The statistical theory of relative errors in floating-point computation

Seems like you haven't provided a review

Don't miss the opportunity to share your thoughts!