Documentation

coder.FixptConfig class

Package: coder

Floating-point to fixed-point conversion configuration object

Description

A coder.FixptConfig object contains the configuration parameters that the fiaccel function requires to convert floating-point MATLAB® code to fixed-point MATLAB code. Use the -float2fixed option to pass this object to the fiaccel function.

Construction

fixptcfg = coder.config('fixpt') creates a coder.FixptConfig object for floating-point to fixed-point conversion.

Properties

ComputeDerivedRanges

Enable derived range analysis.

Values: true|false (default)

ComputeSimulationRanges

Enable collection and reporting of simulation range data. If you need to run a long simulation to cover the complete dynamic range of your design, consider disabling simulation range collection and running derived range analysis instead.

Values: true (default)|false

DefaultFractionLength

Default fixed-point fraction length.

Values: 4 (default) | positive integer

DefaultSignedness

Default signedness of variables in the generated code.

Values: 'Automatic' (default) | 'Signed' | 'Unsigned'

DefaultWordLength

Default fixed-point word length.

Values: 14 (default) | positive integer

DetectFixptOverflows

Enable detection of overflows using scaled doubles.

Values: true| false (default)

fimath

fimath properties to use for conversion.

Values: fimath('RoundingMethod', 'Floor', 'OverflowAction', 'Wrap', 'ProductMode', 'FullPrecision', 'SumMode', 'FullPrecision') (default) | string

FixPtFileNameSuffix

Suffix for fixed-point file names.

Values: '_fixpt' | string

LaunchNumericTypesReport

View the numeric types report after the software has proposed fixed-point types.

Values: true (default) | false

LogIOForComparisonPlotting

Enable simulation data logging to plot the data differences introduced by fixed-point conversion.

Values: true (default) | false

OptimizeWholeNumber

Optimize the word lengths of variables whose simulation min/max logs indicate that they are always whole numbers.

Values: true (default) | false

PlotFunction

Name of function to use for comparison plots.

LogIOForComparisonPlotting must be set to true to enable comparison plotting. This option takes precedence over PlotWithSimulationDataInspector.

The plot function should accept three inputs:

  • A structure that holds the name of the variable and the function that uses it.

  • A cell array to hold the logged floating-point values for the variable.

  • A cell array to hold the logged values for the variable after fixed-point conversion.

Values: '' (default) | string

PlotWithSimulationDataInspector

Use Simulation Data Inspector for comparison plots.

LogIOForComparisonPlotting must be set to true to enable comparison plotting. The PlotFunction option takes precedence over PlotWithSimulationDataInspector.

Values: true| false (default)

ProposeFractionLengthsForDefaultWordLength

Propose fixed-point types based on DefaultWordLength.

Values: true (default) | false

ProposeTargetContainerTypes

By default (false), propose data types with the minimum word length needed to represent the value. When set to true, propose data type with the smallest word length that can represent the range and is suitable for C code generation ( 8,16,32, 64 … ). For example, for a variable with range [0..7], propose a word length of 8 rather than 3.

Values: true| false (default)

ProposeWordLengthsForDefaultFractionLength

Propose fixed-point types based on DefaultFractionLength.

Values: false (default) | true

ProposeTypesUsing

Propose data types based on simulation range data, derived ranges, or both.

Values: 'BothSimulationAndDerivedRanges' (default) | 'SimulationRanges'|'DerivedRanges'

SafetyMargin

Safety margin percentage by which to increase the simulation range when proposing fixed-point types. The specified safety margin must be a real number greater than -100.

Values: 0 (default) | double

StaticAnalysisQuickMode

Perform faster static analysis.

Values: true | false (default)

StaticAnalysisTimeoutMinutes

Abort analysis if timeout is reached.

Values: '' (default) | positive integer

TestBenchName

Test bench function name or names, specified as a string or cell array of strings. You must specify at least one test bench.

If you do not explicitly specify input parameter data types, the conversion uses the first test bench function to infer these data types.

Values: '' (default) | string | cell array of strings

TestNumerics

Enable numerics testing.

Values: true| false (default)

Methods

addApproximationReplace floating-point function with lookup table during fixed-point conversion
addDesignRangeSpecificationAdd design range specification to parameter
addFunctionReplacementReplace floating-point function with fixed-point function during fixed-point conversion
clearDesignRangeSpecificationsClear all design range specifications
getDesignRangeSpecificationGet design range specifications for parameter
hasDesignRangeSpecificationDetermine whether parameter has design range
removeDesignRangeSpecificationRemove design range specification from parameter

Examples

expand all

Create a coder.FixptConfig object, fixptcfg, with default settings.

fixptcfg = coder.config('fixpt');

Set the test bench name. In this example, the test bench function name is dti_test. The conversion process uses the test bench to infer input data types and collect simulation range data.

fixptcfg.TestBenchName = 'dti_test';

Select to propose data types based on simulation ranges only. By default, proposed types are based on both simulation and derived ranges.

fixptcfg.ProposeTypesUsing = 'SimulationRanges';

Convert a floating-point MATLAB function to fixed-point MATLAB code. In this example, the MATLAB function name is dti.

fiaccel -float2fixed fixptcfg dti

Create a coder.FixptConfig object, fixptcfg, with default settings.

fixptcfg = coder.config('fixpt');

Set the name of the test bench to use to infer input data types. In this example, the test bench function name is dti_test. The conversion process uses the test bench to infer input data types.

fixptcfg.TestBenchName = 'dti_test';

Select to propose data types based on derived ranges.

fixptcfg.ProposeTypesUsing = 'DerivedRanges';
fixptcfg.ComputeDerivedRanges = true;

Add design ranges. In this example, the dti function has one scalar double input, u_in. Set the design minimum value for u_in to -1 and the design maximum to 1.

fixptcfg.addDesignRangeSpecification('dti', 'u_in', -1.0, 1.0);

Convert the floating-point MATLAB function, dti, to fixed-point MATLAB code.

fiaccel -float2fixed fixptcfg  dti

When you select to detect potential overflows, fiaccel generates a scaled double version of the generated fixed-point MEX function. Scaled doubles store their data in double-precision floating-point, so they carry out arithmetic in full range. They also retain their fixed-point settings, so they are able to report when a computation goes out of the range of the fixed-point type.

Create a coder.FixptConfig object, fixptcfg, with default settings.

fixptcfg = coder.config('fixpt');

Set the test bench name. In this example, the test bench function name is dti_test.

fixptcfg.TestBenchName = 'dti_test';

Enable numerics testing with overflow detection.

fixptcfg.TestNumerics = true;
fixptcfg.DetectFixptOverflows = true;

Convert a floating-point MATLAB function to fixed-point MATLAB code. In this example, the MATLAB function name is dti.

fiaccel -float2fixed fixptcfg dti

Alternatives

You can convert floating-point MATLAB code to fixed-point code using the Fixed-Point Converter app. Open the app using one of these methods:

  • On the Apps tab, in the Code Generation section, click Fixed-Point Converter.

  • Use the fixedPointConverter command.

Was this topic helpful?