Write a kernel program for matrix multiplication (C=A*B). Assume that matrices are squared. Each thread in the kernel should calculate two elements of matrix C. For example, if dimension of the matrices is 10*10, then 50 threads are launched. Thread zero should calculate Coo and C0, thread one should calculate Co2 and C03, Assume that only one work-group is launched and the threads within the work-group are organized in one dimension. _kernel matrix_mult(const int Mdim, ____global float* A, __global float* B, _global float* C) { }