To show the advantage of Bitonic SIMD Sorting we compare its profiling against the std::sort()
routine provided as part of the C++-17 standard library. The diagram below captured from Vitis Analyzer shows the average # of cycles for each algorithm. A total of 24 random sorting runs were made. The Bitonic sort took the same time of 170 total cycles per invocation. The std::sort()
cycles were dependent on the particular sort data, taking a minimum of 3,197 cycles, a maximum of 7,061 cycles and an average of 5,073 cycles. This gives Bitonic SIMD sorting an advantage of ~30X over std::sort()
.