High Performance Computing: Keystone for Future Opportunities
Shirley Ann Jackson, Ph.D.
President, Rensselaer Polytechnic Institute
EMPAC Concert Hall
8 a.m. Friday, October 28, 2011
Welcome to Rensselaer Polytechnic Institute. We are delighted to host the High Performance Computation Conference. I hope the many sessions this week exploring Industry Use of High-Performance Computing (HPC), Software Challenges, Cross-Sector Collaboration, and Applications have facilitated knowledge sharing and built foundations for cooperation and collaboration.
Today, we have “The Executive Session.” We will hear from two panels of leaders on two key topics: Competitiveness and Future Opportunities.
I readily accepted Mr. Kolb’s invitation to moderate these panels because High-Performance Computation (or HPC) is fundamental to our meeting our challenges in energy, food production, security, healthcare, and protecting our environment. I will add jobs and prosperity to that list because, across the spectrum of activities, we depend, more and more, on High-Performance Computing for innovation and virtually every economist agrees that opportunity, productivity, and wealth depend on innovation.
The essential role of High-Performance Computing is not new. Over ten years ago, as we developed The Rensselaer Plan with the idea of setting our university’s course into the future, we foresaw the rich potential and the vital need for High-Performance Computing. The opportunities presented by bringing the power of parallel processing, simulation, intense computation, and management of extremely large data sets were clear, even then. Further, we understood that, in addition to the intrinsic value these tools bring, even more value could be obtained through partnerships. Working across sectors and across disciplines will allow us to leverage the full potential of High-Performance Computing and that is why we can look with optimism at the competitive advantage and the opportunities HPC will bring.
The building in which we sit, the Experimental Media and Performing Arts Center (EMPAC), could not have been built without tapping the power of High-Performance Computing. EMPAC has full access to the resources of the Computational Center for Nanotechnology Innovations.
But the role of HPC in a building dedicated to active blending of science, engineering, and media artscontinues to evolve.
During this conference, many of the specific opportunities for HPC have been explored. Certainly, it is encouraging to see that the key technologies that undergird HPC continue to be expanded. The horizon of power and performance for these tools and techniques continues to recede, providing astonishing raw capability that will match HPC to the grand challenges of our time.
We also have seen that new vistas for applying HPC, and for combining approaches and adapting solutions are before us. These demonstrate how HPC will change our lives and change our world by helping us to meet major challenges, such as bringing a deeper understanding of genomics to healthcare, more efficiently managing energy sourcing and use, the modeling of complex devices, and the enablement of new scientific discoveries.
In one session, the question of competitiveness was addressed explicitly. At times in the past in HPC, this has been seen mostly in terms of “feeds and speeds.” But for some time now, we have understood that competitiveness requires more than simply more powerful tools. Over and over again, we have seen that competitiveness emerges from sharing great ideas and working together on important problems across a spectrum of endeavors and enterprises in multidisciplinary ways. This requires partnerships that create synergies of talent, resources, and perspectives.
A meeting, such as this one, brings together imaginative, capable people and allows them to begin relationships that can result in new projects, enriched by the contributions of diverse collaborators.
And, when the public sector comes together with the private sector, their different needs and values reveal new things that HPC can do. When we bring together those who have expertise in new techniques with those who have the know-how to scale up their work, whole new solutions become available. When we bring together teams that already are succeeding as they work on important problems, and we make these efforts visible to those who have not yet found partners, we provide examples and approaches that can broaden the appreciation and the necessary skills for working together to achieve new aims.
Such collaborations provide real world experiences that help us to assemble the components of an innovation ecosystem. This can generate competitiveness over the long term by taking the best advantage of available resources and talentand, indeed, attracting more talent and more resources so that these are leveraged with increasing effectiveness.
An innovation ecosystem requires four things:
The first element is strategic focus. Among a world of possibilities, we must choose promising areas to explore and develop, and these must match the talent, resources, and opportunities we have or can attract. Further, our focus needs to be realistic with regard to timing to not work on yesterday’s challenges.
The second element is idea generation. Game-changing ideas tend to arise out of basic research, which pushes the boundaries of human knowledge. Universities are critical players here, because basic research dovetails magnificently with our educational mission. The primary contribution of universities to our ecosystem is the education of bright, motivated people, who ask questions that may take decades to answer. Furthermore, the endpoints of basic research in terms of commercial technologies often cannot be envisaged even by the researchers themselves. Yet, history shows that out of such open exploration, thriving industries are born.
When we fund basic research, we are funding serendipity. Even a sober, frugal, post-recession United States must invest in serendipity, because without it, there is no vitality in the innovation ecosystem. Indeed, there is no innovation.
The third element requires translational pathways that bring discoveries into commercial, or societal, use.
The protection, regulation, and exploitation of intellectual property are the front-end of translation. The Bayh-Dole Act sought to spur this from academic research by giving universities ownership of the results of their federally funded research, and the right to patent and license them, and to share royalties with the researchers. Through the deliberate exploitation of their intellectual property, modern research universities are linked to the marketplace more strongly than ever before.
It may need updating, but Bayh-Dole has been successful, spinning off thousands of new enterprises based on university patents. Patents are important for Intellectual Property protection and exploitation for universities, private enterprises, and government. But University of Michigan Law School Professor Rebecca Eisenberg has warned of a potential “tragedy of the anti-commons,” if the increasing proliferation of patents in upstream research makes it difficult for researchers downstream to access the permission and information they need to build on previous innovations.
Our innovation ecosystem has much to gain if universities fling open such storehouses of knowledge. Industry, too, must beware of controlling its intellectual property so jealously that it stifles innovation.
Intellectual property protection and exploitation does not comprise a translational pathway in and of itself. Many patents emerging from university laboratories are licensed to start-ups, often formed by the researchers themselves, and these fledgling businesses may lack survival skills for the world of commerce. In the wake of Bayh-Dole, business incubators were formed around the country to assist such start-ups. Unfortunately, the standardized services such incubators offer accounting, legal advice, a fax machine and desk while necessary, may not be adequate to launch breakthrough technologies in fields such as synthetic biology or nanomaterials.
Therefore, a robust innovation ecosystem must provide the financial, infrastructural, and human capital to support the development and exploitation of promising new technologies.
We clearly need a new financial model for start-ups, as venture capitalists increasingly prefer to invest in less risky, later-stage enterprises, and entrepreneurs refer to a widening “valley of death,” when no financing is obtainable. Large corporations, too, are not always willing to fund the development of undergirding technological breakthroughs that offer them no exclusive competitive advantage.
Equally important is the physical capital that allows new technologies to be improved and scaled for the marketplace facilities for applied research, for the prototyping and testing of new technologies, for the development of advanced manufacturing processes for modeling and simulation. Emerging technologies in nanoelectronics or bioengineering tend to demand the kind of computational power, instrumentation, robotics, and clean rooms that no single company can afford.
The question is where should such commercial work occur? Universities have facilities that some companies rely on for early-stage commercial explorations.
Universities cannot be turned into early-stage factories for businesses and move in lock-step with their interests and desires, but they do have a role.
What are the alternatives? We need to develop infrastructure that can be shared by nascent industries as a new kind of capital to undergird innovation. Such infrastructure could be based at a university with appropriate safeguards to exploit intellectual property coming out of university research, to support applied research, or to provide critical infrastructure for small and large companies.
The Computational Center for Nanotechnology Innovations (CCNI) offers a potential model. As many of you may know, it is located off our main campus, in the Rensselaer Technology Park a business park owned by the university. A joint project of IBM, New York State, and Rensselaer, it not only hosts one of the world’s most powerful university-based supercomputers, it allows companies of all sizes to perform research, and to tap the expertise of Rensselaer scientists who would otherwise be inaccessible.
The CCNI also provides immense computational power for Rensselaer faculty use in basic research.
Such physical capital need not be exclusively university-based. It can be in proximity to our national laboratories, or be developed by industry consortia. Sematech, the semiconductor consortium formed in 1987, with the support of the U.S. Department of Defense, offers an interesting model. It was created to enable the development of manufacturing processes crucial to the industry that no single company could finance. That consortium laid out a semiconductor research and development roadmap, which has been followed ever since to position the U.S. as a leader in the semiconductor industry in advanced chip design and manufacturing. The Global 450 consortium formed here in the Capital Region of New York State to develop, prototype, and produce new nanoprocesses and chips for high-end computation is a more recent example.
This consortium also illustrates what is required to rebuild our industrial base, i.e., to further develop manufacturing processes which link information systems to physical processes using sensors, actuators, and other key technologies as part of a broadly intelligent advanced manufacturing system.
Abstracted from this might be joint federal and state-sponsored regional advanced manufacturing centers to drive technology and best practices through the supply chain.
In the longer-term, we even can build upon initiatives such as the road-mapping exercise undertaken by the National Science Board in robotics to identify important cutting-edge technologies, relevant across multiple fields, that show the most promise for evolving manufacturing in key fields, such as health care and energy, and then lay out the facilitating framework for deploying such technologies in the United States.
The most crucial of capital required for our ecosystem is human capital. Clearly, a nationwide crisis in our manufacturing sector, which has lost millions of jobs since 2000, has created a skilled, but underemployed, labor force.
The demands of advanced manufacturing require that every player in the ecosystem universities, government at all levels, and businesses contribute to a comprehensive education and retraining effort for that labor force in new technology development and use.
We also must work together to improve mathematics and science education from the very beginning of our children’s educational careers, if we are to have those who will sustain our innovation ecosystem through the next generations. We also must work harder to retain high caliber talent from abroad, especially those obtaining advanced degrees in science and engineering from American universities. Indeed, there is a global competition for talent.
So, within the innovation ecosystem, all sectors are players. Government, business, and academia must work together. No one of them can do it all alone.
This is why this conference is so important not only because of what each participant here does individually, but because it has brought people together from all three legs of the three-legged stool (government, business, and academia), and is the reason our panels this morning are so important.
Source citations are available from the division of Strategic Communications and External Relations, Rensselaer Polytechnic Institute. Statistical data contained herein were factually accurate at the time it was delivered. Rensselaer Polytechnic Institute assumes no duty to change it to reflect new developments.