Language barrier: To take advantage of multicore chips, programmers will need fundamentally new software
Wednesday, March 2, 2011 - 15:30
in Mathematics & Economics
For decades, computer scientists tried to develop software that could automatically turn a conventional computer program -- a long sequence of instructions intended to be executed in order -- into a parallel program -- multiple sets of instructions that can be executed at the same time. Now, most agree that that was a forlorn hope: Code that can be parallelized is too hard to recognize, and the means for parallelizing it are too diverse and context-dependent. "If you want to get parallel performance, you have to start writing parallel code," says MIT computer-science professor Saman Amarasinghe. And MIT researchers are investigating a host of techniques to make writing parallel code easier.