I'm currently using OpenCL to assign a work item to each element of A, with all the work items iterating over every item in B to calculate the result array (same size as A). In the past, I've had ...