Putting this computing power into a small and energy-efficient package, and making it reliable and easier to program, are among the goals of the new DARPA Ubiquitous High Performance Computing (UHPC) initiative. Georgia Tech researchers from three different units are supporting key components of this $100 million challenge, which will require development of revolutionary approaches not bound by existing computing paradigms.
If UHPC meets its ambitious eight-year goals, the new approaches and technologies it develops could redefine the way that computing systems are envisioned, designed and used.
"The opportunity we have is to go far beyond the current product roadmaps," said David Bader, a professor in Georgia Tech's School of Computational Science and Engineering. "We really have the opportunity to change the industry and to design our applications with new computing architectures. For the first time in the history of computing, we will be able to work with a clean slate."
To attain the program's ambitious goals, DARPA funded four groups -- led by NVIDIA Corp., Intel Corp., the Massachusetts Institute of Technology and Sandia National Laboratories -- to develop UHPC prototypes. A fifth group, led by the Georgia Tech Research Institute (GTRI), will develop applications, benchmarking and metrics that will be used to drive UHPC system design considerations and support performance analysis of the developing system designs.
"Our team is developing a set of five difficult problems of a size and scope that the machines they are talking about should be able to accomplish," said Dan Campbell, a GTRI principal research engineer who is co-principal investigator of the benchmarking initiative. "Our challenge is picking the right problems and specifying them at the right level of abstraction to allow innovation and properly represent what the DoD will need in 2018."
The five problems highlight the unique computing needs of the U.S. military:
• Analysis of the vast streams of data originating with widespread sensor systems, unmanned aerial vehicles and new generations of radar systems. The data will be analyzed for nuggets of useful information in ways that are not possible today.
• A dynamic graph challenge, in which many entities interact to create a problem of "connecting the dots." That could mean analyzing relationships in social media to find possible adversaries, or understanding network traffic for cyber-security challenges.
• The decision tree, comparable to a chess game in which many possible interconnected options, each with complex implications, must be analyzed quickly. This could help field commanders or corporate CEOs make better decisions.
• Materials shock and hydrodynamics issues, challenges important to improving future generations of materials.
• Molecular dynamics simulations, which use high-performance computers to understand interactions between very large systems, such as protein folding.
"We need to be able to take in a lot more data and understand it a lot more thoroughly than we can now," said Mark Richards, a principal research engineer in the Georgia Tech School of Electrical and Computer Engineering and co-principal investigator of the benchmarking team. "That might allow us to find adversaries we can't find now because we're unable to tease that information out of the data flow."
While the benefits of making such computing power widely available are obvious, how these machines will be designed, built and reliably operated is not.
"Meeting these very ambitious program goals will pose significant technical challenges," said Bader, who leads application development on the NVIDIA team and is part of the benchmarking group. "The technology roadmaps in such areas as interconnection networks, microprocessor design and technology fabrication will be pushed to their limits."
Meeting power limitations of just 57 kilowatts per rack -- the amount of electricity produced by a portable military generator -- may be the toughest among them. The fastest computer currently in operation requires seven megawatts of power.
"Reducing the power consumption means less energy per computation," noted Richards. "But as we lower the device voltage, we get closer to the physical noise. That will allow more errors due to the physics of the devices, and all kinds of things will have to be done to address that."
And the entire machine will have to fit into a 24-inch wide, 78-inch high and 40-inch deep cabinet.
But the physical implementation of the machines is just one part of the challenge, Bader noted. How people will work with them poses a perhaps more difficult challenge because it will require thinking about computers in a new way.
"Over the past 20 or 30 years, we've taken a single computing design and kept tweaking it through advances like miniaturizing parts," he said. "But we really haven't changed the global nature of how the machine works. To meet DARPA's power efficiency goals, we really will need to change the way we program the machine."
That also affects the humans who interact with these highly-parallel machines, which could have as many as a half-million separate threads operating at the same time. DARPA's initial goal is to build machines capable of petaflop speed -- a trillion operations per second -- which could lead into the next generation of exascale computers a thousand times more capable.
"We will need to find new ways of thinking about computers that will make it feasible for humans to comprehend what is going on inside," Campbell said. "It's a huge programming challenge."
To encourage collaboration in solving these complex problems, DARPA has embraced the idea of open innovation. It expects the organizations to work together on common critical topics, creating a collaborative environment to address the system challenges. New technology generated by the program -- believed to be today's largest DoD computing research initiative -- is likely to move quickly into industry.
"There is certainly an expectation among the companies that what they are doing in this project is going to change how we do mainstream computing," Bader said. "The technology transfer implications are certainly obvious."
]]>College of Computing
404-312-6620
]]>Similar to thePetit Institute for Bioengineering and Bioscience, IPaT will catalyze researchactivities, create new economic development opportunities, and addressimportant societal problems. It will support various college research centersthat collectively pursue transformations in healthcare, education, consumermedia, and other complex human enterprises by integrating advances inhuman-centered computing, architectural and digital design, policy, and systemscience and engineering.
"IPaTwill create an innovation crossroads where Georgia Tech faculty, students,industry partners, government partners and other stakeholders meet toco-innovate, collaborate and pave the road for Georgia Tech research that addressescomplex societal challenges,” stated Cross. “IPaT will focus onengaging the Tech community and our external partners in far-reachingleadership goals through its investment in unique research platforms, throughour living laboratories and datasets, through partnership with Georgia Tech'sEnterprise Innovation Institute (EI2), and through alignment with the many ongoingresearch activities on campus.”
The Institutefor People and Technology will be led by Executive Director Beth Mynatt, Collegeof Computing professor and former director of the GVU Center. Mynatt is arenowned researcher in human-computer interaction, health informatics andubiquitous computing. She is a member of ACM's SIGCHI Academy, is a Sloan andKavli research fellow and serves on Microsoft Research's Technical AdvisoryBoard. She has published more than 100 scientific papers and recently chairedthe CHI 2010 conference, the premier international conference in human-computerinteraction. Prior to joining the faculty in 1998, she was a member of theresearch staff at Xerox PARC, working with the founder of ubiquitous computing,Mark Weiser. Mynatt earned both her master’s and PhD in computer science fromGeorgia Tech.
"I'm thrilled that Beth has agreed to lead the Institute forPeople and Technology," said Cross. "Her vast understanding of the complexitiesinherent in today's social and business institutions and the transformativerole of technology in those enterprises is a critical foundation for this newinstitute. Moreover, her track record in bringing together diverse stakeholdersto accomplish difficult interdisciplinary and translational research and toprovide solutions for today’s challenges is a prized asset for leading aninstitute that will connect so many parts of the Georgia Tech community."
The IPaTleadership team includes Deputy Director Jeff Evans, deputy director of the Georgia Tech Research Institute (GTRI)Information and Communications Laboratory. In addition to being a graduate ofGeorgia Tech, Evans has been a research engineer for GTRI for more than 20years, directing more than 70 successful external sponsored research programsin networked systems, performance applications and emerging wireless services.He has collaborated extensively across campus in leading and supportingresearch efforts in the Georgia Centers for Advanced Communications Technology (GCATT) and several interdisciplinary centers, establishingnetwork-based platforms and test beds to support a wide range of researchefforts. Evans will continue the development of technology convergence testbeds for IPaT, and his leadership will promote a broad set of researchopportunities for a diverse group of government and industry interactions withGeorgia Tech.
Ron Hutchins(OIT), Bill Rouse (ISyE/COC), and Renu Kulkarni (GTRI/EI2) will also play importantleadership roles. IPaT will be supported by EI2’s Strategic Partners Office.
IPaT willbring together many notable, cutting-edge and renowned research activities,including:
Theseexisting centers represent every college at Georgia Tech as well as GTRI andEI2.
]]>Workingwith Dr. John Xerogeanes, chief of the Emory Sports Medicine Center, AllenTannenbaum, the Julian Hightower professor of bioengineering at the GeorgiaInstitute of Technology, has developed 3-D MRI technology that allows surgeons topre-operatively plan and perform anatomic ACL surgery.
Tannenbaumand student researchers in the Wallace H. Coulter Department of BiomedicalEngineering at Georgia Tech and Emory University created the 3-D MRI technologythat allows the surgeon to see from one point of the knee to another duringligament replacement. All of themedical imaging processing and algorithmic work was done through Tannenbaum’sMinerva Research Group at Georgia Tech.
“Thedevelopment of this interactive computer software allows much safer repair ofthe ACL in young athletes with a much smaller chance of complications,”Tannenbaum said. “It is an excellent example of how 3-D MRI data, inconjunction with state-of-the-art image processing, can help the practicingsports surgeon in a key image-guided surgery and surgical planning task.”
The ACLis one of the four major ligaments in the knee, and ACL tears are one of themost common injuries in children who participate in contact sports such asfootball, basketball, soccer and gymnastics.
Traditionaltreatment for ACL injuries in kids has been rehabilitation, wearing a brace andstaying out of athletics until the child stops growing - usually in themid-teens - and ACL reconstruction surgery can safely be performed.
“Theproblem with doing surgery on a young child is that if you damage the growthplate, you can cause a growth disturbance,” said Xerogeanes, an associate professorin the Department of Orthopaedics at Emory University School of Medicine.
The ACLis like a rubber band that attaches at two points to stabilize the knee. Inorder to replace the ligament, surgeons create a tunnel in the upper and lowerknee bones, slide the new ACL between those two tunnels and attach it to bothends. The new ligament is typically taken from either a hamstring tendon orallograft tissue, which is donor material.
Prior tousing the 3-D MRI technology, ACL operations were conducted with extensive use ofX-rays in the operating room and left too much to chance when working aroundgrowth plates, researchers said.
With thisnew technology, surgeons can actually see from one point to the other on eitherside of the knee and can correctly position the tunnels where they will placethe new ligament. The surgery can be done in less time than the traditionalsurgery and with complete confidence that the growth plates in young patientswill not be damaged.
Kids whoundergo this type of operation will still have at least one year of recoverytime, Xerogeanes said. The good news is that it does allow them to eventuallypursue normal activity.
Xerogeanesand his colleagues at Emory are performing the anatomic ACL reconstructiontechnique on adult patients as well as pediatric patients. He hopes that anotheradvantage of this new anatomical procedure will be that it helps preventre-injury in the future for all athletes who have suffered from ACL tears.
Tannenbaum also has an appointment in Georgia Tech's School of Electrical and Computer Engineering.
]]>