Teach Parallel

This talk will revolve around a small set of "common strategies" (tricks) that practitioners use to achieve regular patterns: tiling, data structure padding, data transposition, data binning, locality based layout, hierarchical data structures, and loop transformations. These tricks manifest themselves differently in different types of applications and different types of hardware architectures. These strategies complement the "dwarfs" view of the world. The tricks often conflict with software engineering practice. I currently cannot see an elegant way of injecting the course into the mainstream undergraduate CS curriculum. However, I have a feeling that these tricks cannot be ignored by mainstream CS education for much longer since they are the reality of effective parallel programming.