Columbus, Ohio -- March 24, 1997 -- Capital University in Columbus and Ohio Northern University in Ada are often recognized for their programs in music, law or a traditional arts program, and...computer science? Now they are. Both universities now offer courses in parallel computing to undergraduates. These types of classes are usually offered only at major research universities.
"Coming to such a small university, I didn't expect to be able to develop parallel code here, let alone run programs on supercomputers," said David Koppenhofer, a senior from Luckey majoring in computer science and mathematics at Capital. "I know that many small universities, or larger ones for that matter, do not have such programs, facilities or curriculums."
Ohio Northern's computer science program is quite advanced, according to Nathaniel E. Baughman, a junior from Dover majoring in computer science and mathematics.
"In fact, Ohio Northern University's Math and Computer Science Department is extraordinarily proficient at offering interesting, cutting-edge classes, in order to give its students exposure to many fields," Baughman said.
Parallel processing is a new form of computing that involves using more than one computer, or processor, to solve a problem. These processors can work on the same instructions simultaneously or execute different instructions, communicating through some type of network.
"It's quicker than doing simulations, which take too much time," said Ignatios Vakalis, instructor of Capital's high performance computing courses. "What takes several minutes to do in a regular computer lab can be done in less than a few seconds on the supercomputers."
Capital received a grant from the National Science Foundation to establish an Advanced Computational Laboratory, which opened in the fall, and to add parallel computing courses to the undergraduate computer science curriculum.
The incentive to fund schools like Capital and ONU comes from their size and student-to-teacher ratios, said Vakalis, associate professor of mathematics and computer science and co-director of the Advanced Computational Laboratory.
Learning parallel processing is important to individuals doing computational research in engineering, computer science, mathematics and other academic areas, because supercomputers can solve more complex problems at a faster rate than personal computers.
"Many courses in the computer science curriculum isolate one area of the discipline, programming, for example," said David Hudak, associate professor of computer science and mathematics and instructor of Ohio Northern's parallel courses. "This course forces students to realize they need to consider the entire system in problem solving-- the problem, algorithm, programming environment, system software and hardware."
But what will these courses do for students in the long run? From the instructors' perspective the courses are necessary if they want to do science in the 21st century. Not only that, but knowledge of these systems is essential for students when they get into the workforce.
Students like Brandy Ruthsatz agree.
"As a future engineer, I feel that I will need these concepts in my line of work," said Ruthsatz, a senior from Castalia majoring in mathematics and computer science at Capital. "I recently took a two-week intensive course in engineering and I was already analyzing how the modeling, for example, could have been done in parallel."
The class projects at both schools required many of the machines at the Ohio Supercomputer Center in Columbus, including the Cray T3D and YMP, IBM SP2, the Convex Exemplar, and the Silicon Graphics Power Challenge.
"I am truly excited about these courses," said M. Jason Smiley, a junior from Hilliard majoring in computer science at Capital. "I feel very privileged to have the opportunity to take them. Another thing I really like is the fact that I have accounts on some of the most sophisticated computers in the world, like the ones at the Ohio Supercomputer Center."
Students at both schools felt they learned so much about high performance computing
because their instructors were so knowledgeable about the subject.
"I was impressed with the amount of material covered in Dr. Hudak's class," said Ohio Northern's Baughman. "Dr. Hudak is amazingly proficient with computer hardware. What's more, he is able to present his knowledge and understanding to the class."
Some students at Capital found their first experience using parallel challenging, but thrilling, and currently are taking a second course in high performance computing.
"Taking this class is almost like looking into the future," Ruthsatz said. "I think that high performance computing is the way of the future."
The response at Ohio Northern is not much different.
"I would recommend the high performance computing class to anyone with serious interest in the field of computer science," Baughman said. "High performance computing is not a fad that will blow over in time."
This is great news to professors like Vakalis and Hudak, who say there is a shortage of students interested in fields that use parallel computing.
"We need more students interested in science, physics and biology because the research they do affects our future," Vakalis said. "These classes teach them how to work together. In other words, we need to have more interdisciplinary education at our colleges and universities."
In the meantime, the success stories at both universities might just be enough incentive for other privates to start offering classes in high performance computing to undergraduates.
The Ohio Supercomputer Center is a state-funded resource serving Ohio's higher education community. The Center facilitates discoveries that enhance Ohio's economic development and supports statewide technological advancement and education. The Center's networking initiative provides Internet access to faculty, staff, and students at Ohio's colleges and universities.