$1.$Write a CUDA program for finding the maximum value using a single block over an array of size n using reduction algorithm. Give $($i$)$ the kernel call from the host main function $(2$ points$),($ii$)$ skeleton of the device kernel function $(3$ points$),$ and $($iii$)$ give the detailed code for the device function performing reduction employing the CUDA threads. The number of threads would typically be larger than n$,$ but your code should work for smaller n$.$