You are an agent for the office of strategic national alien planning. Your task is to complete projects focusing on code incription, computer optimization and pin collecting. Failure means obliteration of your grade because this is an assignment for one of the University of Oregon’s most innovative classes.
This is an actual assignment in CIS 410/510 Parallel Computing, an experimental course that premiered last spring was brought into existence through technology company Intel, via a $100,000 donation to the University of Oregon’s Computer Information Science department (CIS).
The class focuses on programming for devices, which use multiple processors or cores. The cores determine a computer’s speed and power, according to Daniel Ellsworth, a researcher who was involved with the construction of the class and its implementation. Most devices generally have one core, but using multiple cores in a computer can make it more efficient without the risk of overheating. Each core works together on a single task, the only problem is programing it so it works together.
“Your computer is like a five-year-old. It’s very eager to do what you tell it, but like a five-year-old it isn’t very smart,” said Ellsworth.
Ellsworth explained that parallel programming allows for all these “five-year-olds” to work together on a single task. Professor Allen Malony, the main developer and teacher for the class, describes parallel computing as having to compute the interest in a hundred bank accounts at the same time.
The 25 person class took place in a computer lab that was specifically built for the course, said Malony. The 16 computer stations were built inside a conference room in Straub Hall, a set-up that took over two weeks to construct. The room has since been reverted back to its original form.
Intel recently created what the company calls the “Intel Parallel Computing Centers” in numerous institutions across the United States. These centers are designed with the goal of improving parallel computing technology to improve computers as a whole, according to a press release from Intel.
Assistant CIS Professor Boyana Norris and Hank Childs also facilitated the program’s design and implementation. Norris explained how vital it is for future computer programmers to learn and understand parallel computing. “If you want to be one of the people creating software, you have to take a class like this,” she said.
A textbook on the subject titled Structured Parallel Programing was written by Intel and used in the class. People can visit the class website for more information on the curriculum and class structure.
The course will be made available again to students this coming spring with the hopes of offering it year round in the future.
“This is how you will be programing computers in the future. In some sense, the future is now,” said Malony.
Follow Eric Schucht on Twitter @EricSchucht