Control Data Cyber 170 State Hardware Reference Manual page 106

CP Floating-Point Arithmetic Instructions
42ijk Floating double-precision product
of ( ~ j ) and (Xk) t o
X i
This i n s t r u c t i o n reads operands from two X r e g i s t e r s , o p e r a t e s on them t o form
a double-precision, floating-point product, and d e l i v e r s t h e lower h a l f of t h i s
r e s u l t t o a t h i r d X r e g i s t e r . The operands f o r t h i s i n s t r u c t i o n a r e i n
X j
and
Xk.
These operands a r e i n f l o a t i n g - p o i n t format and a r e not n e c e s s a r i l y
normalized. The lower h a l f of the double-precision product i s d e l i v e r e d t o
X i
i n floating-point format and i s not n e c e s s a r i l y normalized.
The operands a r e not rounded i n t h i s operation. The two operands a r e unpacked
from floating-point format. The exponents a r e added t o determine t h e exponent
f o r t h e r e s u l t . The r e s u l t exponent i s e x a c t l y 48 l e s s than t h e exponent f o r
a
40 i n s t r u c t i o n . The c o e f f i c i e n t s a r e m u l t i p l i e d a s signed i n t e g e r s t o form a
96-bit i n t e g e r product. The lower h a l f of t h i s product i s e x t r a c t e d t o form
the c o e f f i c i e n t f o r t h e r e s u l t . I f t h e o r i g i n a l operands a r e normalized and
t h e double-precision product has only 95 s i g n i f i c a n t b i t s ,
a
l - b i t l e f t s h i f t
i s done t o normalize t h e r e s u l t c o e f f i c i e n t .
The r e s u l t i n g exponent i s reduced
by one count i n t h i s case.
I f both operands a r e not normalized, t h e r e s u l t i n g double-precision product has
l e s s than 96 s i g r d f i c a n t b i t s .
N o t e s t is made f o r t h e p o s i t i o n of t h e most-
s i g n i f i c a n t b i t . The lower 48 b i t s a r e always read from t h e 96-bit product
r e g i s t e r .
This i n s t r u c t i o n i s used i n multiple-precision, f l o a t i n g - p o i n t c a l c u l a t i o n s .
This i n s t r u c t i o n a l s o provides f o r i n t e g e r m u l t i p l i c a t i o n c a p a b i l i t i e s where
both operands have an exponent value of plus o r minus zero, and n e i t h e r
c o e f f i c i e n t has been normalized. The i n t e g e r r e s u l t s e n t t o X i i s 48 b i t s w i t h
60-bit s i g n extension. I f t h e r e s u l t exceeds 48 b i t s , t h e hardware does not
d e t e c t an overflow. A n overflow check can be made by executing a 40
i n s t r u c t i o n using t h e same two operands.
I f t h e r e s u l t i s nonzero, overflow i s
then i n d i c a t e d . A n i n t e g e r multiply o p e r a t i o n i s not intended f o r use with
normalized operands. I n f i n i t e (3777xxx
...
x and 4000xxx
...
x )
o r i n d e f i n i t e
(1777xxx...x
o r 6OOOx xx... x ) operands cause corresponding e x i t c o n d i t i o n s t o
s e t i n t h e CP f o r e x i t mode a c t i o n .
For f u r t h e r information, r e f e r t o Floating-point Arithmetic under CP
Programming i n chapter 5.