Copyright (c) 1993, Memorial University of Newfoundland. Do not use in whole or in part without permission. Postscript version available by anonymous ftp as follows: > ftp triton.cs.mun.ca > cd /pub/miminis > get munsup.ps. For further information, contact: george@garfield.cs.mun.ca or gayle@morgan.ucs.mun.ca High Performance Computing (HPC) Technology in Education An evaluation of needs, status and potential for Memorial University of Newfoundland, & potential benefits for education in Newfoundland March, 1993 A report prepared for Dr. Kevin Keough, Vice-President (Research), Memorial University of Newfoundland, by George Miminis, Department of Computer Science and Gayle Tapper, Department of Computing and Communications. PART I Executive Summary 1. Introduction 2. Use, need and support for HPC in education at Memorial University 3. High Performance Computing in K-12 4. Implementation Appendix A: Process of research for this report and list of references Appendix B: Grand Challenges and problem areas well suited to HPC technology Appendix C: Definition of terms PART II (separate document) Letters and correspondence relating to the development of this report (MUN internal document) Executive Summary Use of High Performance Computing technologies in education is a desirable step for our university if we wish to produce top-quality graduates and for our province to enhance the educational process, to encourage young people to pursue careers in science and technology, and to support those career choices. There is potential for benefit at the graduate and undergraduate levels at Memorial University, at other post-secondary educational facilities, within the K-12 sector and for continuing education and evolution of the province's workforce. It is a well-established premise that our province must diversify its economy. HPC technologies can contribute to the shift in the Newfoundland economy as our province diversifies from the traditional resource-based economy to one that is science- and knowledge-based. Many reports that are directly relevant to education in Newfoundland strongly emphasize that the well-being of our future economy will depend upon a workforce trained in information technologies, science and mathematics. If we accept this premise, then use of leading-edge computing technology represents the pinnacle of efforts towards achieving such goals. Supercomputing can inspire and capture the imagination and increase understanding. What could be a better way to enhance the educational process? This report concludes that the establishment of a Centre for High Performance Computing as a distinct entity, with part of its specific mandate to support educational integration, promote awareness, facilitate a program in computational science and to have outreach to the province's educational sector - K-12 teachers and students, would have substantial, positive impact on education at Memorial and in our province. In fact, negligence to secure and use such a valuable tool might contribute to our falling behind other institutions, and loss of some of our best students, researchers and educators, and less ability to attract others. This report further concludes that Memorial should actively seek and promote use of HPC technologies in education, and that the primary funding source for this should be federal/provincial, both for new initiatives and on an on-going basis. A national commitment to science and technology applications and education requires vision, and the determination to cause positive action through funding and program creation. Canada currently lacks this political drive and coordination, although there is tremendous amount of talent, interest and vision from some individuals across the country. It is more important now than ever that we be on the cutting edge of scientific advancement, while we have the opportunity to be leaders in this field. Provinces, institutions and individuals must urgently take initiatives if we wish to remain scientifically competitive with other countries. HPC will have an increasingly significant impact on research, education, economies and the quality of life as we approach the 21st century. 1. Introduction "It is widely accepted that education is critical to the development of the provincial economy. Therefore the university is considered both by the government and the public at large to be of great importance as part of the province's developmental infrastructure." - Arthur May, President, "Future Focus," 1992 "The transformation from a resource-based and industrial society to one in which information and the ability to use information are the most prized commodities is progressing at an accelerating pace. The consequences of under-education can become only more serious as we move into the twenty-first century... The leading edge of most areas of economic activity, from the fishing industry to the development of computer software, from forestry to the provision of health services, involves he application of science and technology to increase productivity, develop new products, or solve problems. It is now widely acknowledged that knowledge is becoming the most prized commodity in the emerging global economy. Knowledge resides in people, and people are much less mobile than the other means of production. Those societies which are most successful in educating their people to a level which is attractive in the global marketplace are much more likely to be successful economically than those which continue to rely on natural resources or on unsophisticated labour." - Robert Crocker, "Towards an Achieving Society," 1989 "It seems inevitable that the digital world of knowledge must transform the process of education, so that we can produce a scientifically and technologically literate society. Many school children today are totally at home in an interactive computer-based video world. The challenge we face is to create interactive software that is as compelling as video games, but that teaches science, history, music, mathematics, or social values. The National Science Foundation supercomputer centers are engaged in a wide range of education-related activities such as producing just such software. In the future, schools whose personal computers are connected over the network to supercomputers will have `virtual' physics, chemistry, and biology laboratories created out of software in their classrooms." - Larry Smarr, "Supercomputing and the Transformation of Science," 1992 "Technological leadership has become a strategic factor in determining the economic performance of a country, a region or a province. Economic well being and job creation, now and in the future, depend on the growth of knowledge in our institutions and industries and on the application of technology to keep established industries competitive, and create new ones. Canada lags far behind its industrial competitors, and Atlantic Canada and Newfoundland lag behind other provinces in the production and retention of the human resources that will ensure this growth. On a per capita basis, Canada employs approximately half the engineers and scientists of major competitor countries. The windows of time and opportunity are narrowing, and Canada must move quickly to upgrade its skills and capabilities." - David Strong, Oceans 2000 Proposal High-speed networks, visualization, supercomputers, workstations, very large databases or data sets, complex problem solving, computational science and modelling comprise the modern notion of High Performance Computing, or HPC. A set of complex problems requiring at least the amount of computing power available today, and even more (such as environmental modelling or drug design) are known as "Grand Challenges." Leading edge computing technologies can help students to understand scientific process, and encourage them to pursue further education in science and technology fields. Most educators are not familiar with HPC and therefore they do not understand how using supercomputing technology can improve education. We must overcome our narrow view of how technology can improve educational processes and understanding. We must gain new vision of the possibilities, and aggressively take action. Support for use of HPC technology in research, and the value of computational science (see Appendix C) are made clear in the Information Technology Plan for Memorial University. The Gartner Group report [6] states that "In future years, computational science is virtually certain to be increasingly critical to national competitiveness and quality of life." Memorial's IT Plan reflects a trend in universities for greater need to access HPC, and by many governments to make a commitment to HPC in support of scientific competitiveness. Memorial University must facilitate growth in this important scientific method. Leading edge scientific exploration is becoming increasingly dependant on the latest computing technologies. For example, computer simulation is a tool used to predict the influence of various factors on a data set through a time frame, and often uses visualization methods to help us understand what happens with this data set. An example of this is the modelling of the formation of a storm cloud. Such models can be very complex and thus heavily dependant on the power of supercomputing. The education of young people in science is a reflection of how we do our science. Let us compare the teaching of medicine with the teaching of science in general. Educators in medicine certainly do not introduce archaic tools to their students for practical use. A trained physician must understand use of the most modern medical devices. Similarly, we must introduce science students to the latest "tools of the trade." In the case of scientific computing, keeping pace with the latest tools of the trade is conceptually more difficult, because the pace of change in computing technology has been so rapid. Consequently, more aggressive steps are needed to integrate the technology than to simply let experience and expertise flow from one generation to the next. Computing technology has had a tremendous rate of development, so that, by today's standards, we might consider 5- or 10-year-old technology to be "archaic," such as an IBM PC-XT or a Cray X-MP. Supercomputing power has become cheaper and more available, and with this transition has come many new applications, the largest of which are known as Grand Challenges (see Appendix B). Many of the Grand Challenge problems are of particular interest to us. The Grand Challenges have driven the quest for a great deal more computing power than even that which is available today. By the year 2000, a typical supercomputer user will have a trillion times the computing power than what was available in the middle of this century, according to Larry Smarr in his address to Supercomputing '92. Cost per peak megaflop in 1980 was $100,000, while in the year 2000 it is forecast to be less than $100 [6]. Many countries have new strategies for high-performance computing and see the use of supercomputing as critical to remaining competitive in science and technology. The United States has taken aggressive action, which has been bolstered by the $5 billion High Performance Computing and Communications Initiative (HPCCI), a bill sponsored by Senator Al Gore. In a report prepared for the U.S. Department of Energy and Los Alamos National Laboratory (1991), the Gartner Group stated "The proposed Federal HPCC Program will increase the Gross National Product (GNP) of the United States by $172.5 to $502.6 billion over the next decade." Because of a perceived need for greater ubiquity of HPC technology, the National Science Foundation supercomputer sites and the National Labs in the U.S. have a plan to link their sites across the country, creating a "National Metacomputer Centre" - connected with high-speed networks, the "MetaCenter" will place a virtual supercomputer on the desktop of researchers around the country. The European community has strategic goals regarding HPC as do a number of other countries. In Canada, a number of provinces have on-going initiatives to address high performance computing, including major financial commitments in Ontario, Alberta, Quebec and B.C. There are also initiatives underway or interest expressed in New Brunswick and Nova Scotia. Although NSERC recently made a financial commitment in support of established supercomputer centres, there is NO coordination or strategic planning on a Canadian scale, and NO national commitment to HPC in education. The amount of attention that Canada has given to a national HPC strategy is abysmal, and with respect to HPC in education, Canada's record is non-existent. We are just now at the threshold of true empowerment of individuals by technology, from the microcomputer level to high-speed networks and the largest supercomputers. This is a very exciting and historical threshold. Perhaps the challenge to integrate this technology, spur research, and develop new scientists, is the "Grandest" Challenge for us here in Newfoundland. 2. Use, need and support for HPC in education at Memorial University "There is virtually no field of science or engineering that will be untouched by breakthrough advances in supercomputing." - Larry Smarr, Keynote address to Supercomputing '92 There are 4 primary aspects in the consideration for use of HPC in education at our university or other institutions of higher education: 1. Development of HPC use by graduate students in their study and research. Graduate students in sciences that depend on computational resources require access to system resources in the same way as researchers. As stated elsewhere, HPC is increasingly important to research activities. 2. Use of HPC in the current Department of Computer Science curriculum for academic courses, both graduate and undergraduate. Access to different architectures is always of interest. Specific examples of courses that could use HPC are outlined under the section "Current activities at Memorial University and summary of input received," and in Part II. 3. Development of a multi-disciplinary program of study in computational science. Computational science and computer simulation are modern tools of scientific investigation, but have limited results without the support of supercomputing. These tools increase scientific understanding at both the graduate and undergraduate level. At the graduate level, access to these tools may be a vital component of research and study. This is actually the focal point of a report on "Computational Science Education" by Charles Swanson [12], who states that "Computer simulation has been established as a basic methodology of doing scientific research, joining theory and experimentation. Evidence for this is given by the proliferation of supercomputers throughout government, academic, and industrial organizations. It is appropriate, therefore, that the knowledge and techniques required to perform computer simulations, i.e. computational science, be taught in our colleges and universities." Computational Science (or Scientific Computing) is the application of computing technology to scientific problem solving. There are currently a small number of universities that offer graduate programs in computational science. Memorial has an opportunity to be a leader in the study of computational science, by integrating these techniques into existing programs, or by creating a new, inter-disciplinary program in computational science. There certainly exists at Memorial both the talent and the enthusiasm necessary to implement such a program. 4. Support of existing use of HPC in sciences and specific courses which use the technology heavily at present to enhance and enrich the undergraduate programs, and helping faculty members understand how this technology could benefit their courses and how they can develop curricula incorporating HPC. The faculties/departments which could derive the greatest benefit from use of HPC technology in education are Chemistry, Physics, Computer Science, Earth Science, Engineering, Medicine, Biochemistry and Mathematics. Use of HPC in education has historically taken a back-seat to research activities surrounding equipment. However needs have been perceived and are being addressed by several individuals and departments at Memorial University. There is substantial support for the idea of coordination of HPC educational activities, actively pursued by individuals with immediate need, those with vision for the potential and who wish to keep Memorial's graduates top-calibre in their fields of study. Curriculum development can be supported by the implementation of special programs, exposing faculty to the applications of supercomputing to their disciplines. Some examples of special programs with this purpose come from the National Science Foundation sites in the United States. "Supercomputing and Undergraduate Education" was the subject of a workshop sponsored by the National Science Foundation in 1992. The rationale for this program was to help faculty to "introduce undergraduates to the latest computing technology, prepare them for graduate study, and give them the formal background to make decisions in our technological world." SDSC began a program in 1991 to foster undergraduate curriculum development in advanced computing. The National Center for Supercomputing Applications at the University of Illinois, Urbana-Champaign conducts a program of Research Experiences in Computational Science for undergraduate students. The authors encountered varied levels of support for the use of HPC in education around the university. While some expressed keen interest and need, others had not given any thought to the idea, or felt that their needs were satisfied. Most were astonished at the notion of using supercomputers in the K-12 sector. In this regard, the process of generating this report has helped to create a broader base of awareness on our campus. In a world where people are very busy, information is drawn from various sources - conferences, journals, newsletters, awareness campaigns, etc. Because of interest in the subject of HPC in education, and a mandate from the Vice-President (Research), the authors of this report have sought out information, thus become aware of many things which helped to form the report. Awareness and information must be actively delivered, or real progress will only be made by a small handful of "pioneers." Awareness and information distribution would be one of the most valuable benefits of coordinating efforts to acquire HPC on an institutional, provincial or national scale. Current activities at Memorial University and Summary of Input Received The following text summarizes input received from around the university, either by way of verbal communications, or in response to a letter inviting input into this process, which was distributed to all deans, directors and department heads. Chemistry The Chemistry Department is currently addressing access to leading edge computing technologies. The need for understanding the applications of modern computing is the rationale behind their proposal "Chemistry Calculations Center - A new initiative in chemistry teaching." In this proposal, it is stated that "excellent computational facilities will attract both undergraduate and graduate students. This initiative will permit MUN Chemistry to be a leader in the integration of computer skills into the undergraduate and graduate curriculum." This centre, comprised of Pcs networked into workstation servers and the campus VAXen is seen as the first step to fully integrate computer applications into their undergraduate and graduate programs. "Access to HPC on the MUN campus would establish MUN Chemistry as a world center for computational chemistry. Graduates would readily find employment and some would be sufficiently skilled to establish companies to market their skills. Students and faculty would be attracted to MUN. Bright students would not have to move away to learn or to work at the cutting edge of technology." (submission from Drs. M. Brooker and R. Poirier, Coordinators, Chemistry Computer Initiative.) Memorial University has the potential to develop an outstanding program in computational chemistry where use of computational science as a method of discovery is an integral component of the program. Computer Science A number of faculty members in the Department of Computer Science have expressed interest in high-performance computing technology in order to deal with courses in image processing, scene analysis, numerical methods, computer vision, parallel machine architectures, VLSI and graphics, and feel that new courses could be offered if facilities were available. There were a variety of responses received from individual faculty members in the Department of Computer Science. A course developed and taught by Miminis (co- author of this report), entitled "Numerical Algorithms for Supercomputers" explored how to use different supercomputer architectures to solve numerical problems. In order to be able to give this course, Miminis participated in a course at the Argonne National Laboratory (Argonne, IL) on how to use their high performance computing facilities. Because of problems in network access and support requirements it became impractical for students of this course to use that remote installation. Problems that one can run into are: bandwidth and support. According to Pearce [10], this course was one of the earliest undergraduate courses of its kind in Canada. Physics The Department of Physics is also working actively to increase the level of competency and exposure to computing technologies for its students, and use of workstation-level systems and high end scientific computing is regarded as a natural and essential part of doing physics. The Physics Department is actively working to expose students to network and workstation-based computing, and the revised undergraduate curriculum requires at least one course in Numerical Methods (CS3731, where supercomputing concepts are introduced.) "The primary purpose of the development of a departmental computing network, primarily for undergraduate use, is to provide our students with state of the art computational facilities. The revised curriculum requires that they utilise them effectively throughout their undergraduate and graduate training. Given that budgetary constraints constitute the primary impediment to accomplishing this goal, the department would therefore welcome any initiative by the University that would assist the department in achieving this end." (from submission by Dr. J.P.Whitehead, Chairman, Computer Affairs Committee, Physics.) Graduate Studies The idea of an initiative in high-performance computing technology in education was heartily endorsed by the Dean of Graduate Studies, John Malpas, in his response our letter inviting input: "I do wish to add weight... to your argument that access to leading-edge computing technology must be available for graduate students as a research tool." Graduate research in a variety of disciplines could benefit from access to high performance computing, as discussed elsewhere. There were also responses received from Sir Wilfred Grenfell College (SWGC), Nursing, Political Science and Medicine. The full contents of all responses is contained in Part II of this report. 3. High Performance Computing in K-12 "It can take ten years or more for new ideas in science to be incorporated into the curriculum of schools. Given the rapid advances in science, new ways must be found to link researchers, educators and learners. The technologies that make possible the accelerated pace of scientific advances are also a key to creating the needed linkages with education." [4] "Integrating high performance computing into the curriculum of public schools will be America's vital link to a successful, competitive future." - Allan Bromley, Director of the Office of Science and Technology Policy Keynote Address to Supercomputing '91 There is a vision in the United States that HPC can be used by children and educators in the K-12 sector with the following goals:  understanding scientific endeavour and process.  inspire the minds of young people, attracting them to careers in science and technology. The rationale for programs in the U.S. is twofold:  low performance and interest in math/science in young people.  future shortage of graduates in science and technology. This has led to a number of specific educational programs to introduce supercomputing in the K-12 sector evolving in the United States, some of which are outlined in the subsection entitled "Examples of special programs in K-12" below. These are familiar echoes in our own province. The Crocker Report [1] makes a number of recommendations that could be addressed in part by an outreach program for Supercomputing in K-12. As stated in the Summary Report, "the most general conclusion reached ... is that the educational system is in the midst of a crisis of low expectations. [...] Unfortunately, a decrease in expectations has occurred just at a time when education is becoming more crucial than ever to the economic and social well-being of the province and the nation." The need to address such issues is in part, the rationale behind the development of STEM~Net (Science, Technology Education and Mathematics network), a province-wide educators' network . Development of STEM~Net is but one link in a chain. STEM~Net can help to deliver programs for innovation in the use of HPC in education to educators around the province. Through discussion with the author of the STEM~Net proposal, there is agreement that a program for HPC in K-12 and STEM~Net could augment the functions of each. For example, geography classes distributed around the province could gather data at their local sites and share findings over the network. By linking to a remote supercomputer, they could get a visual representation of their findings, access other weather data, make predictions, compare historical data and discuss results among themselves. As stated in the STEM~Net proposal, "The importance of education in general, and science, technology, and mathematics education in particular, to the social and economic vigour of individuals and regions has been highlighted in many provincial and national studies and reports in recent years. However, in spite of progress, there are many obstacles that continue to impede educational advancement, and keep this province and its citizens from achieving their educational and economic potential. One of the impediments that bears on many of the problems identified in recent studies and reports is the professional isolation experienced by teachers, and particularly by those in rural and small schools. A second impediment is associated with low levels of teacher experience with technology in general and with computer and communications technologies in particular." The evolution of STEM~Net is a definitive step to help us overcome geographic isolation, and move us more rapidly to the information age. Although this step is bold and visionary, it is certainly not radical - integration of networking and information infrastructure into the K-12 sector, and the subsequent cultural and educational integration is on going at many places around North America and around the world. The goals we must set for ourselves, as Newfoundlanders, or as Canadians, are scripted in report after report. We refer the reader to the STEM~Net Proposal, Section 2.2, "Education for Personal, Community, Provincial and National Benefit" for an excellent overview of the background that convinces us of the importance of integration of HPC into education. In particular, the reader's attention is drawn to the following quote from that section... "More recently in 1992, [the Canadian] Task for on Challenges in Science, Technology and Related Skills, in a background report to the Prosperity Initiative, titled Prosperity Through Innovation, gave the following within the context of the challenges facing Canadians: Our national prosperity requires people with the ability to change, to adapt and to learn continually throughout their lives. We need to have a flexible population, capable of handling the technological and social aspects of work and life. Canada must prepare for success in a knowledge-intensive economy by designing a strategy to benefit from its most significant competitive advantage: its people. In doing so, it can build on and support the many innovative initiatives under way across the country to strengthen learning institutions, including broadly based efforts to interest youth in technology and sciences." In today's language of technology and global competition, we hear new terms in reference to the availability of technology to our society such as "the have and have-not" of information/network access, or a "crisis of technological illiteracy." Whether we choose to see it or not, there is undoubtedly a challenge in front of us, as educators and administrators, and especially as the guardians of our children's future. The challenge is to invest into the future - to help to build that future, and not to simply be passive recipients of some benefits at some later date. Examples of special programs in K-12 The United States is a good model for use of supercomputing in the educational process, driven by the NSF Supercomputer sites (Cornell, NCSA, Pittsburgh, John von Neumann, San Diego), the DOE Labs and the HPCCI. Technological competitiveness is promoted aggressively in the United States and considered crucial to economic good health. A program called SuperQuest, funded by the National Science Foundation is offered by the National Center for Supercomputing Applications (NCSA), Northwest Regional SuperQuest Centre, San Diego Supercomputer Center, Cornell Theory Center, Sandia National Laboratories & the Alabama Supercomputer Network sites, and funding is provided by NSF for students and teachers to participate in these programs. The SuperQuest program is designed to "put leading edge technology into the hands of tomorrow's scientists today." A brochure from the San Diego Supercomputer Center on the SuperQuest program states "HPCC is the highway to the future." Students use supercomputers to investigate such problems as "when 2 balls collide," wave-tracing, modelling an x-ray telescope and predicting traffic jams. Nora Sabelli, assistant director of education at NCSA states that "One of the problems in science education is that the students are talked to about science. That doesn't teach science... You learn science by doing science. In the next ten years, computer modelling and mathematical experimentation will be an accepted mode of forming scientific hypotheses, along with theory and physical experimentation. The sooner we introduce young people to supercomputers, the better." The San Diego Supercomputer Center also has a program to introduce high performance computing and communications to teachers and students at the primary and secondary levels. Their brochure states "Integrating HPCC into the curriculum of primary and secondary schools is critical for the development of technicians, scientists, and engineers of the future." The U.S. National Energy Research Supercomputer Center offers the "National Education Supercomputer Program," with a mission to help America's teachers motivate and educate the nation's students in Math and Science. EarthVision is a program of the Environmental Protection Association and Saginaw Valley State University for high school teachers and students to "help high schools develop environmental research programs using computational science and access to high speed computers." Alabama Supercomputer Authority offers a training course for secondary and elementary teachers entitled High Performance Computational Methods in Mathematics and Science. In the brochure for this program, the preface states "Computational science is of key importance to the United States if we are to remain world leaders in high technology. One new area of national significance at both the secondary and post-secondary level is the application of high performance computing in mathematics and science." The introduction states "A very common reaction by people is to ask why high schools need access to such powerful and expensive equipment. Can these students be solving problems which are so computationally intensive as to require a Cray? The answer is a profound - Yes!" Pittsburgh Supercomputer Center also has a program, to "introduce teachers to the advantages of supercomputing for student projects and curriculum enhancements, and excite and motivate high school students to pursue careers in science, mathematics and engineering." Organizers have found that both students and teachers have renewed interest in science and mathematics. "How are rainbows formed? Where does garbage go when we throw it away? How does an electric field pull apart a protein?" are the unusual questions this program prompts high school students to ask. The Minnesota Supercomputer Center, Inc.'s SuperTrek program has run for several years. Some of the student projects have included the Big Bang Theory, Twin Primes (pairs of primes with a difference of 2, such as 5 and 7), neural networks, fractals and chaos. The Supercomputer Computations Research Institute (funded by U.S. Dept. of Energy and the state of Florida) sponsors a summer workshop for high school science and math teachers at Florida State University, entitled "The Science Connection -from Supercomputer to PC." The goal of this workshop is "to share with teachers the excitement of using supercomputers in scientific research, and to develop microcomputer-based examples for use in the high school science classroom, by exploring how computers, from supercomputers to Pcs, are used to study real life problems." New Mexico High School Supercomputing Challenge is a program designed with goals to "enhance knowledge, promote careers in science and engineering, encourage academic competition, and to link the business, scientific and education communities in a common endeavour for the benefit of all." A program offered by Cray Computer, called "Cray Academy" operates each summer with the mandate to improve K-12 mathematics, science and technology education. This program has operated for 5 years, and in 1992 had almost 1,200 teachers participating. At Supercomputing '92, there was a series of workshops and sessions especially for high school teachers, including an introduction to climate modelling, ray tracing, computational methods of studying chemical reactions, Internet, parallel computing in K-12 and a visualization lab. Note: We have not gathered sufficient information from other countries to determine if there are similar programs at their national supercomputer centres, although many industrialized countries have formed national strategy for high performance computing. 4. Implementation It is beyond the mandate of this report to suggest a model for implementation, however there are some points that merit consideration. At Memorial University, we have found the following: (i) there is strong support for the concept of using high-end workstations, (ii) there is particular support for the concept of a centre with its own identity, and support staff, (iii) there is active debate over which technology (architecture) would be most desirable (i.e. workstation cluster v.s. massively parallel v.s. vector system, etc.) and (iv) there are serious funding concerns. There is a general wish for coordination of such resources on a campus-wide basis, and frustration at trying to reach perceived levels of adequacy in equipment availability on a departmental scale, and with limited budgets. Individuals, and some whole departments now struggle independently to bring benefits of HPC technologies into the educational process. There would be substantially more benefit if these efforts could be given support, and new initiatives encouraged. From our discussions with others, and studies of sites elsewhere, the following issues have emerged. Inter-provincial, inter-departmental or inter-personal agendas often conflict, and the result is a variety of conflicting opinions, most often underscored by competition for funds, or a lack of overview to a "broad picture." There is agreement that the NSF sites in the U.S. have been very successful, and could serve as a good model for Canada to pursue in delivery of supercomputing on a national scale. Such centres could form part of a national infrastructure in support of R&D. A centre would have a primary thrust of research support. It would therefore be important to clearly establish a mandate for such a centre to support and promote educational activities, awareness and programs. We should seek to create a Centre for High Performance Computing with the following goals:  Research in "Grand Challenge" problems particularly relevant to us, such as ocean population modelling or other cold ocean and environmental studies.  Support access to HPC technology by experienced researchers.  Outreach to MUN faculty and undergraduates to help them understand and integrate the technology, and to other post- secondary institutions.  Outreach to K-12 or other educational sectors in Newfoundland.  Exploration of use of technology by small business, but not for-profit use. While there would be an advantage to dedicating support staff to the integration of HPC in education regardless of where equipment is located, it is more advantageous for us if the equipment is located here. It is our belief that financial support for the idea of integrating HPC technologies into education should come from Memorial University and the provincial and federal governments. The federal government and its agents such as NSERC and NRC in particular, should be encouraged to show vision and leadership in policy planning, and offset burdens of negative publicity, ownership and financial constraint. Memorial University and/or our province would have to compete for federal dollars in this regard. Appendix A Background reports and information The following steps were taken in the preparation of this report: Meetings and discussions with individuals at the Memorial University campus regarding their needs and insights into this issue. Invitation to submit input to all deans, directors and department heads. Attendance of IEEE/ACM Supercomputing '92 Conference. Research, including papers and brochures from particular sites, and electronic news from bulletin boards, Supernet International, and reports including: [1] "Towards an Achieving Society." Task Force on Mathematics and Science Education, Robert Crocker, 1989. [2] A Large Scale Computation Evaluation Study for the Council of Ontario Universities. Gillespie, Folkner & Associates. [3] Information Technology Plan for Memorial University of Newfoundland, 1993. [4] Grand Challenges 1993: High Performance Computing and Communications, FYI 1993 U.S. Research and Development Program Report by the Committee on Physical, Mathematical, and Engineering Sciences Federal Coordinating Council for Science, Engineering, and Technology. To Supplement the President's Fiscal Year 1993 Budget. [5] Report of the ad hoc Committee on Research Computation (CORC), NSERC Report, September 1990, and National Initiative for Scientific Computing (NISC) - A Proposal for Implementation. NSERC, June, 1991. [6] High Performance Computing and Communications: Investment in American Competitiveness. The Gartner Group, Inc. [7] OCEANS 2000 Proposal, Part II. D.F.Strong. [8] Supercomputing and the Transformation of Science (Kaufmann/Smarr). [9] STEM~Net PROPOSAL: A Proposal to Create a Computer Network for K-12 and College Educators in Newfoundland and Labrador. Harvey Weir on behalf of the Vice-President (Academic) of Memorial University of Newfoundland, February, 1993. [10] Feasibility Study of High Performance Computing Facilities in Canada, Elizabeth Pearce, November, 1992. [11] Perspectives on the National Information Infrastructure: CSPP's Vision and Recommendations for Action, publication of the Computer Systems Policy Project, January 12, 1993. [12] Computational Science Education. Charles D. Swanson, Cray Research, Inc., April, 1992. [13] Proposal to establish centralized high-performance computing facility for Memorial University of Newfoundland. Gayle Barton, Department of Computing and Communications, May, 1992. Conversations with Jim Stacey, (Consultant, The Collaboratory), Charlie Folkner (Consultant, Gillespie & Assoc.), Gorden Stokell (Ontario Council of Universities) Lynn Pammett (Hypercomp), Ron Wittig (University of Calgary), Anna Pezacki (Director, HPC, U of T), members of the (former) OCLSC, Harvey Weir (STEM~Net) and others. Appendix B Grand Challenges and problem areas well suited to HPC Applications particularly significant for Memorial University or the province of Newfoundland Ocean sciences, including current modelling and population dynamics (fish stocks) Computational Chemistry Summary of examples outlined under "Need for High Performance Computing," Feasibility Study on High Performance Computing Facilities in Canada, Elizabeth Pearce, 1992. Seismic research Thinning ozone Preservation of wheat and grain against moisture and mould Acid rain and toxic pollution of soil and groundwater Toxic contamination and noise pollution Better treatments and a cure for cancer New drug therapies for asthmatic children, AIDS patients, Alzheimer's victims and others Genetic development of better plant and animal species Better methods of metal forming Analysis for safer aircraft Applications itemized in the Gartner Group, Inc. report "High Performance Computing and Communications: Investment in American Competitiveness." Materials Science Semiconductor Design Vehicle Dynamics Transportation Turbulence Superconductivity Efficiency of Combustion Oil and Gas Recovery Nuclear Fusion Design of Pharmaceuticals Structural Biology Astronomy Quantum Chromodynamics Speech Vision Vehicle Signature Undersea Surveillance Engineering Computational Chemistry Film Animation Bond Bidding Human GenomePrediction of Weather and Global Climate Change Computational Ocean Sciences Appendix C Definition of terms High Performance Computing - This term refers to leading edge computing technologies, particularly those of supercomputing, workstations, algorithms and high-speed networks, and the integration of these technologies for scientific problem solving. Computational Science - application of numerical methods and computer algorithm development and implementation with equipment to scientific problems. The Gartner Group report states "broadly speaking, computational science entails the construction of large, complex computer models of physical systems, then observing the behaviour of the models under various conditions of change. One could think of computational science as the application of computer science to actual scientific problems, and therefore requires and understanding of both the item of investigation and the methodology. Workstation cluster or RISC farm - a group of workstations networked together, which may take advantage of specialized programming methods to work in parallel, and providing attractive price/performance returns for specific computational tasks. FLOP - Floating point instructions per second. (megaflop=billion FLOPs, gigaflop=trillion FLOPs) MPP - Massively Parallel Processing, or many processors in a single computer system (> 1024) Supercomputer - A "supercomputer" is one of a class of computer systems that provides the greatest amount of computational power of computer systems at a discreet point in time. The power of a "supercomputer" of yesterday is quite literally available on the desktop today. A supercomputer is a very expensive piece of equipment, noted by the integration of very-high speed internal communications, heat-dispersal technology, memory size and floating point operations per second. A supercomputer by today's standards would have memory size in the range of 32 megabytes to 8 gigabytes, and be capable of greater than one gigaflop of double precision performance. Some systems exist which are capable of peak (not sustained) performance of a teraflop, and there is general consensus that a sustained teraflop of computing performance will be common by the end of this decade.