MATLAB Coder

Generate C and C++ code from MATLAB code

MATLAB Coder™ generates C and C++ code from MATLAB^® code for a variety of hardware platforms, from desktop systems to embedded hardware. It supports most of the MATLAB language and a wide range of toolboxes. You can integrate the generated code into your projects as source code, static libraries, or dynamic libraries. The generated code is readable and portable. You can combine it with key parts of your existing C and C++ code and libraries. You can also package the generated code as a MEX-function for use in MATLAB.

When used with Embedded Coder^®, MATLAB Coder provides code customizations, target-specific optimizations, code traceability, and software-in-the-loop (SIL) and processor-in-the-loop (PIL) verification.

To deploy MATLAB programs as standalone applications, use MATLAB Compiler™. To generate software components for integration with other programming languages, use MATLAB Compiler SDK™.

Get Started:

Run Anywhere
Deploy C++ Code from MATLAB
Deploy Deep Learning Networks and Machine Learning Models
Prototype on Hardware
Integrate with Software
Accelerate Algorithms

What Is MATLAB Coder?

Run Anywhere

Generate readable and portable ANSI C/C++ source code. Deploy code royalty-free.

Deploy Algorithms Royalty-Free

Use any C/C++ compiler to compile and run your generated code on any hardware, from desktop systems to mobile devices to embedded hardware. The generated code is royalty-free—deploy it in commercial applications to your customers at no charge.

Generating C/C++ Code from MATLAB Code (4:34)

Generated code for matrix multiply.

MATLAB Coder Success Stories

Learn how engineers and scientists in a variety of industries use MATLAB Coder to generate C/C++ code for their applications.

BMW Deploys Machine Learning Algorithms to Detect Oversteering onto ECU

Qualcomm Develops 5G RF Front-End Components and Associated Control Algorithms

Delphi Develops Radar Sensor Alignment Algorithm for Automotive Active Safety System

IDNEO Develops Embedded Computer Vision and Machine Learning Algorithms for Interpreting Blood Type Results

VivaQuant Accelerates Development and Validation of Embedded Device for Ambulatory ECG Sensing

dorsaVi developed motion analysis algorithms for medical, sports, and occupational safety applications using MATLAB Coder.

Delphi develops radar sensor alignment algorithm for automotive active safety system.

Supported Toolboxes and Functions

MATLAB Coder generates code from a broad range of MATLAB language features that design engineers use to develop algorithms as components of larger systems. This includes over 2500 operators and functions from MATLAB and companion toolboxes.

Supported Functions and Toolboxes

Deep Learning with MATLAB Coder

For Algorithms Using Additional Functions and Features, Consider Also Using MATLAB Compiler SDK

Generate code from computer vision algorithms for applications such as detecting and tracking faces using the KLT algorithm

Deploy C++ Code from MATLAB

Integrate generated code with object-oriented C++ source code.

Generate C++ Code with Namespaces

MATLAB Coder can generate C++ code in a namespace, making it easy to integrate with other source code that might have identical function or data type names. The code generator packages all generated functions and type definitions into the namespace.

C++ Code Generation

Integrate Multiple Generated C++ Code Projects

Generated code that integrates variables with identical data type names using namespaces.

Generate C++ Classes from MATLAB Classes

MATLAB Coder produces C++ classes from classes in your MATLAB code, including value classes, handle classes, and system objects. The generated code can be compiled into C++ libraries or executables and can be integrated into your existing C++ source code.

Generate C++ Classes for MATLAB Classes

Generate C++ Code with Class Interface

Generate C++ Classes from MATLAB Classes Using MATLAB Coder (4:13)

Generating Object-Oriented C++ Code from MATLAB Algorithms

Use Dynamically Allocated C++ Arrays in Generated Function Interfaces

Generate C++ code for MATLAB functions that accept or return an array with an array size unknown at compile time, or whose bound exceeds a predefined threshold. In the generated code, memory for the array is dynamically allocated and implemented as a class template named coder::array. In addition to exception-safe memory deallocation, coder::array provides APIs to access and manage the dynamic array.

Use Dynamically Allocated C++ Arrays in Generated Function Interfaces

Pass dynamically allocated arrays to generated functions.

Deploy Deep Learning Networks and Machine Learning Models

Generate code from trained deep learning networks and machine learning models.

Deploy End-To-End Deep Learning Algorithms

Deploy a variety of trained deep learning networks such as ResNet-50 and MobileNet-v2, as well as LSTM and other layers from Deep Learning Toolbox™ to Intel^® and ARM^® Cortex^® CPUs. Generate code for preprocessing and postprocessing along with your trained deep learning networks to deploy complete algorithms.

Supported Networks and Layers

Deep Learning on an Intel Processor with MKL-DNN (1:36)

Image Classification on ARM CPU: SqueezeNet on Raspberry Pi (4:22)

Code Generation and Deployment of MobileNet-v2 Network to Raspberry Pi

Classify images using SqueezeNet on Raspberry Pi

Generate Optimized Code for Deep Learning Inference

Because MATLAB Coder generates only the code needed to run inference with your specific algorithm, the code is faster and uses less memory than other deep learning solutions. The generated code calls optimized libraries, including Intel MKL-DNN for Intel processors and ARM Compute Library for ARM Cortex processors. Use GPU Coder™ to accelerate or deploy algorithms by generating CUDA^® code that runs on any modern NVIDIA^® GPU.

Generate C++ Code for Object Detection Using YOLO v2 and Intel MKL-DNN

Deep Learning Inference for Object Detection on Raspberry Pi (2:04)

Defect Detection Inference on Arm Cortex A from MATLAB (4:37)

GPU Coder

Pedestrian detection running on Intel Xeon processors

Deploy End-to-End Machine Learning Models

Deploy statistics and machine learning models by generating C/C++ code for your entire machine learning algorithm, including preprocessing and postprocessing. Update parameters of deployed models without regenerating the C/C++ prediction code.

Key Techniques for Running Your Machine Learning Model on Embedded Devices (2:30)

Code Generation for Machine Learning Models

Code Generation for Image Classification

Update Model Parameters for Code Generation

Generate Fixed-Point C/C++ for SVM Prediction

Code generation workflow for machine learning models.

Prototype on Hardware

Get to hardware fast with automatic conversion of your algorithm to C/C++.

Prototype on Desktop and Cloud Platforms

Use the MATLAB Coder app or equivalent command-line functions to quickly generate code for your signal processing, computer vision, deep learning, control systems, or other application and then compile the code for your hardware.

Generate Executable for Prototyping Using MATLAB Coder (2:57)

Deep Learning on an Intel Processor with MKL-DNN (1:36)

Prototyping algorithms quickly on desktops and cloud platforms for image processing algorithms such as Sobel edge detection.

Prototype on Embedded and Mobile Platforms

Target any device by manually integrating the generated code with your application. Automate the process for Raspberry Pi using MATLAB Support Package for Raspberry Pi.

Deploying to iPhone and iPad Apps Using MATLAB Coder (0:26)

Raspberry Pi Support from MATLAB, Part 2: Deploy MATLAB Algorithms on Raspberry Pi (11:11)

Defect Detection Inference on Arm Cortex A from MATLAB (4:37)

Image Classification on ARM CPU: SqueezeNet on Raspberry Pi (4:22)

Prototyping algorithms quickly on embedded and mobile platforms.

Move from Prototyping to Production

Use MATLAB Coder with Embedded Coder to generate code that takes advantage of processor-specific intrinsics that can execute faster than standard ANSI/ISO C/C++ code.

Accelerate Using Processor-Specific Intrinsics for ARM Cortex Processors

Profile Execution Time of the Generated Standalone Code

Interactively Trace from MATLAB Code to Generated Code and Vice Versa

Profile execution time of the generated standalone code.

Integrate with Software

Reuse MATLAB algorithms as C/C++ code within your software environment.

Generate Code with Simple Interfaces That Are Easy to Integrate

Generated code uses C/C++ types in a natural way, simplifying integration with external code. You can integrate generated code as source code or libraries. Trusted C/C++ libraries or components can be brought into MATLAB for higher-fidelity testing and are automatically called from generated code as well.

Example Main Function Shows How To Invoke Generated Code

Integrate Code into Visual Studio Using MATLAB Coder (4:17)

Unit Testing C Code Using MATLAB and MATLAB Coder (4:19)

Interactive traceability report using MATLAB Coder with Embedded Coder.

Optimize the Performance of Generated Code

Apply optimizations to adjust tradeoffs between execution speed, memory usage, readability, and portability. Use profiling tools to identify bottlenecks. To further boost performance, generate multicore OpenMP code and call optimized libraries such as LAPACK, BLAS, and FFTW when available.

MATLAB Coder Optimizations in Generated Code

Multicore Code Generation

Profile MEX Functions by Using MATLAB Profiler

Example of generated code with calls to OpenMP.

Reuse MATLAB Tests on Generated Code Prior to Integration

Reuse existing MATLAB tests to verify the behavior of generated code in the interactive MATLAB environment. Use the MATLAB unit test framework to quickly develop a rich set of regression tests that can be used to verify the generated C/C++ code.

MATLAB Unit Test Framework

Unit Testing Your Generated Code Using MATLAB Coder (2:58)

Verify Numerical Behavior of the Final Generated Code Using Embedded Coder

Verifying behavior of generated code before integrating with your application.

Accelerate Algorithms

Generate C/C++ code and compile it for use inside MATLAB.

Accelerate Algorithms on CPUs

You can call generated code as MEX functions from your MATLAB code to speed execution, though performance will vary depending on the nature of your MATLAB code. You can profile generated MEX functions to identify bottlenecks and focus your optimization efforts.

Accelerating MATLAB Algorithms and Applications

Accelerate Image Compression Algorithm Using MATLAB Coder (4:21)

Profile MEX Functions by Using MATLAB Profiler

Profile MEX functions to identify bottlenecks in performance.

Accelerate Algorithms Using GPUs

Use Parallel Computing Toolbox™ to accelerate algorithms running in MATLAB. Use GPU Coder to generate CUDA code for acceleration or deployment that runs on any modern NVIDIA GPU.

Parallel Computing Toolbox

GPU Coder