target.FloatingPointDataType Class

Namespace: target

Describe floating point data type implemented by compiler for target hardware

Since R2021b

Description

Use a target.FloatingPointDataType object to describe a floating point data type that is implemented by the target hardware compiler but is not specified by a C language standard.

To create a target.FloatingPointDataType object, use the target.create function.

The target.FloatingPointDataType class is a handle class.

Properties

Size — Size in bits

0 (default) | integer

Size of the data type, specified in bits.

Attributes:

GetAccess	public
SetAccess	public

TypeName — Data type name

character vector | string scalar

Name of floating point data type.

Attributes:

GetAccess	public
SetAccess	public

LiteralSuffix — Numeric literal suffix

character vector | string scalar

Describe the suffix appended to numeric literal in C or C++ code, which indicates data type. For example, some ARM^® compilers support the use of the literal suffix f16 for the _Float16 data type, which allows the specification of literals of the data type in C or C++ code, say, 1.23f16.

Attributes:

GetAccess	public
SetAccess	public

Includes — Required C/C++ header files

cell of character vectors | string array

For the data type, specify the include files that are required in the C or C++ translation unit.

Attributes:

GetAccess	public
SetAccess	public

SystemIncludes — Required C/C++ system header files

cell of character vectors | string array

For the data type, specify the system include files that are required in the C or C++ translation unit.

Attributes:

GetAccess	public
SetAccess	public

Name — Identifier

string scalar

Identifier for data type description.

Attributes:

GetAccess	public
SetAccess	public

Examples

Specify Half-Precision ARM Data Type

Specify a 16-bit floating point ARM data type.

languageImplementation = target.create('LanguageImplementation', ...
                                       'Name', 'ARM with half', ...
                                       'Copy', 'ARM Compatible-ARM Cortex');
 
customHalf = target.create('FloatingPointDataType', ...
                           'Name', 'ARM Half Float', ...
                           'TypeName', '_Float16', ...
                           'LiteralSuffix', 'f16', ...
                           'Size', 16, ...
                           'SystemIncludes', "arm_fp16.h");
languageImplementation.DataTypes.NonStandardDataTypes = [customHalf];

target.add(languageImplementation);

Version History

Introduced in R2021b