By Neal Singer
A new supercomputer rating system will be released by an international team led by Sandia on Nov. 17 at the upcoming Supercomputing Conference 2010 in New Orleans.
The rating system, Graph500, tests supercomputers for their skill in analyzing large, graph-based structures that link the huge numbers of data points present in biological, social, and security problems, among other areas.
“By creating this test, we hope to influence computer makers to build computers with the architecture to deal with these increasingly complex problems,” says Richard Murphy (1422).
“The very careful and thoughtful definition of this new competitive standard is both quite subtle and tremendously important, as it may heavily influence computer architecture for decades to come,” says Rob Leland, director of Computations, Computers, and Math Center 1400.
The group isn’t trying to compete with Linpack, the current standard test of supercomputer speed, says Richard. “There have been lots of attempts to supplant it, and our philosophy is simply that it doesn't measure performance for the applications we need, so we need another, hopefully complementary, test.”
(Many scientists view Linpack as a ‘plain vanilla’ test mechanism that tells how fast a computer can perform basic calculations, but has little relationship to the actual problems the machines must solve.)
The impetus to achieve a supplemental test code came about at “an exciting dinner conversation at Supercomputing 2009 [conference],” Richard says. “A core group of us recruited other professional colleagues, and the effort grew into an international steering committee of over 30 people.” See www.graph500.org.
Many large computer makers have indicated interest, says Richard, who says there’s been buy-in from Intel Corp., IBM, AMD Inc., NVIDIA Corp., and Oracle Corp. “Whether or not they submit test results remain to be seen, but their representatives are on our steering committee.”
Each organization has donated time and expertise of committee members, he says.
While some computer makers and their architects may prefer to ignore a new test for fear their machine will not do well, the hope is that a large-scale demand for a more complex test will be a natural outgrowth of the greater complexity of problems.
How does it work?
Large data problems are very different from ordinary physics problems.
Unlike a typical computation-oriented application, large data analysis problems often involve searching large, sparse data sets performing very simple computational operations.
To deal with this, the Graph500 benchmark creates two computational kernels: a large graph that inscribes and links huge numbers of participants and a parallel search of that graph.
“We want to look at the results of ensembles of simulations, or the outputs of big simulations in an automated fashion,” says Richard. “The Graph500 is a methodology for doing just that. You can think of them being complementary in that way — graph problems can be used to figure out what simulation actually told us.”
Performance for these applications is dominated by the ability of the machine to sustain a large number of small, nearly random remote data accesses across its memory system and interconnects, as well as the parallelism available in the machine.
Five problems for these computational kernels could be cybersecurity, medical informatics, data enrichment, social networks, and symbolic networks. “Many of us on the steering committee believe that these kinds of problems have the potential to eclipse traditional physics-based high-performance computing over the next decade,” says Richard.
While general agreement exists that complex simulations work well for the physical sciences, where lab work and simulations play off each other, there is some doubt simulations can solve social problems that have essentially infinite numbers of components. These include terrorism, war, epidemics, and societal problems.
“These are exactly the areas that concern me,” Richard says. “There’s been good graph-based analysis of pandemic flu. Facebook shows tremendous social science implications. Economic modeling this way shows promise.”
Studies show that moving data around (not simple computations) will be the dominant energy problem on exascale machines, the next frontier in supercomputing and the subject of a nascent DOE initiative to achieve this next level of operations within a decade. (Petascale and exascale represent 10 to the 15th and 18th powers, respectively, operations per second.)
Part of the goal of the Graph500 list is to point out that in addition to more expense in data movement, any shift in application base from physics to large-scale data problems is likely to further increase the application requirements for data movement, because memory and computational capability increase proportionally. That is, an exascale computer requires an exascale memory.
“In short, we’re going to have to rethink how we build computers to solve these problems, and the Graph500 is meant as an early stake in the ground for these application requirements,” Richard says.
“We’re all engineers and we don’t want to over-hype or over-promise, but there’s real excitement about these kinds of big data problems right now,” he says. “We see them as an integral part of science, and the community as a whole is slowly embracing that concept. However, it’s so new we don't want to sound as if we're hyping the cure to all scientific ills. We’re asking, ‘What could a computer provide us?’ but ignoring the human factors in problems that may stump the fastest computer. That’ll have to be worked out.” -- Neal Singer
Tracy Solomon’s business cards depict three photographs of uniformed veterans — himself, his father, and his grandfather — but this service-disabled Navy veteran doesn’t dwell on the injuries he sustained during the first Gulf War. He’d rather talk about how his small business can help Sandia researchers get the test and measurement equipment they need and how his company’s services add value to their purchases.
VETERAN AND SMALL BUSINESS owner Tracy Solomon, right, of TEVET LLC, and Kent Childs (1748) look over some equipment that Solomonís company supplied to Sandia. (Photo by Randy Montoya)
Solomon’s company, TEVET LLC — short for Test Equipment Veterans — is a service-disabled veteran-owned small business (SDVOSB) that provides equipment from manufacturers such as Agilent Technologies.
TEVET is one of 120 veteran-owned companies doing business at the Labs whose owners have started companies after serving in the military or being wounded while serving.
The owners are known as “vetrepreneurs,” and under federal law, Sandia can set aside procurement contracts for veteran-owned companies and SDVOSBs, recruit them, and advocate for them.
Toni Leon Kovarik, an advocate in Small Business Utilization Dept. 10222 for veteran-owned businesses, says contract-related payments to veteran-owned small businesses were more than $20 million in fiscal year 2009, including about $3.1 million to SDVOSBs.
Toni recruits veteran-owned firms to bid on contracts at Sandia, attending events and holding office hours at the Veterans Procurement Assistance Center in Albuquerque.
John Smatana, vice president of sales for Factory Express, a certified SDVOSB, was introduced to Toni through the center. That initial contact eventually resulted in a contract for the Albuquerque office supply firm to supply and service high-security paper shredders, as well as other supplies and equipment repair, at Sandia.
Consistency of work makes a difference
Smatana and CEO David Zimpelman, who was diagnosed with a serious medical condition while serving at Kirtland Air Force Base in the 1980s, say the consistency of the work with Sandia has helped the company weather the recent economic downturn and helped it keep a full-time service technician on board, which helps all its New Mexico customers.
Smatana says the company carries the red, white, and blue SDVOSB logo on its website and letterhead.
Being registered as an SDVOSB “adds legitimacy to the company for the government purchasers. I think the next step is making sure that they have the opportunities for the service-disabled veteran-owned companies. I think they’re out there, but I think there’s a lot more that could be set aside,” he says.
Toni also works with small businesses to prepare them to bid on projects at the Labs and to work with Sandia, which typically has more business requirements than the private sector, and sometimes encourages companies to partner to provide goods and services to the Labs.
“Working with veterans is an honor,” Toni says. “We have some people who have just returned from service. To be able to help point them in the right direction, toward services that will help them grow their companies, means a lot.”
Set-asides are competitive in marketplace
Toni also helps veteran-owned companies and SDVOSBs through her work inside the Labs by learning about pending contracts that such firms could bid on and by working with Sandia’s contracting representatives to encourage them to set aside contracts for such companies.
Jeff Miller (10248), a contracting representative, recently set aside a Just In Time (JIT) agreement for SDVOSBs to provide Cisco networking hardware to the Labs. A decision on the agreement, estimated to be worth about $3.5 million for up to seven years, is expected to be made later this month.
Because he’s spending taxpayer money, Jeff says he needs to make sure set-asides are still competitive by ensuring that the Labs can select from a pool of highly qualified companies.
“We don’t think there will be any detriment to Sandia because there are enough qualified companies out there that can offer the service we need, as well as good competitive pricing,” he says.
Zimpelman agrees that set-aside contracts cannot be handouts. Veteran-owned businesses need to earn their business and offer competitive prices and quality, but he thinks programs to help such companies are good public policy.
“The reason to support that is that we need to have some patriotism. Let’s take care of those who took care of our country,” he said. “I think it’s good that this society as a whole supports those who are making a sacrifice.”
Solomon agrees: “A young man or woman who enters military service is making a conscious decision to step forward and sign themselves over as property to their government to do with them what they will and to be placed in harm’s way, if necessary, in order to protect and preserve this great nation. So I think it’s important, especially in the veteran community, that the government does something for those who were willing to sacrifice.”
Son and grandson of veterans
Solomon, the son of a Vietnam veteran wounded in the Tet Offensive and the grandson of a World War II veteran who was wounded after escaping from capture during the Bataan Death March, continues to suffer from pain due to his own injuries incurred during the first Gulf War, but overcame those obstacles and led a successful corporate career before starting TEVET. Both his father and grandfather received Purple Hearts.
He could have located the new company anywhere, but he chose to start his business to help those in his rural Appalachian hometown of Greeneville, Tenn., a Historically Underutilized Business Zone (HUBZone), in 2003.
For companies like TEVET — which obtained the first SDVOSB and HUBZone competitive set-aside procurement at Sandia — these set-asides have meant economic development and job creation.
Sixty percent of TEVET’s workforce is veterans and the company just secured a 7,500-square-foot facility in Tennessee, Solomon says. The company also has seen its revenue grow each year, nearly doubling in 2010.
When TEVET began working with Sandia in 2009, the company opened an office in Albuquerque, and Solomon expects to expand further in 2011.
“Our goal is to grow our footprint in New Mexico, hiring people here locally,” Solomon says.
The National Cancer Institute recently announced two five-year awards totaling nearly $4 million for a partnership between the University of New Mexico Cancer Center and Sandia. One $1.95 million grant will fund the creation of a joint Cancer Nanotechnology Platform Partnership, and another $1.8 million grant will pay for a new Cancer Nanotechnology Training Center to train a new generation of multidisciplinary scientists. In addition, the State of New Mexico is providing another $2 million to build a lab supporting Sandia Fellow and UNM professor Jeff Brinker’s (1002) research, which is devoted to nano-biomaterials and nanomedicine. UNM donated more than 4,500 square feet of lab space in the new Centennial Engineering Building for the project. Construction is scheduled to begin December 2010.
“Sandia is proud to be a part of this important undertaking,” says Steve Rottler, Chief Technology Officer and VP of Science and Technology and Research Foundations Div. 1000. “Pairing Sandia’s expertise in materials science with UNM Cancer Center researchers’ knowledge of cancer biology, oncology, and clinical attributes provides an ideal setting in which to move forward in our nation’s fight against cancer.”
Second phase of alliance
The awards comprise the second phase of the NCI’s Alliance for Nanotechnology in Cancer Program, which initiated an investment of more than $30 million per year for the next five years to establish Centers of Cancer Nanotechnology Excellence, Cancer Nanotechnology Platform Partnerships, training grants, and the Nanotechnology Characterization Laboratory. The alliance was founded in 2004 to leverage specific advantages of nanotechnology to improve cancer diagnosis, treatment, and prevention. Since then, the alliance has facilitated the discovery of many novel technologies, some of which are currently undergoing commercialization and clinical trials.
Jeff, who is coprincipal investigator on the grant with Dr. Cheryl Willman, a physician and director and CEO of the UNM Cancer Center, and a Distinguished Professor of Chemical and Nuclear Engineering at UNM, used a Sandia-funded Laboratory Directed Research and Development grant to conduct preliminary research in nanofabrication. The nanostructures he developed form the basis for delivering drugs directly to a wide variety of cancer targets, a method that increases the drug’s effectiveness and reduces side effects.
“The technology I developed with now-Truman Fellow Carlee Ashley  and UNM colleagues is really a generic platform to target any arbitrary cancer, so we’ve already written other proposals and are interacting with other cancer research centers throughout the US and Canada, to go after low-outcome cancers like breast, lung, pancreatic, and liver,” Jeff says.
Jeff is also a team leader in the Cancer Nanotechnology Training Center, which is focused on training multidisciplinary scientists at both Sandia and UNM. One such student, Carlee, started working as an undergraduate student in Jeff’s lab in Sandia’s Advanced Materials Laboratory. Jeff then served as her graduate coadviser with David Peabody of UNM, and Carlee recently earned one of Sandia’s prestigious Truman Fellowships.
“Carlee’s experience was the model for the training grant. Her work was used as the basis for what we’d like to do in the future,” Jeff says. “She went from biochemistry to training in my lab and then chemical engineering. That kind of interdisciplinary training is something that Sandia and UNM are actively encouraging.”
The next five years of the NCI’s Alliance program will focus on rapidly advancing new nanotechnology discoveries and speeding their transformation into cancer-relevant applications in clinical practice; aiding nanoparticle characterization and standardization of characterization methods to enable technology transfer from university laboratories to companies that bring these technologies to patients; and developing the next-generation of cancer researchers in the area of nanotechnology. - Stephanie Hobby