Richard H. Goodman, Alan Feldstein, Joaquín Bustoz: Relative error in floating-point multiplication. Computing 35(2): 127-139 (1985)