MATLAB Examples

# Generate HDL-compatible lookup table function replacements using 'coder.approximate'

This example shows MATLAB code generation from a floating-point MATLAB® design that is not ready for code generation. We use 'coder.approximate' function to generate a lookup table based MATLAB function. This newly generated function is ready for HDL code generation (not shown in this demo).

## Introduction

The MATLAB code used in the example is sigmoid function, which is used for threshold detection and decision making problems. For example neural networks use sigmoid functions with appropriate thresholds to 'train' systems for learning patterns.

## MATLAB Design

```design_name = 'mlhdlc_approximate_sigmoid'; testbench_name = 'mlhdlc_approximate_sigmoid_tb'; ```

Lets look at the Sqrt Design

```dbtype(design_name) ```
```1 function y = mlhdlc_approximate_sigmoid( x ) 2 3 % Copyright 2014-2015 The MathWorks, Inc. 4 5 y = 1./(1+exp(-x)); 6 end ```

## Simulate the Design

It is always a good practice to simulate the design with the testbench prior to code generation to make sure there are no runtime errors.

```mlhdlc_approximate_sigmoid_tb ```

We can use coder.approximate to generate a lookup-table based replacement function for 'mlhdlc_approximate_sigmoid'

## Create a New Folder and Copy Relevant Files

Execute the following lines of code to copy the necessary example files into a temporary folder.

```mlhdlc_demo_dir = fullfile(matlabroot, 'toolbox', 'hdlcoder', 'hdlcoderdemos', 'matlabhdlcoderdemos'); mlhdlc_temp_dir = [tempdir 'mlhdlc_fixpt_approximate']; % create a temporary folder and copy the MATLAB files cd(tempdir); [~, ~, ~] = rmdir(mlhdlc_temp_dir, 's'); mkdir(mlhdlc_temp_dir); cd(mlhdlc_temp_dir); copyfile(fullfile(mlhdlc_demo_dir, [design_name,'.m*']), mlhdlc_temp_dir); copyfile(fullfile(mlhdlc_demo_dir, [design_name,'_design.m*']), mlhdlc_temp_dir); copyfile(fullfile(mlhdlc_demo_dir, [testbench_name,'.m*']), mlhdlc_temp_dir); ```

## Generate fixed-point lookup-table replacements

```repCfg = coder.approximation('Function','mlhdlc_approximate_sigmoid','CandidateFunction',@mlhdlc_approximate_sigmoid,...
'NumberOfPoints',50,'InputRange',[-10,10],'FunctionNamePrefix','repsig_');
coder.approximate(repCfg);
```

First the fixed-point conversion completes with appropriate function replacements, and following console message,

```### Generating approximation for 'sigmoid' : repsiglookuptable.m
### Generating testbench for 'sigmoid' : repsiglookuptable_tb.m
### LookupTable replacement for function 'sigmoid' used 50 data points```

This should generate the MATLAB files 'repsig_lookuptable_tb', and 'repsig_lookuptable' containing the testbench and design respectively.

## Test the replacement functions

To visually see the degree of match between lookup-table based replacement function and the original function use the testbench,

```repsig_lookuptable_tb();
```

## Clean up the Generated Files

You can run the following commands to clean up the temporary project folder.

```mlhdlc_demo_dir = fullfile(matlabroot, 'toolbox', 'hdlcoder', 'hdlcoderdemos', 'matlabhdlcoderdemos');
mlhdlc_temp_dir = [tempdir 'mlhdlc_fixpt_approximate'];
clear mex;
cd (mlhdlc_demo_dir);
rmdir(mlhdlc_temp_dir, 's');
```