As part of its mission to support researchers across Tufts University, UIT Academic Technology (AT) is committed to training today the researchers of tomorrow by providing students access to Tufts’ state-of-the-art research technology infrastructure including Tufts’ high-performance computing research cluster and the Center for Scientific Visualization.
This spring semester, AT had the opportunity to do just that when it co-sponsored with the Tufts Computer Science Department two courses that introduced students to the worlds of high-performance computing and scientific visualization. In both courses, students learned about and used Tufts research technology tools and experienced how research technology applications touch all aspects of life.
The Tufts High-Performance Computing Cluster
High-performance computing (HPC) is a branch of computer science that concentrates on architecting supercomputers and developing software for these parallel platforms to solve complex problems that require large computing power. A main area of this discipline is the development of parallel processing algorithms and associated software: in other words, programs that can be divided into little pieces so that each piece can be executed simultaneously by separate processors that constantly communicate with each other through a high-performance local network.
In his course Introduction to High Performance Computing: Tools and Algorithms, Leopold Grinberg, a Senior Research Associate with the Division of Applied Mathematics at Brown University and a Visiting Lecturer at Tufts, gives students the opportunity to use supercomputers and learn the fundamentals of HPC and parallel computing. A major goal of his course is for students to have sufficient knowledge to understand what a parallel program is doing in order to evaluate it, modify it and create their own models and additions to the software.
During the course, the students were able to use three different HPC clusters: the UIT Research Cluster which Dr. Grinberg found to be a very nice and powerful machine that is an incredible resource for students; Kraken, the third largest HPC cluster in the world, based at the National Institute for Computational Sciences; and Ranger, the ninth largest HPC cluster, at the Texas Advanced Computing Center. Many of the students had no prior experience working with supercomputers, so Dr. Grinberg’s first goal was to get them comfortable with the machines. At the beginning of the course, they went straight from working with a laptop to a supercomputer and by the end of the course they will be applying it to everyday research in a final research project.
Dr. Grinberg points out that his course is not just for computer science people but for anyone who does some aspect of modeling using a mathematical formula. He feels it is extremely important that students understand HPC as its use cuts across all disciplines and fields. Meteorologists use HPC to predict the weather. In neuroscience, HPC helps to map the blood flow in the human vascular network, and in the world of fashion design HPC is being used to show how fabric moves.
Scientific visualization is another important research tool that has wide application. Creative visualizations have been proposed to aid understanding of data, processes, structures, and concepts in a wide range of fields. Well-designed visualizations are powerful tools for communication and support recall, inference, and decision-making. As Sara Su, Visiting Assistant Professor of Computer Science, explains, “visualization creates a visceral, emotional connection to the results and by putting data into a spatial context you are able to convey more information.”
Visting Assistant Professor Sara Su teaching the course on Visualization
Professor Su is teaching a visualization course this semester that introduces students to the fundamentals of visualization and to visualization technologies which includes the visualization walllocated in the Tufts Center for Scientific Visualization at Anderson Hall on the Medford campus. Students in her course are learning the techniques for creating effective visualizations, applying principles of graphic design, data analysis, and cognitive and perceptual psychology.
A major component of the course is a final project in which students work individually or in pairs to identify open visualization problems and develop solutions. These projects can take the form of: prototyping a new software visualization tool; implementing an existing technique and applying it to a new data set; significantly extending a homework assignment; or conducting a user study to evaluate or compare visualization techniques.
Professor Su has been encouraging students to apply visualization techniques to datasets of personal interest and to real problems. She also has been collecting project ideas from Tufts researchers who have need for visualization tools in their work. Project ideas have been proposed that range from A Day in the Life of the MBTA, Visualizing Fluid Motion, to Mapping Company Donations.
Professor Su is very excited by the collaborations that are developing in the class. She has 35 students, both undergraduates and graduates, who come from a broad range of backgrounds – Spanish, Math, Biomedical Engineering and Engineering Psychology. She is finding that the more experienced students are collaborating with undergraduates on the projects and that the students are drawing upon their different backgrounds. The projects will be demonstrated on the last day of class and the presentations are open to the public. The demos will include a combination of posters and the use of the visualization wall. Some of these projects will also be shown at the April Open House that is held each year for newly admitted students to the Tufts School of Arts and Sciences.
Rebecca Sholes, Senior Faculty Development Consultant, UIT Academic Technology.