Resource Conscious Reuse-Driven Tiling for GPUs
Prashant Singh Rawat, Changwan Hong, M. Ravishankar, Vinod Grover, Louis-Noël Pouchet, Atanas Rountev, Ponnuswamy Sadayappan · 2016
Computations involving successive application of 3D stencil operators are widely used in many application domains, such as image processing, computational electromagnetics, seismic processing, and climate modeling. Enhancement of temporal and spatial locality via tiling is generally required in order to overcome performance bottlenecks due to limited bandwidth to global memory on GPUs. However, the low shared memory capacity on current GPU architectures makes effective tiling for 3D stencils very challenging -- several previous domain-specific compilers for stencils have demonstrated very high performance for 2D stencils, but much lower performance on 3D stencils.