directorate those users are in,” explains Thigpen.

Powers adds: “The ‘P’ in PBS stands for ‘portable,’ and it allows us to run this on any architecture. We’ve had PBS on fat node architectures, on thin node clients and on IBM architectures. PBS has been able to adapt to all those computing environments. This has allowed our users to have a consistent set of batch scripts across these different environments. They only have to learn one thing. So one, it’s flexible; two, we can use it on any architecture; and three, it’s easy for users to learn.”

Your Own HEC Environment

According to Thigpen, HEC is an enabling technology that allows a company to build products that can meet their customers’ requirements in a cost-effective manner: “By spending a relatively small amount on a system, they can run through hundreds or thousands of alternatives before building a physical prototype. This will allow for a better product with lower production costs.”

However, there are many issues to address when considering whether an HEC environment is the right choice for an enterprise. “There has to be a balance between the cost of the resources, the technology they enable, the increased productivity of their staff, the potential return on their investment and what their competition is doing,” Thigpen concludes.

 

Footnote: As originally published in Altair’s Concept to Reality magazine’s 2011 Fall/Winter issue. Actual stats updated as reflected on NASA’s Pleiades web page, http://www.nas.nasa.gov/hecc/resources/pleiades.html.

Subscribe to HPCwire's Weekly Update!

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

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire