Fig. 4From: Parallel acceleration of CPU and GPU range queries over large data setsShown here are (a) the reduction pattern used by the COA method and (b) the mapping of thread blocks within thread grids to the WAH query dataBack to article page