Table of Contents
Advertisement
23 22 21 20 19 18 17 16 15 14 13 12 11 10
9
8
7
6
5
4
3
2
1
0
FULL
OPERAND
Hardware
GEN 1
GEN 2
GEN
3
OP
Instruction octal
X
Index word address
OP
000101
010101
100101
*
Relative addressing indicator
Y
Operand
address (a few instructions use Y
as the operand; others
use
Y as the address
of the operand)
x
x
x
x
*
G
K
s
D
I
s
y
G
1
K
I
D
I
G
K
Micro-coded GEN 1 subcommand
Bit position of A-register or length of
shift
GEN 2 subcommand
I/O device address (bits 9, 10, 11
=
7/
8
for high speed
I/
O)
GE/
PAC 4020
Compute
r Instruction Formats
Single-word fixed-point
23
22
0
I
I
~~~~~~~~~~---.~~~~~~~--'
Sign
0=+
1
= -
Data
(Two's complement if negative)
Range: +8
,
388,607 to -8,388,608
Double-word fixed-point
23 22
0
23 22
0
I I
I
r--r~-----.
SIGN
DATA
Range:
+70,368,744,177,633 to -70,368,744,177,664
The
GE/
PAC 4020 computer's floating-point
representation uses a sign and absolute value notation
for the fractional part (mantissa) and an offset ex-
ponent for the characteristic.
This
is a self-
normalizing floating-point, to preserve maximum
precision without extra programming effort.
Two modes of floating-point arithmetic are
available. The single-length mode is adequate
for
· _ .
__
:
the simpler mathematical process and control re-
·
lationships, since it provides
17
bits of precision,
equivalent to about five
decimal digits. The double-
length
mode is
useful
for more sophisticated
models,
linear programs, and other operations involving
large matrices
and provides about
11 digits.
Both
8
. .
modes can be made available in the same central
processor, selectable by the floating mode shift
(FMS) command, though it remains the programmer's
responsibility to use the modes and formats con-
sistently and correctly. Accompanying this quasi
package are fix and float conversion commands
;
they
could be used with an FMS to convert between for-
mats
,
or a new quasi could be written and assigned
one of the unused quasi operation codes.
Single-word floating-point
23
22
Sign of
Fraction
0=+
1
= -
17
Power of two
exponent, +40 8
16
0
Absolute value of
fraction
Range (approximate): ±2.15 x 10 ± 9
Double-word floating-point
23 22
14 13
0
23 22
. . . . . . . . . . , 1...---1 -,
t><1
0
I
'--./~---~~~~~--....,~~~--
Sign of
Power of two
Fraction
fraction
exponent, +400 8
Range
(approximate): ±5.8 x
10
±
76
Advertisement
Table of Contents
