By David McNally, U.S. Army Research Laboratory Public Affairs
ABERDEEN PROVING GROUND, Md.– In the past few weeks, computer scientists and technicians unboxed and installed two powerful new supercomputers at the U.S. Army Research Laboratory.
The upgrades at the ARL DOD Supercomputing Resource Center are part of a modernization program to bring new capabilities to researchers. These massive room-size computers deliver lightning fast processing speeds and the power necessary to run complex simulations and computations.
“Our most recent acquisitions are the Hellfire system and the Centennial system, which is aptly named after Aberdeen Proving Ground’s 100th anniversary,” said Lee Ann Brainard, ARL DSRC deputy director “Both of these resources are expected to be operational later this summer.”
More than 60 Army programs of record depend on the service’s use of advanced computing and networking resources.
“Supercomputing provides a technological advantage for DOD projects,” said Dr. Raju Namburu, ARL DSRC director. “It helps to reduce costs by reducing reliance on expensive and destructive live experiments and prototype demonstrations. We empower researchers to solve the most difficult military operational challenges through advanced computing.
Namburu said researchers also use supercomputers to analyze vehicle animations that can be rotated, taken back and forward in time, manipulated and improved.
Computer scientists at the Army Research Laboratory’s DOD Supercomputing Resource Center use high performance computing to increase mission effectiveness through test and experimentation savings and discovering lethality and protection solutions through technology. (Photo Credit: U.S. Army)
“It takes a fraction of the time it would take to conduct a series of physical experiments,” he said. “You also are left with more interesting data to work with in determining what just happened, why, and how to improve.”
Increasing capabilities allows researchers to use different aspects of applications so they can design and improve systems at a faster pace to increase the nation’s security, he said.
“High performance computing is a proven, critical enabling technology for the Army’s research, development, test and evaluation community,” Namburu said.
Between 2003 and 2007, when there was an average of 27 fatalities each month caused by IED attacks in Iraq, Army investments in supercomputers enabled the Army Underbody Blast Research Program to design and rapidly deploy kits to mitigate the threat.
“The Army ground vehicle fleet was equipped with armor reinforcements in less than four months by leveraging the center’s resources,” said David Lyon, chief of the Protection Division for the laboratory’s Weapons and Materials Research Directorate. “It accelerated the research and development of the new armor designs and paved way for a remarkably rapid fielding of a technology that has saved countless lives.”
Army researchers used high performance computing to develop a vehicular armor upgrade kit, known as Frag 6 Kits, for Humvees in Iraq. It successfully mitigated IED Threats.
“High performance computing resources and expertise enabled a 20 percent reduction in fatalities and resulted in three to four times faster deployment rate for the kits,” Lyon said.
In the future, officials project demand and capacity for the Army’s supercomputers to increase dramatically.
“There is an insatiable need for high performance computing and advanced networking infrastructure by the DOD RDT&E community,” he said.
One of the Army Research Laboratory DOD Supercomputing Resource Center’s newest supercomputers is the Centennial system, which is named after Aberdeen Proving Ground’s 100th anniversary. Researchers plan to have the system fully installed and operational later this summer. (Photo Credit: U.S. Army)
HISTORY
Since the 1940s, the Army sought advanced computers to accomplish things like calculating artillery firing tables. According to 50 Years of Army Computing, a publication from 1996 that recounted a two-day anniversary celebration, the tables showed Soldiers “what angle of elevation was required for a specific projectile to impact a target at a specified range with a given propellant charge.”
Historically, the Ballistic Research Laboratory (ARL’s predecessor) played a significant role in the evolution of scientific computing architectures and technologies. The ENIAC, was the world’s first general purpose electronic digital computer to be designed with an internally stored program.
Over the years, the Army has invested in supercomputers with advanced capabilities.
In response to a 1992 Congressional inquiry as to why the United States was falling behind in high performance computing, the DOD created the High Performance Computing Modernization Program, recounted Charlie Nietubicz, who went on to become the first director of the ARL Major Shared Resource Center (predecessor to the ARL DSRC). Nietubicz retired as director in 2010.
“It was really the right thing to do at the time,” Nietubicz said. “The program goals included the creation of high performance computing centers, which continue to show a significant return on investment.”
The DOD established the centers to support focused military research, development, test and evaluation with high performance computing resources and technology, he said. Today there are five such centers across the nation within the DOD.
The Army center leverages in-house research initiatives in advanced computing architectures, tactical high performance computing and advanced networks.
“In the tactical high performance computing research domain, the Army needs mobile, adaptable, cost-effective, low power high performance computing solutions to efficiently process the voluminous data from sensors in the digital and networked battlespaces,” said computer scientist Bob Sheroke. “We truly have a world-class supercomputing infrastructure and computational science research portfolio to support the U.S. warfighter.”
“With these resources, the quality of modeling and simulations has increased to provide support to various complex computational science and engineering applications,” said Lee Ann Brainard.
The computers are getting more powerful and physically larger. For example, the room-size DSRC Cray XC-40, known as Excalibur, debuted as number 19 on the world’s top 500 list of most powerful supercomputers in 2015.
The ENIAC, in the Ballistics Research Laboratory, Building 328. Left: Glen Beck Right: Frances Elizabeth Snyder Holberton. (Photo Credit: K. Kempf)
SUPERCOMPUTING IN RESEARCH
Researchers working on vehicle protection face challenging ballistic scenarios where there’s no single armor solution. Dr. Robert Doney, a computational physicist with ARL’s Weapons and Materials Research Directorate, said they must consider many potential materials and strategies.
“To improve on an armor design, we need to know how it performs when struck by any of the many anti-armor threats,” Doney said. “High performance computing has been revolutionary in digging into the kinematics and providing insight to these problems by providing sub-microsecond temporal resolution and access to variables unavailable from experimentation.”
Doney and his colleagues use Army supercomputers to run hundreds to thousands of armor simulations per year that use several hundred million hours of CPU time.
“To understand how big of a number that is, one hundred million hours is about 11,400 years,” he said. “If all of our simulations were represented by one simulation, run on a computer with one CPU, you would have had to start the simulation in 9,400 BC for it to be finishing about today –assuming no interruptions.”
Research requirements are growing rapidly and high performance computing gives scientists greater confidence in the quality of their solutions and the robustness of a supercomputer-improved armor design, he said.
Research mechanical engineer, Dr. Luis Bravo, ARL Vehicle Technology Directorate, is the principal investigator for a five-year DOD Frontier project that involves the teaming of High Performance Computing Modernization Program resources with DOD laboratory and test center domain expertise. Bravo uses supercomputers to create high-resolution 3-D simulations of the diesel fuel injection process.
“Supercomputing enables cutting-edge propulsion research in ever more extreme conditions and regimes that are far less understood and more difficult to diagnose experimentally,” Bravo said. “Such insights are invaluable for guiding the development of stable, reliable systems providing major savings in time and cost.”
For researchers looking at the next generation of vertical lift, supercomputers are critical to the success of their work. Dr. Rajneesh Singh, research aerospace engineer and a team leader at VTD, does a lot of research for the DOD’s Joint Multi-Role Technology Demonstrator program.
“Because the program is looking at future helicopters that can achieve very high speed, the designs are unique, innovative and nontraditional,” Singh said. “A lot of our conventional knowledge does not apply.”
The team uses high performance computing to look at simulations in a variety of operational environments.
“Even on high performance computers, each simulation takes days,” Singh said. “If we didn’t have access to these resources our research would just not be possible.”
Modeling and simulation identifies numerous key technology challenges within the Joint Multi-Role Technology Demonstrator program. Army researchers use supercomputers to look at the next generation of vertical lift. (Photo Credit: U.S. Army illustration)
UPGRADES
The DOD High-Performance Computing Modernization Program typically invests in supercomputing upgrades at the facility every two years.
Engineers are in a constant state of planning for the complex system upgrades. Brainard said they prepare for the improvements more than a year before a new system is expected.
The sheer amount of electricity and cooling infrastructure to support the supercomputers is staggering. For example, the Excalibur has more than 101,000 CPUs.
“Robust engineering is required to balance high availability with the massive scale of power and cooling required’, said Tom Kendall, technical director at the center.
“To enable weeklong computations at full system scale to routinely complete, facility and system engineering team need to work together’, he said.
In 2012, The Laboratory maintained its daily missions while moving to a new facility with increased floor space for the massive computers. In 2014, the laboratory partnered with the U.S. Army Installation Management Command to add a 1,200-ton water plant to keep the computer systems water-cooled in a way that uses far less energy than traditional cold water designs.
The Centennial system augments the center’s Excalibur supercomputer and extends the total unclassified high performance computing capability to nearly 6.3 petaflops. A petaflop is a measure of the computer’s processing speed. One petaflop is thousand trillion operations per second.
Computer scientists at the Army Research Laboratory’s DOD Supercomputing Resource Center use high performance computing to increase mission effectiveness through test and experimentation savings and discovering lethality and protection solutions through technology. (Photo Credit: Jim Nelson)
PARTNERSHIPS
In the future, the Army plans to continue using powerful supercomputers to save lives and money.
The center recently partnered with the Army Test and Evaluation Command during a Network Integration Evaluation exercise. ATEC engineers were gathering massive amounts of sensor data of up to one terabyte a day. With the organization’s small-scale computational resources, it took up take 60 hours of processing to analyze and make assessment of the data.
“Since 2012, the DOD Supercomputing Resource Center, co-located at Aberdeen Proving Ground, has been instrumental in supporting the Aberdeen Test Center in meeting mission by providing access to skilled data scientists and huge computing power,” said John Wallace, ATC technical director. “Data computations, WIN-T for example, used to take over 80 hours on available test center computers can now be completed in just four hours with HPC.”
Wallace said this increased efficiency and provided a quality data product.
“Access to and collaboration with the data scientists at the HPC was critical in this and many other successes,” he said.
“Our ARL team worked hand-in-hand with ATC to develop tools and algorithms to enable rapid and efficiently processing of the data,” said Brian Panneton, ARL Computational and Information Sciences Directorate. “This is the tip of the iceberg in terms of using HPC as a game-changing technology for the Army’s test and evaluation community.”
The center also instituted an outreach program to inform and educate users across the DOD and their research partners the capabilities and resources available with the Army supercomputers.
“We anticipate significant growth in the use of these resources that will really enhance Army readiness,” Namburu said. “Our vision is to accelerate research, development, technology and engineering breakthroughs, and it’s something we are achieving.”
This article was originally published on Army.mil.
Subscribe to Army AL&T News, the premier online news source for the Acquisition, Logistics, and Technology (AL&T) Workforce.
RELATED LINKS
U.S. Army Materiel Command
U.S. Army Research, Development and Engineering Command
U.S. Army Research Laboratory
ARL DOD Supercomputing Resource Center
DOD High Performance Computing Modernization Program
50 Years of Army Computing\History of Army Supercomputing
ARL on Facebook
ARL on Twitter
Joint Multi-Role Technology Demonstrator program
