Fastest way to transpose a matrix stored in an 1D array in NumPy?

Question

Given a vector v of length N^2 that holds the entries of a NxN matrix M, what is the fastest way to compute the transpose