Concepts like:
Sum Reduction

CudaStream_t

Considering thrust, the undesirablity was the fact that I had to choose one of two approaches below when using thrust:
1- using keys for a partitioned input volume(like inclusive_scan_by_key).
2- using streams.
while its easier to use, it takes time to create keys or create streams regarding using some of this library, so I used custom defined kernels.

Thank you sincerely for being with me