# Super sparsity

The fraction of distinct nonzero values in a matrix. If a matrix has z nonzeros -i.e. z = sparsity × m × n, then the super-sparsity is number of distinct values divided by z. Although related to the idea of sparsity, a matrix can be highly super-sparse yet not sparse. For example if $LaTeX: \textstyle A_{(i, j)} = 1$ for all $LaTeX: i,j,$ then sparsity is 0 but super-sparsity is $LaTeX: 1/mn.$