Parallel Cache-Efficient Algorithms On Gpus