DECIMAL
ARITHMETIC
INSTRUCTIONS
INTRODUCTION
DATA FORMATS
Packed Format
Zoned
Format
Description
of
Formats
• Decimal
arithmetic
is
performed
on
data
in packed
format.
In
this
format,
two decimal digits
are
placed in one
byte
(four
bits
each).
The
operands
may
be variable
in
length,
and
must
contain a sign
in
the
right-
most
four
bits.
All decimal
instructions
are
two-address, SS-type
format.
The
instruc-
tion
set
includes addition, subtraction, comparison, multiplication,
and
divi-
sion. Since
data
sent
to,
and
from,
external
devices
are
usually
in
zoned
(unpacked)
format
(one digit in one
byte),
there
are
also
instructions
for
converting
to,
and
from, packed
and
zoned
format.
All decimal
arithmetic
instructions
are
standard
features
of
the
70/46
Processor.
•
The
formats
for
decimal
data
in
high-speed
memory
are:
Byte
Byte Byte
Digit
:
Digit
Digit:
Digit
Digit:
Digit
I I I
Byte
I
Digit
I
Digit
I
Byte
I
Digit
I
Digit
I
Byte
I
Digit
:
Sign
In
packed
format,
one byte
represents
two decimal digits. The
right-
most
half-byte (4
bits)
of
a field
represents
the
sign.
Byte
Byte
I
Zone:
Digit
Byte
I
Zone
I
Digit
I
Byte
Byte
Byte
I
Zone
I
Digit
In
zoned
format,
the
low-order
four
bits
of
each eight-bit
byte
contain
the
decimal digit
and
the
high-order
four
bits
contain
the
zone. The
high-order
four
bits
of
the
rightmost
byte
of
a field contain
the
sign
of
the
field.
• Decimal
arithmetic
instructions
operate
from
right
to left.
The
addresses
specify
the
leftmost
byte
of
the
operand,
and
the
length specifies
the
addi-
tional
number
of
bytes
that
are
to
the
right
of
the
addressed byte.
The
fields specified
by
the
addresses can be variable in length
beginning
at
any
byte
in
main
memory
and
consisting
of
from
1 to 16 eight-bit bytes. Results
of
operations
are
always placed in
the
first
operand
field. The
result
never
exceeds
the
limits
set
by
the
address
and
length
of
the
first
operand
field.
If
a decimal
arithmetic
operation
results
in a
carry
outside
the
operand
limits, a decimal overflow
interrupt
occurs.
If
the
first
operand
is longer
than
the
second,
the
second
operand
is extended
with
high-order
zeros
up
to
the
length
of
the
first
operand
during
operation
execution
(in
addition
and
subtraction
only).
This extension
never
changes
main
memory.
Because
the
code configurations
of
digits
and
sign
are
verified while
arithmetic
operations
are
performed,
improper
overlapping
of
fields is
recognized
as
a
data
error.
The
arithmetic
instruction
set
(except
Pack,
Unpack, Move
with
Offset) should
not
specify
overlapping
fields unless
the
rightmost
byte
of
the
fields coincide.
In
the
move-type
instructions
of
this
set
(Pack,
Unpack, Move
with
Offset), no checking is made
for
valid codes. Consequently,
overlapping
is
permitted
without
any
restrictions. (Although unusual
results
are
possible,
overlapping is dangerous.)
154
To Next Page
Loading...
Need help?
General
