A Virginia Tech professor is once again in the national spotlight for his work with supercomputers.
Professor Wu-Chun Feng, director of Tech's Synergy lab, and his team has received a $3.5 million grant to research simulating micro-air vehicles, MAVs, for the U.S. Air Force.
Wu will be using the supercomputers to simulate fluid dynamics of the micro-air vehicles. The computations required are so detailed that the machine being used to do them has to be capable of processing very large amounts of information.
"We have to figure out the right way to take the application, decompose it to its fundamental algorithms, extractions and map them appropriately onto a CPU or GPU," Wu said.
CPUs and GPUs are the key to Wu's success, they are two parts of a computer that his work brings together, tapping into previously untapped potential. According to Wu, GPUs and CPUs essentially mirror a human brain.
"You have two kind of brains: the CPU kinds of brains and the GPU kinds of brains, and they're good at different things — much like our left and right brain," Wu said.
Wu compares CPUs and GPUs to sport utility vehicles and drag racing cars, respectively. The CPUs are able to do a multitude of actions relatively well, but don't have the greatest performance, while GPUs are able to do single tasks with extreme efficiency and speed.
Wu gained fame last year when he was able to build the world's most power efficient supercomputer using the technique of