Motorola DSP96002 User Manual page 723

IEEE Double Precision
Using Software Emulation
B.4.1 ADDITION
B.4.2 SUBTRACTION
B.4.3 MULTIPLICATION
Non-IEEE Double Precision
Using Software Emulation
B.5.1 CONVERT NUMBERS:
-TO DPLIB FORMAT
(IEEE2DPLIB)
-TO IEEE FORMAT
(DPLIB2IEEE)
B.5.2 ABSOLUTE VALUE
(DP_ABS)
B.5.3 ADDITION
(DP_ADD)
B.5.4 CLEAR
(DP_CLR)
B.5.5 COMPARE
(DP_CMP)
B.5.6 COPY SIGN
(DP_COPYS)
B.5.7 DIVISION
(DP_DIV)
B.5.8 ROUND TO AN INTEGER
(DP_INT)
B.5.9 MAC
(DP_MAC)
B.5.10 COPY A NUMBER
(DP_MOVE)
B.5.11 MULTIPLICATION
(DP_MPY)
B.5.12 NEGATE
(DP_NEG)
B.5.13 SCALE
(DP_SCALE)
B.5.14 SQUARE ROOT
(DP_SQRT)
B.5.15 SUBTRACTION
(DP_SUB)
B.5.16 TEST A NUMBER
(DP_TST)
Note: typical execution times are the average of all possible
paths through the algorithm.
B-204
TYPICAL
6.86 us
7.01 us
13.58 us
TYPICAL
Figure B-1. Standard Benchmark Summary (continued)
DSP96002 USER'S MANUAL
WORST CASE FULLY TESTED
29.1 us
29.2 us
39.5 us
Instruction Cycles
WORST CASE
23 33
20 23
5 5
69 86
9 9
22 30
13 16
852 1020
16 20
109 149
15 15
109 51
14 14
11 13
1317 1413
103 120
13 14
YES
YES
YES
BEST CASE
6
16
5
22
9
14
11
32
10
61
15
27
14
8
9
56
12
MOTOROLA