Chapter 2
C Language Reference
2-28
ni.com
language and replace the supplied macros (or functions) with your
(assembly) functions so that you can take full advantage of the processor’s
arithmetic capabilities.
Generated Code with Fixed-Point Variables
Code generated for models using fixed-point variables—such as the
examples provided in this chapter—will differ from code generated for
models using floating-point, integer, or logical signals in the following
areas:
•
Signal and variable type declarations will reflect fixed-point types.
•
Arithmetic operators
+
,
–
,
*
, and
/
will be replaced by fixed-point
arithmetic macro calls (or function calls based on the interface used).
•
Relational operators
>
,
>=
,
<
,
<=
,
==
, and
!=
will, when necessary, be
replaced by fixed-point relational macro calls.
•
Floating-point literals used in SystemBuild models will be replaced by
the scaled integer counterpart.
•
Macros (or procedures) for converting between various fixed-point
types will be invoked when necessary.
Fixed-Point Data Types
Fixed-point type definitions are provided in the system-specific files
src
directory. Files
sa_types.h
,
sa_defn.h
,
sa_fxscale.h
, and
sa_fxlimit.h
use
typedef
statements to represent fixed-point types
and related constants for 8-bit, 16-bit, and 32-bit data. All fixed-point types
have an associated radix position value and a sign (signed or unsigned).
The radix position value is clearly related to the data type scale factor
scale factor =
2
–(radix position)
. To perform any arithmetic operation, both the
value and the radix position scalar are required. The table below lists the
data types generated by AutoCode/C. For information on ranges and
accuracy of each type, refer to the
SystemBuild User Guide
.
Table 2-6.
AutoCode/C Data Types
Data
Type
Number
of Bits
Signed or
Unsigned
Data Type Name
byte
8
unsigned
RT_UBYTE
(radix 00)
RT_UBYTExx (xx = radix 48 to –16)
signed
RT_SBYTE
(radix 00)
RT_SBYTExx (xx = radix 48 to –16)
