Code Generation

Home > MPLAB Device Blocks for Simulink > User Guide > Code Generation

🏠 Home

User Guide Topics

The MCHP Blockset automatically generates optimized C code from Simulink models for deployment on Microchip microcontrollers.

Code Generation Workflow

Model Design - Create algorithm in Simulink
Configuration - Set up target hardware and compiler
Build - Generate and compile code
Deploy - Program device with generated firmware

Model Configuration Parameters

System Target File

Select the appropriate MCHP system target file for your device family:

MCHP_dsPIC_stf.tlc / MCHP_dsPIC_PIC24.tlc / MCHP_dsPIC_TC_stf.tlc - classic dsPIC (30F / 33F / 33E / 33C)
MCHP_dsPIC33A_stf.tlc / MCHP_dsPIC33A_noTMF_stf.tlc - dsPIC33A / PIC32A
MCHP_pic32_stf.tlc - PIC32 (MIPS)
MCHP_ARM0p_stf.tlc / MCHP_ARM4F_stf.tlc / MCHP_ARM7_stf.tlc - SAM (ARM Cortex-M)

The correct target is selected automatically for the device chosen in the Master block.

Code Generation Settings

Parameter	Recommended Value	Description
System target file	one of the MCHP `*.tlc` above	Selected automatically from the Master block device
Language	C	Generate C code (default)
Code interface packaging	Reusable function	For standalone deployment
Generate code only	Unchecked	Enable compilation

MCHP Target Settings

Configure in Code Generation → MCHP Target section:

Device - Select target microcontroller
Compiler Version - Select installed compiler
Optimization Level - -O0 (debug) to -O2 (free license); -O3 / -Os require a paid PRO license
Build Action - Build, Build and Download, etc.

Optimization Settings

Compiler Optimization Levels

Level	Flag	Description	Use Case
0	-O0	No optimization	Debug, code analysis
1	-O1	Basic optimization	Development testing
2	-O2	Moderate optimization	Balanced speed/size
3	-O3	Aggressive optimization	Maximum performance
s	-Os	Size optimization	Memory-constrained devices

Model Optimization

Solver - Fixed-step, Discrete (no continuous time)
Eliminate dead branches - Enable for code size reduction
Inline parameters - Enable for performance
Reusable functions - Enable for modular code

Custom Code Integration

Adding Custom C Code

Use Code Generation → Custom Code section:

Include directories - Add header file paths
Source files - Add .c files to compilation
Additional include files - Specify headers to include
Custom code - Add code snippets (initialization, etc.)

Legacy Code Integration

Use Legacy Code Tool to wrap existing C functions:

Create legacy code specification structure
Register function interface with Simulink
Generate S-Function wrapper
Use S-Function block in model

Code Generation Process

Build Steps

TLC Processing - Model converted to Target Language Compiler format
Code Generation - C code generated from TLC
Compilation - XC compiler builds object files
Linking - Object files linked to create executable
Post-processing - Generate .hex, .elf, .lst files

Generated Files

File Type	Extension	Description
Source Code	.c, .h	Generated C source and headers
Executable	.elf	ELF format with debug symbols
Programming File	.hex	Intel HEX for device programming
Listing	.lst	Assembly listing with addresses
Map File	.map	Memory map and symbol table

Fixed-Point Code Generation

Fixed-Point Toolbox Integration

For DSP and motor control applications:

Use fi() objects for fixed-point signals
Configure word lengths: 8, 16, 32 bits
Set fraction lengths for precision
Enable fixed-point code generation in model settings

dsPIC Assembly Optimization

The blockset includes hardware-optimized assembly code replacements (classic dsPIC 30F/33F/33E/33C):

Trigonometric - sin, cos, sincos, atan2
Square root - 16-bit and 32-bit fixed-point
Saturation - optimized type conversions with overflow handling
Arithmetic - absolute value, mixed signed/unsigned multiply

dsPIC33A / PIC32A additionally use the hardware FPU and single-precision math replacements (see release notes).

Note: Assembly replacements are automatically used when available for the target device family.

Memory Management

RAM Allocation

Stack - Configured in linker script
Heap - Avoid dynamic allocation if possible
Global Variables - Minimize for better performance
DMA Buffers - Use aligned memory sections

Flash Optimization

Use const for constant data
Enable compiler size optimization (-Os)
Minimize duplicate code with reusable functions
Use lookup tables for complex math

Build Diagnostics

Common Build Errors

Error	Cause	Solution
Compiler not found	Path not configured	Set compiler path in model settings
Symbol not defined	Missing library/header	Add to custom code includes
Memory overflow	Code/data too large	Enable optimization, reduce model size
Unsupported block	Block not compatible	Replace with supported alternative

Build Warnings

Address these warnings for robust code:

Unused variables - Remove or comment out
Type mismatches - Use explicit casts
Uninitialized variables - Add initialization
Deprecated functions - Update to current API

Code Review and Analysis

Generated Code Inspection

Review generated code for:

Efficiency - Check loop structure, function calls
Safety - Verify array bounds, pointer usage
Readability - Enable code generation comments
Compliance - MISRA-C, coding standards

Code Metrics

Use Code Generation Report to analyze:

Lines of code generated
ROM/RAM usage
Function complexity
Stack depth estimates

Best Practices

Use discrete solver with fixed time step
Avoid dynamic memory allocation
Minimize global variable usage
Use appropriate data types (int16, uint32, etc.)
Enable code generation report for documentation
Test with PIL before final deployment
Use version control for model files

Navigation