Floating-Point Specification in Simulink
Because single precision numbers use only 32 bits, they require less memory than double-precision numbers while still offering many of the benefits of floating point arithmetic. The half-precision data type occupies only 16 bits of memory, but its floating-point representation enables it to handle wider dynamic ranges than integer or fixed-point data types of the same size.
Classes
Simulink.NumericType | Specify floating-point, integer, or fixed-point data type |
target.FloatingPointDataType | Describe floating point data type implemented by compiler for target hardware (Since R2021b) |
Functions
ssGetDataTypeIsDoubleSingleorHalf | Determine whether registered data type is double, single, or half-precision data type |
ssGetDataTypeIsHalfPrecision | Determine whether registered data type is half-precision data type |
ssRegisterDataTypeHalfPrecision | Register half-precision data type and return its data type ID |
target.add | Add target object to internal database |
target.create | Create target object |
target.get | Retrieve target objects from internal database |
target.remove | Remove target object from internal database |
Namespaces
target | Manage target hardware and build tool information |
Topics
- What Is Half Precision?
Introduction to the half-precision data type in MATLAB® and Simulink®.
- The Half-Precision Data Type in Simulink
Simulink software features supported for the half-precision data type.
- Generate Native Half-Precision C Code from Simulink Models
Generate C code for embedded hardware targets that natively supports half precision floating-point data types.
Featured Examples
Half-Precision Field-Oriented Control Algorithm
Compare system response of a FOC in half precision and single precision.
Image Quantization with Half-Precision Data Types
Image quantization using half-precision data types.
Digit Classification with Half-Precision Data Types
Classify images in half-precision.
Convert Single Precision Lookup Table to Half Precision
Halve the memory size of a Lookup Table block while keeping the desired system performance by converting the lookup table to use half precision.
