An auto-tuning framework for parallel multicore stencil computations