Motorola DSP96002 - Manual (Page 350)

A - 162 DSP96002 USER’S MANUAL MOTOROLA
FSCALE.X Scale FSCALE.X
a Floating-Point Operand
Operation:
2
S.H
* D
→
ROUND(SEP)
→
D
(parallel data bus move)
2
nn
* D
→
ROUND(SEP)
→
D
Assembler Syntax:
FSCALE.X S,D
(move syntax - see the MOVE instruction description.)
FSCALE.X #byte,D
Description:
Scale the destination operand D according to the scale factor contained in the 11 LSBs of the high portion
of the source register S, round to single extended precision and store the result in the destination operand
D. An 8-bit Immediate Short scaling factor, sign-extended to 11 bits, may also be used. The scale factor
is a signed 2’s complement 11-bit integer.
As an example of the use of FGETMAN, GETEXP, and FSCALE; consider decomposing a floating-point
number into its mantissa and unbiased exponent and then recreating the original floating-point number.
FGETMAN D0, D1 ;extract normalized mantissa
GETEXP D0,D2 ;extract unbiased exponent
MOVE D2.L, D2.H ;move unbiased exponent
FSCALE.S D2.H, D1 ;scale original mantissa
Input Operand(s) Precision: SEP Floating-Point.
Output Operand Precision: SEP Floating-Point.
CCR Condition Codes:
C - Not affected.
V - Not affected.
Z - Set if result is zero. Cleared otherwise.
N - Set if result is negative. Cleared otherwise.
I - Set if result is infinity. Cleared otherwise.
LR - Not affected.
–
R - Not affected.
A - Not affected.