Multiplication of n*n matrix takes order of n^3 with the standard multiplication algorithm. We can improve the time component of matrix multiplication to order of n^w where w= log(7)/log(2).
Using Strassen (divide and conquer) algorithm we can improve the time to the exponent of w<=2.38. But this algorithm has a large constant and is not used in practice.
How can we implement a matrix multiplication algorithm that has a good performance for n*n matrix for 10<=n<=1000 by reducing the data stucture of divide and conquer algorithm or by any other improvisation in the divide and conquer algorithm.
