map a CPU based C++11 algprithm to Cuda version running on GPU

an algorithm clearly discussed, documented, and implemented with c++11 on CPU and is "short" in terms of code lines. we need to implement this on GPU written with Cuda and compile with nvcc compiler (cuda 11.0 and above).