Sorting is one of the most important problems in computer science and the ability to sort large amounts of data efficiently is absolutely critical. Sorting algorithms were traditionally been implemented on CPUs and they work very well there, but on the flipside implementing them on GPUs can be challenging. In the OpenCL programming model, we have both task and data parallelism and getting a sorting algorithm to work on the OpenCL model can be challenging, but mostly from the algorithm point of view, that is, how to create an algorithm that takes advantage of the massive data and task parallelism that OpenCL offers.