2
o
[RR, Short Multiplier and Multiplicand,
Long Product]
8 12 15
ME
[RX, Short Multiplier and Multiplicand,
Long Product]
'7C'
8 12 16
MDR [RR, Long
o 8 12 15
MD [RX, Long
o 8 12 16
MXDR Rt,R:z
o
[RR, Long Multiplier and Multiplicand,
Extended Product]
'27'
MXD
2 ,B:z)
[RX, Long Multiplier and Multiplicand,
Extended Product]
'67'
8 12 16
MXR R
2
[RR, Extended
The normalized product of the second
operand (the multiplier) and the first
operand (the multiplicand) is placed at
the first-operand location.
Multiplication of two floating-point
numbers consists in exponent addition
and fraction multiplication. The oper
ands are first normalized to eliminate
leading hexadecimal zeros. The sum of
the characteristics of the normalized
operands, less 64, is used as the char
acteristic of the intermediate product.
The fraction of the intermediate product
is the exact product of the normalized
operand fractions. When the
intermediate-product fraction has one
leading hexadecimal zero digit, the
fraction is shifted left one digit posi
tion, bringing the contents of the
guard-digit position into the rightmost
position of the result fraction, and the
intermediate-product characteristic is
reduced by one. The fraction is then
truncated to the proper result-fraction
length.
For MER and ME, the
multiplicand fractions have six hexade
cimal digits; the product fraction has
the full 14 digits of the long format,
with the two rightmost fraction digits
always zeros. For MDR and MD, the
multiplier and multiplicand fractions
have 14 digits, and the final product
fraction is truncated to 14 digits. For
MXDR and MXD, the multiplier and multi
plicand fractions have 14 digits, with
the multiplicand occupying the high
order part of the first operand; the
final product fraction contains 28
digits and is an exact product of the
operand fractions. For MXR, the multi
plier and multiplicand fractions have 28
digits, and the final product fraction
is truncated to 28 digits.
An exponent-overflow exception is recog
nized when the characteristic of the
final product would exceed 127 and the
fraction is not zero. The operation is
completed by making the characteristic
128 less than the correct value. If,
for extended results, the low-order
characteristic would also exceed 127,
it, too, is decreased by 128. The
result is normalized, and the sign and
fraction remain correct. A program
interruption for exponent overflow
occurs.
Exponent overflow is not recognized when
the intermediate-product characteristic
is initially 128 but is brought back
within range by normalization.
An exponent-underflow exception exists
when the characteristic of the final
product would be less than zero and the
fraction is not zero. If the exponent
underflow mask bit is one, the operation
is completed by making the character
istic 128 greater than the correct
value, and a program interruption for
exponent underflow occurs. The result
is normalized, and the sign and fraction
remain correct. If the exponent-
Chapter 9. Floating-Point Instructions 9-13