We are developing software in C++ that needs to interface with MATLAB, predominantly the reading and writing of MAT-files. We want to use the C++ Mx and MAT API without having to redistribute all of the MATLAB DLL files. Are there smaller libraries that we can leverage for this application?
MATLAB: How to redistribute C++ applications that read and write MAT files
bundlecdeploymatMATLABredistribute
Related Solutions
You can test your build environment first using the "mex" command. You can also get a sample file "main.cpp" file from the "Test Your Build Environment" documentation page in:
Below is a guideline on building C++ Engine program using Visual Studio 2017 which uses the "main.cpp" from the aforementioned documentation link.
Create a new Visual C++ project from existing code
Click on "File > New > Project From Existing Code..." in the menu of Visual Studio 2017.
When it asks for:
- What type of project would you like to create? Choose "Visual C++".
- Project file location? Choose the path to a directory which contains only "main.cpp" file and no other files, e.g. "E:\Project".
- Project name: Choose any name, e.g. "Project".
- Use Visual Studio: Choose "Console application project"
Then change the platform to x64, see the screenshot below:
Set the correct properties for the project
Inside Solution Explorer, right click on the project and choose "Properties".
Set the path to the MATLAB header files
Then click on "Configuration Properties > C / C++ > General", and set (see screenshot below):
- Additional Include Directories: the path to "<matlabroot>\extern\include", e.g. "E:\MATLAB64\R2018b\extern\include"
Set the path to the MATLAB static library files
Then click on "Configuration Properties > Linker > General" and set (see screenshot below):
- Additional Library Directories: the path to the "<matlabroot>\extern\lib\win64\microsoft", e.g. "E:\MATLAB64\R2018b\extern\lib\win64\microsoft"
Set the dependencies in the Visual Studio linker
Then click on "Configuration Properties > Linker > Input" and set (see screenshot below):
- Additional Dependencies: "libMatlabEngine.lib;libMatlabDataArray.lib;%(AdditionalDependencies)"
Then finally setup the environmental variable for the debugging
Click on "Configuration Properties > Debugging > Environment" and set (see screenshot below):
- Environment: "PATH=<matlabroot>\extern\bin\win64;%PATH%"
- e.g. "PATH=E:\MATLAB64\R2018b\extern\bin\win64;%PATH%".
Building and running the project
In "main.cpp" line 34 (it says "return 0;") put a break point.
Press F5 or the "Local Windows Debugger" button to start the debugger. It will build the project automatically and start the compiled console application, see the screenshot below.
If you have built the project in the past, you can rebuild it by going to "Build > Rebuild Solution".
We have a documentation page which mentions what is needed to be done in order to compile MATLAB Engine applications in an IDE. This is available in the documentation page in:
You can link in the MAT-file library by following the directions in the documentation for loading MAT-file. Executing the following command at the MATLAB prompt will open the documentation on how to go about using the built-in .DLL libraries to write .MAT files from third-party applications.
web([docroot '/techdoc/matlab_external/f19027.html#f20073'])Related Question
- Do I get an error when deploying a MEX file created using MATLAB 7.7 (R2008b) to another 64-bit Windows machine
- How to exclude a file from the automatic dependency analysis, which was introduced in R2016b
- Does MATLAB crash upon start with error code 0xC0150004 or give a System Error:14001 stating that the side-by-side configuration is incorrect when trying to start
- Where can I place a configuration file for a MATLAB-compiled Application
- Are there C++ libraries available to access MATLAB generated MAT-files from the C++ application
- Do I receive an “application has failed to start” error message when loading DLL into MATLAB with LOADLIBRARY
Best Answer