A Parallel Dual-Pivot QuickSort Algorithm with Lomuto Partition

Sorting is one of the basic problems in computer Science and big data analytics. This paper presents a parallel Dual Pivot QuickSort (PDPSort) for manycore CPU systems. PDPSort makes use of the classic Lomuto partioning algorithm with two pivot values in parallel. It is developed in C++ and linked with OpenMP 4.5. PDPSort is compatible with the Standard Template Library sort (STLSort). The comparison includes run time, Speedup over STLSort and CPU utilization. PDPSort is faster than STLSort by 6.13×,3.54× and 2.99× on an 8-core 16-thread AMD R7-1700, an 8-core 8-thread AMD FX-8320 and a 4-core Intel i7-2600 Linux systems, respectively.