FSCALE FSCALE
Scale Exponent
Operation:
FPn x INT(2 Source) ~ FPn
Assembler
FSCALE.<fmt> <ea>,FPn
Syntax: FSCALE.X FPm,FPn
Attributes: Format= (Byte, Word, Long, Single, Double, Extended, Packed)
Description:
Converts the source operand to an integer (if necessary) and adds that
integer to the destination exponent. Stores the result in the destination floatin~q-point
data register. This function has the effect of multiplying the destination by 2~ource,
but is much faster than a multiply operation when the source is an integer value.
The FPCP assumes that the scale factor is an integer value before the operation is
executed. If not, the value is chopped (i.e., rounded using the round-to-zero mode) to
an integer before it is added to the exponent. When the absolute value of the source
operand is (>~) 214, an overflow or underflow always results.
Operation Table:
e In Range Zero Infinity
+ - + -- + --
In Range + Scale Exponent FPn I NAN 2
Zero + + 0.0 + 0.0 NAN2
-0.0 -0.0
Infinity + + inf + inf NAN2
- inf - inf
NOTES:
1. Returns the value FPn before the operation. However, the result is processed by the normal
instruction termination procedure to round it as required. Thus, an underflow and/or inexact
result may occur if the rounding precision has been changed to a smaller size since the FPn
value was loaded.
2. Sets the OPERR bit in the FPSR exception byte.
3. If the source operand is a NAN, refer to 4.5.4 NANs for more information.
Status Register:
Condition Codes:
Affected as described in 4.5.5.1 SETTING FLOATING-POINT
CONDITION CODES
Quotient Byte: Not affected
Exception Byte: BSUN
SNAN
OPERR
OVFL
UNFL
Cleared
Refer to 4.5.4 NANs.
Set if the source operand is (+ or -)infinity;
cleared otherwise
Refer to 6.1.4 Overflow.
Refer to 6.1.5 Underflow.
MC68881/MC68882 USER'S MANUAL
FREESCALE
4-93
To Next Page