I have two (large) sets of points represented as the columns of two 3xN matrices.

is there an easy way to take each column of the respective matrices so that they are multiplied together by A.col(k)*B.col(k).transpose() and summed which produces a 3x3 matrix that represents the sum of all of these outer products?

Does A * B.transpose() do the trick? If not, I do not understand what you want.

Oh yes, wow. That was a dumb question in retrospect.