Error in ulps of the multiplication or division by a correctly-rounded function or constant in binary floating-point arithmetic.

Nicolas Brisebarre, Jean-Michel Muller, Joris Picot

Browse the full ARITH paper archive.