Smashing: Folding Space to Tile through Time

Partial differential equation solvers spend most of their computation time performing nearest neighbor (stencil) computations on grids that model spatial domains. Tiling is all effective performance optimization for improving the data locality and enabling course-grain parallelization for such computations. However, when the domains are periodic, tiling through time is not directly applicable due to wraparound dependencies. It is possible to tile within the spatial domain, but tiling across time (i.e. time skewing) is not legal since no constant skewing call render all loops fully permutable. We introduce a technique called smashing that maps a periodic domain to computer memory withont creating any wrap-aronnd dependencies. For a periodic cylinder domain where time skewing improves performance, the performance of smashing is comparable to another method, circular skewing, which also handles the periodicity of a cylinder. Unlike circular skewing, smashing call remove wrap-aronnd dependencies for an icosahedron model of earth's atmosphere.