
Floating-Point Multiplication
435
8
.~<,,~ ~xr %~..
Floating-Point Multiplication
We now consider the algorithm and implementations for floating-point multi-
plication. The algorithm is given in generic terms, whereas the implementations
are geared to the IEEE standard.
Let x and y be the operands represented by (M], Ex) and (My,
Ey),
respec-
tively. The significands are signed and normalized, and the result
z--xxy
8.53
is represented by
(M z, Ez),
where M z is also signed and normalized. The high-
level description of this operation is composed of the following four steps:
1. Multiply significands and add exponents.
M z -- M x x