Limitations for MATLAB Compiler
This topic presents limitations to consider when working with MATLAB® Compiler™ and MATLAB Runtime, including packaging limitations and other restrictions.
To see the full list of
MATLAB
Compiler
limitations, visit: https://www.mathworks.com/products/compiler/compiler_support.html.
Packaging MATLAB and Toolboxes
MATLAB Compiler supports nearly the full MATLAB language and toolboxes based on MATLAB except:
Most prebuilt graphical user interfaces included in MATLAB and its companion toolboxes
Functionality that cannot be called directly from the command line
For a list of functions not supported by MATLAB Compiler, see Functions Not Supported for Compilation by MATLAB Compiler and MATLAB Compiler SDK.
Compiled applications can run only on operating systems that run MATLAB. Also, because MATLAB Runtime is approximately the same size as MATLAB, applications built with MATLAB Compiler need specific storage memory and RAM to operate. For the most up-to-date information about system requirements, see MATLAB System Requirements.
For guidelines on multiplatform deployment, see Ensure Multiplatform Portability for Compiled Applications (MATLAB Compiler SDK).
MATLAB Compiler Generated Components
Components generated by MATLAB
Compiler SDK™ or MATLAB
Compiler are intended for deployment with MATLAB Runtime and cannot be used in MATLAB directly. You can test generated standalone applications in MATLAB using the system command or bang operator. For more
information on testing applications in MATLAB, see Set MATLAB Library Paths for Testing Deployed Applications.
Older Neural Networks Not Deployable with MATLAB Compiler
Loading a neural network created using an older version of Deep Learning Toolbox™ versions requires some initialization routines that are not deployable with MATLAB Compiler. Therefore, you need to update the network first before deploying it.
For example, deploying a network created using Deep Learning Toolbox Version 5.0.1 (2006b) and MATLAB Compiler Version 4.5 (R2006b) yields the following errors at run time:
??? Error using ==> network.subsasgn
"layers{1}.initFcn" cannot be set to non-existing
function "initwb".
Error in ==> updatenet at 40
Error in ==> network.loadobj at 10
??? Undefined function or method 'sim' for input
arguments of type 'struct'.
Error in ==> mynetworkapp at 30Restrictions on Parallel Pools
Deployed applications that use a shared MATLAB Runtime instance or singleton runtime cannot use more than one parallel pool. To use multiple parallel pools, refactor your code to create multiple MATLAB Runtime processes.
Parallel Computing Toolbox™ functionality is not supported in MATLAB
Compiler SDK C++ shared libraries using the MATLAB Data API interface running in
MATLABApplicationMode::OUT_OF_PROCESS mode.
Restrictions on Calling PRINTDLG with Multiple Arguments in Packaged Mode
When code runs in deployed mode, you can include only one argument in a call to the
MATLAB
printdlg function (for example,
printdlg(gcf)).
If you make a call to printdlg with multiple arguments in
MATLAB, you do not receive an error. However, if you deploy a compiled application
containing the multiple-argument call, the action fails, and the software displays the
following error message:
Error using = => printdlg at 11 PRINTDLG requires exactly one argument
Opening File Using which and open Does Not Search Current Working Folder
Using the which function, as in the example below, does not cause the
current folder to be searched in deployed applications. In addition, using this function
might cause unpredictable behavior of the open
function.
function pathtest
which myFile.mat
open('myFile.mat') Use one of the following solutions as an alternative:
Use
loador a specialized function for your particular file type.loadexplicitly checks for the file in the current folder.load myFile.mat
Use the
ctfrootfunction to explicitly point to the file location relative to the deployable archive.load(fullfile(ctfroot,'..','datafiles','data1.mat'));
Include the file using one of the following options: in the Files required for your application to run area of the Compiler app, the
AdditionalFilesoption of acompiler.buildfunction, or the-aflag ofmcc. You can then locate the file using thewhichfunction.
For more information on including data files with your application, see Include and Access Files in Packaged Applications.
Suppress Warnings on UNIX System
Several warnings might appear when you run a standalone application on UNIX® systems.
To suppress the libjvm.so warning, set the dynamic library path
properly for your platform. For details, see Set MATLAB Runtime Library Paths for Deployment.
You can also use the compiler option -R -nojvm to set your
application's nojvm run-time option, if the application is capable of
running without Java®.
If your application includes graphics and you compile it using the
-nojvm option, you get a run-time error.
No MATLAB File Help for Packaged Functions
If you create a MATLAB function with self-documenting online help and then package it, the results of the following command are unintelligible:
help filename
Note
For performance reasons, MATLAB file comments are removed before MATLAB Runtime encryption.
Using C++ SetData to Dynamically Resize mwArray
You cannot use the C++ SetData method to dynamically resize an
mwArray.
For instance, if you are working with the following array, you cannot use
SetData to increase the size of the array to a length of five
elements.
[1 2 3 4]
Accepted File Types for Packaging
The table below lists the valid and invalid file types for packaging using deployment apps.
| Target Application | Valid File Types | Invalid File Types |
|---|---|---|
Standalone applications | MATLAB MEX files, MATLAB scripts, MATLAB functions, and MATLAB class files. These files must have a single entry point. | Protected function files ( |
C/C++ shared libraries, .NET assemblies, Java classes, Python® packages, and COM components | MATLAB MEX files, MATLAB functions, and MATLAB class files. These files must have a single entry point. | MATLAB scripts, protected function files ( |
MATLAB Production Server archives | MATLAB MEX files and MATLAB functions. These files must have a single entry point. | MATLAB scripts, MATLAB class files, protected function files ( |
You can add other types of files to the packaged code archive, such as data files. For more information, see Include and Access Files in Packaged Applications.
