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

By Jan Rowell

December 1, 2016

In April 2015, U.S. Department of Energy Undersecretary Franklin Orr announced that Intel would be the prime contractor for Aurora: a leadership-class, 200 petaflops supercomputer to be installed at Argonne National Laboratory in 2018. The new system—part of the Collaboration of Oak Ridge, Argonne, and Lawrence Livermore (CORAL) initiative—aims to be 20 times more powerful than Argonne’s current Mira system. Power consumption will top out at a peak of 16.5 megawatts, just 3.4 times Mira’s energy level. Among other requirements, Aurora must support mission-critical computing from day one, bolster US competitiveness, and move the industry significantly closer to exascale computing. It needs to deliver breakthrough capabilities for both traditional HPC and data analytic applications. And it must mesh with future product plans for both Intel and Cray, Intel’s key collaborator on the initiative.

This article is the first in a series that goes behind the scenes to explore the making of the forthcoming Aurora supercomputer. We’ll share what we learned in interviews with more than a dozen Intel technologists, plus insights from individuals at Cray and the DOE’s Argonne Leadership Computing Facility (ALCF), where Aurora will reside and where technologists are co-designing the system.

You’ll hear how the three organizations are working together to manage the risks and deliver an optimal solution for DOE science. You’ll get a peek under the hood and a glimpse of the technology highlights. And you’ll see the human side. What draws people to such a high-stakes project? What keeps them awake at night? What do they hope their work will mean for DOE—and for HPC users everywhere?

Opportunities and Risks

Aurora is a landmark supercomputer that represents significant new opportunities for Intel, Cray, DOE, and the industry. Aurora will not only enable a new wave of scientific and technical computing for DOE and its research partners, its innovations, delivered at smaller scales by Intel and Cray, will also help drive HPC’s continued expansion into enterprise data centers around the world. These smaller-scale systems will no doubt spur breakthroughs for researchers and commercial users alike.

With Intel as the prime contractor, Aurora offers a high-visibility showcase for Intel Scalable System Framework (Intel SSF)—a chance to demonstrate the full power of integrating Intel’s latest technology advances within a flexible, unified architecture. Aurora will incorporate the forthcoming Intel Xeon Phi processor, codenamed Knights Hill, deployed on next-generation 10nm process technology. The system’s interconnect will utilize a next-generation Intel Omni-Path fabric implemented in silicon photonics, along with an advanced memory hierarchy. The file system, based on Intel Enterprise Edition for Lustre Software, will provide more than 150 petabytes capacity, with throughput of more than a terabyte per second. Intel HPC Orchestrator will help simplify integration, deployment, and operations.

Aurora also advances Intel’s ability to excel at the systems thinking that’s essential to keeping HPC on its faster-than-Moore’s-Law performance curve. “Aurora is the flagship to show what Intel SSF is capable of,” says Jake Wood, CORAL program manager at Intel Federal. “It’s all about systems-level thinking and unprecedented technology integration.”

ALCF Intel Cray leadership meeting
(left to right) Rick Stevens, Argonne; Susan Coghlan, Argonne; Al Gara, Intel; Mark Seager, Intel; Raj Hazra, Intel; Barry Bolding, Cray; Gary Geissler, Cray; Paul Messina, Argonne; Mike Papka, Argonne (Source: Intel)

But there are a range of risks, as well—from the basic financial responsibilities of the prime contractor, to designing with leading-edge technologies, to managing such a large and complex undertaking. The Aurora design includes next-generation Intel technologies that were only on the drawing boards when the project started. The undertaking involves deep collaboration among DOE, Cray, and Intel, and the breadth of innovations means that Intel is drawing resources from multiple divisions across the company. The CORAL team at Intel bridges enterprise and government, chip design and software solutions, platform engineering and system packaging, and more. The scale of the project presents new challenges for everything from software design to system validation.

And although Intel has been part of numerous government contracts, you have to look back to the vaunted Accelerated Strategic Computing Initiative (ASCI) Red supercomputer—the first teraflops system—to find such a large-scale program where Intel was the prime contractor rather than ingredient supplier.

Meet the Maestro

While Al Gara is the architect of the Aurora supercomputer, Jake Wood is the program manager and project foreman, responsible for ensuring that all the moving parts line up correctly, the juggling pins stay up in the air, and the system comes together as specified. He is also program manager for Intel’s other DOE collaborations, including R&D for exascale computing.

jake-wood-500x
After 18 years in manufacturing operations, Jake Wood became CORAL program manager for Intel. His job is much the same: get in the middle of super-smart people and provide a framework for them to solve big, complex problems.

Wood is a 20-year Intel veteran who grew up a town or two over from Intel’s Jones Farm campus west of Portland, Oregon. He headed to Chicago for college, earning BS degrees in Industrial Engineering and Manufacturing Engineering from Northwestern University. But Wood was eager to get back to the camping and fishing opportunities of the Northwest. Intel was staffing up its Fab 5 fabrication facility in Aloha, Oregon, and Wood signed on. His first position was night shift operations manager, helping to deliver first-generation Intel Pentium processors.

After advancing through a variety of manufacturing engineering positions, Wood took advantage of an opportunity that he says ended up transforming his professional life. In 2005, Intel sent him and his family to MIT for two years to participate in MIT’s Leaders for Manufacturing Program, now called the Leaders for Global Operations. Wood came away with a Master’s in Civil Engineering, an MBA from MIT’s Sloan School of Management, and a vastly expanded set of skills and insights.

“MIT gave me more context and insight around the business side of manufacturing—more business acumen and critical thinking in the manufacturing space,” Wood recalls. When he returned to Oregon, he moved into manufacturing management and leadership positions at Intel’s D1D facility, overseeing the work of hundreds of technicians and their managers, and ensuring their work kept the factories operating on Intel’s precision cadence.

A Framework for Brilliant People to Solve Problems

The CORAL position came at a time when Wood was looking to stretch in a new direction, and Intel was looking for someone with a unique skillset to drive the execution of the Aurora program plan. “Intel knew this program needed a different skillset to go drive a physical deployment, handle system integration, and take on all the other aspects of being the prime contractor,” Wood says. “It was a match, and I Jumped in with both feet.”

Wood thought he was making a 180-degree turn from his leadership roles in manufacturing management. In fact, although he’s on the other end of the manufacturing pipeline now, his current position is remarkably similar at its core to his previous ones.

“My job is still to get in the middle of the best and brightest technical people in the world in their fields, and help them solve very complex, interrelated problems,” he says. “I’m not there to be smarter than them or to solve their technical issues, but to provide context, give them a framework to more easily solve their problems, and help them drive into the space where we all need to be to meet our objectives. Take away the brilliant person with a PhD in some aspect of materials engineering and swap in the brilliant person with a PhD in some aspect of high performance computing, and it’s very, very similar.”

And, he adds, “I’m absolutely loving it.”

Checkpoint: Theta and Beyond

When we spoke with Wood in October, he said Aurora is on track for first-rack delivery in Q4 2018, and the CORAL teams at DOE, Intel, and Cray are celebrating the on-time installation and acceptance of the Theta supercomputer. Theta is based on the Cray XC40 platform with Intel Xeon Phi processors, previously codenamed Knights Landing.

Theta is one of the key scientific platforms for the CORAL program and serves as an important early Aurora production system for the ALCF. At 8.5 petaflops, the system is a bridge between the Mira and Aurora systems, and an advancement of Cray’s leadership in adaptive supercomputing. “Theta represents a huge success,” Wood says. “It’s a small system relative to Aurora, but it you look at it on its own, it’s very much a major supercomputer. It’s a significant platform for DOE science.”

The Theta installation went very smoothly, according to Susan Coghlan, deputy division director of ALCF. “Argonne drove an aggressive timeline to have the system installed and accepted, and we all had to work as a very tight team to make it happen,” she says. “We all pulled together, and with a lot of planning and preparation, we had the hardware installed in an afternoon. It was unbelievably smooth and insanely fast—one of the best I’ve ever seen. We also learned a lot about how to work together, and that’s going to benefit us on the Aurora install.”

Theta is now supporting ALCF’s Early Science computing and paving the way to Aurora. ALCF put on an four-day workshop for participants in the Theta Early Science Program to get their hands on the system, experience the development and runtime environments, work with ALCF staff to fine-tune their key applications, and assess their scalability. By the workshop’s end, Coghlan reports, all 12 early science teams had their codes up and running on Theta, some at full scale.

Now, the teams are using Theta to address real-world research problems. “Theta is a fabulous stepping stone to Aurora, and Aurora is going to be a really big step toward exascale and the leaps in science that it is going to make possible,” says Coghlan. “It’s exciting.”

­­­­­­­­­­­­­­­­­­­_________________

Watch for part 2 of our Aurora series, which explores some of Aurora’s technology challenges and how Intel is stepping up its game to meet them. Subsequent articles will drill down on the close collaboration among Intel, Cray, and Argonne to create Aurora, and the system’s likely impacts.

About the Author

Jan Rowell covers technology trends in high-performance computing, life sciences, healthcare, and other fields.

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!

Can Cerabyte Crack the $1-Per-Petabyte Barrier with Ceramic Storage?

July 20, 2024

A German startup named Cerabyte is hoping to solve the burgeoning market for secondary and archival data storage with a novel approach that uses lasers to etch bits onto glass with a ceramic coating. The “grey ceramic� Read more…

Weekly Wire Roundup: July 15-July 19, 2024

July 19, 2024

It's summertime (for most of us), and the HPC-related headlines aren't as plentiful as they once were. But not everything has to happen at high tide-- this week still had some waves! Idaho National Laboratory's Bitter Read more…

ARM, Fujitsu Targeting Open-source Software for Power Efficiency in 2-nm Chip

July 19, 2024

Fujitsu and ARM are relying on open-source software to bring power efficiency to an air-cooled supercomputing chip that will ship in 2027. Monaka chip, which will be made using the 2-nanometer process, is based on the Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI projects.). Other tools have joined the CUDA castle siege. AMD Read more…

Quantum Watchers – Terrific Interview with Caltech’s John Preskill by CERN

July 17, 2024

In case you missed it, there's a fascinating interview with John Preskill, the prominent Caltech physicist and pioneering quantum computing researcher that was recently posted by CERN’s department of experimental physi Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to communities across the globe. As climate change is warming ocea Read more…

Can Cerabyte Crack the $1-Per-Petabyte Barrier with Ceramic Storage?

July 20, 2024

A German startup named Cerabyte is hoping to solve the burgeoning market for secondary and archival data storage with a novel approach that uses lasers to etch Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI pro Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to com Read more…

Shutterstock 1886124835

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical c Read more…

Shutterstock 2203611339

NSF Issues Next Solicitation and More Detail on National Quantum Virtual Laboratory

July 10, 2024

After percolating for roughly a year, NSF has issued the next solicitation for the National Quantum Virtual Lab program — this one focused on design and imple Read more…

NCSA’s SEAS Team Keeps APACE of AlphaFold2

July 9, 2024

High-performance computing (HPC) can often be challenging for researchers to use because it requires expertise in working with large datasets, scaling the softw Read more…

Anders Jensen on Europe’s Plan for AI-optimized Supercomputers, Welcoming the UK, and More

July 8, 2024

The recent ISC24 conference in Hamburg showcased LUMI and other leadership-class supercomputers co-funded by the EuroHPC Joint Undertaking (JU), including three Read more…

Generative AI to Account for 1.5% of World’s Power Consumption by 2029

July 8, 2024

Generative AI will take on a larger chunk of the world's power consumption to keep up with the hefty hardware requirements to run applications. "AI chips repres Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

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_1687123447

Nvidia Economics: Make $5-$7 for Every $1 Spent on GPUs

June 30, 2024

Nvidia is saying that companies could make $5 to $7 for every $1 invested in GPUs over a four-year period. Customers are investing billions in new Nvidia hardwa Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Leading Solution Providers

Contributors

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l 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…

IonQ Plots Path to Commercial (Quantum) Advantage

July 2, 2024

IonQ, the trapped ion quantum computing specialist, delivered a progress report last week firming up 2024/25 product goals and reviewing its technology roadmap. 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…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh 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…

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…

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…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire