Tuesday, January 24, 2012
IST Lunch Bunch
Gerard Holzmann, Chief Scientist, Laboratory for Reliable Software, JPL/NASA
Within the last few years, multi-core systems have become ubiquitous, as expected. Virtually every desktop and laptop systems sold today has at least a dual-core chip inside, and a graphics card that may have hundreds of powerful processing engines -- all capable of processing massive amounts of data in parallel. If only we could learn how to program these systems well, we could tackle many more highly interesting practical problems. <br><br> One such fundamentally important problem is to develop new highly scalable algorithms for program analysis, but there are many others. Computing science education has traditionally focused only on sequential languages and sequential algorithms, which are increasingly becoming the technologies of the past. In this talk, I'll discuss what we need to do to prepare for our parallel future, and what types of problems we are up against.