Clear Filters
Clear Filters

How to use OpenMP properly from C++ static library?

6 views (last 30 days)
I am trying to incooperate a static library written in C++ in Matlab. The C++ static library uses OpenMP for parallel processing. ;
I have built the library interface following I am using Windows.
When I call the C++ library from Matlab, it works, but parallel processing is not working and it only runs in serial. In Matlab, the compiler has not been changed. MEX configured to use 'Microsoft Visual C++ 2019 (C)' for C language compilation
How do I setup Matlab or generate library interface properly to use OpenMP of the C++ library?
hH on 11 Mar 2022
Thanks for your reply.
Yes, I check the number of threads avaiable from the system and try to run with that many thread. I use the following set up.
The nThreads value printed out by fprintf is multiple (e.g. 4).
The library also seems to work with OpenMP properly if I am calling the library from another C++ application.
int nThreads = omp_get_max_threads() - 1;
fprintf(stderr, "Num. threads = %d\n", nThreads);
#pragma omp parallel for num_threads(nThreads)
James Tursa
James Tursa on 11 Mar 2022
Can you show the code for how you are calling the library functions?

Sign in to comment.

Answers (1)

Infinite_king on 10 Jan 2024
As per my understanding you are able to use the library from MATLAB but the operations are not being performed in parallel.
To enable OpenMP support for your C++ library when using it in MATLAB through MEX, you need to ensure that the OpenMP flags are correctly set during compilation. Since you're using Microsoft Visual C++ 2019, you need to include the OpenMP flag (/openmp) while compiling.
Follow the below troubleshooting steps to possibly resolve the issue.
  1. Make sure to compile the library with the same compiler that you are using with MATLAB.
  2. Pass the OpenMP flag (/openmp) while compiling Interface.
  3. Change the Mex configuration file or pass the OpenMP flag while compiling mex. Refer the following link for more information,
  4. Checkout the following answer, you may find it useful,
Hope this is helpful.


Find more on Execution Speed in Help Center and File Exchange




Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!