Recursive vectorized computation of the vector $p$-norm

Vedran Novaković

Recursive algorithms for computing the Frobenius norm of a real array are proposed, based on hypot, a hypotenuse function. Comparing their relative accuracy bounds with those of th…