Integer Data Formats
The three Integer data formats (byte, word, and long word) are the standard data for-
mats supported in the TS68000 Family architecture. Whenever an integer is used in a
floating-point operation, the integer is automatically converted by the TS68882 to an
extended precision floating-point number before being used. For example, to add an
integer constant of five to the number contained in floating-point data register 3 (FP3),
the following instruction can be used:
FADD.W #5.FP3
The ability to effectively use integers in floating-point operations saves user memory
since an integer representation of a number, if representable, is usually smaller than the
equivalent floating-point representation.
Floating-point Data
Formats
The floating-point data formats single precision (32-bits) and double precision (64-bits)
are as defined by the IEEE standard. These are the main floating-point formats and
should be used for most calculations involving real numbers. Table 8 lists the exponent
and mantissa size for single, double, and extended precision. The exponent is biased,
and the mantissa is in sign and magnitude form. Since single and double precision
require normalized numbers, the most significant bit of the mantissa is implied as one
and is not included, thus giving one extra bit of precision.
Table 8. Exponent and Mantissa Sizes
Data Format
Single
Exponent Bits
Mantissa Bits
23 (+1)
52 (+1)
64
Bias
127
8
Double
11
15
1023
16383
Extended
The extended precision data format is also in conformance with the IEEE standard, but
the standard does not specify this format to the bit level as it does for single and double
precision. The memory format on the TS68882 consists of 96 bits (three long words).
Only 80 bits are actually used, the other 16 bits are for future expandability and for long-
word alignment of floating-point data structures. Extended format has a 15-bit exponent,
a 64-bit mantissa, and a 1-bit mantissa sign.
Extended precision numbers are intended for use as temporary variables, intermediate
values, or in places where extra precision is needed. For example, a compiler might
select extended precision arithmetic for evaluation of the right side of an equation with
mixed sized data and then convert the answer to the data type on the left side of the
equation. It is anticipated that extended precision data will not be stored in large arrays,
due to the amount of memory required by each number.
24
TS68882
2119A–HIREL–04/02
ATMEL [ ATMEL ]
