Store 
Simulink Fixed Point 6.2
Float- to Fixed-Point Conversion Using Fixed-Point Advisor
Learn how to explore fixed-point models and use
automated logging and scaling tools to refine the fixed-point data
types for more optimized results.
Contents
Demo Overview
The video shows how to apply float- to fixed-point conversions on models created using:- Simulink®
- Stateflow®
- Embedded MATLAB™ functionality
- On-target rapid-prototyping
- Embedded systems
- Mass production
- C
- C++
- Microcontrollers
- Digital Signal Processors
Exploring the Model
The Fixed-Point Advisor helps to reduce your development
time by automating many of the manual steps needed for conversion and
initial scaling. We’ll use, as an example, a floating-point model of a
fault-tolerant fuel control system created using Simulink blocks and
Stateflow charts. We expect some error between the floating- and
fixed-point results after the initial scaling. In the follow-up
recording, we will show how to optimize the fixed-point scaling using a
companion tool called the Fixed-Point Tool.
Preparing Model for Conversion
The first set of checks validates model-wide settings
and creates the floating-point reference data that will be used to
compare the fixed-point results. We highlight one check that finds
blocks that do not support fixed-point data types and gives advice on
correcting this issue.
See
how we check for blocks that do not support fixed-point data types.
Preparing for Data Typing and Scaling
In the second set of checks, we remove data type inheritance so we can specify individual scaling for all blocks later. We also explicitly set the design min and max that are used as an input for the tool to propose the initial scaling.
See how we set the design min and max values.
Performing Data Typing and Scaling
In the third set of checks, we propose and set initial
scaling for individual blocks and resolve any scaling conflicts that
may occur. We also plot the difference plots for the golden reference
floating-point results against the initially scaled fixed-point
results.
See how we set initial scaling and view fixed-point conversion results
graphically.
Preparing for Code Generation
In the last set of checks, the tool helps to identify settings that may lead to non-optimal generated code. We highlight one check that finds blocks with questionable settings such as the unevenly spaced look-up tables that lead to additional code and increased ROM consumption.
See
how we find blocks with questionable settings.
Additional Resources
Advanced Fixed Point Tips Document and Models
Technical Evaluation Kit
Related Video Demos
Related Demos
Products Used in this Demo
- MATLAB®
- Simulink
- Stateflow
- Simulink Fixed Point
- Real-Time Workshop
- Real-Time Workshop Embedded Coder
- Stateflow Coder™
