Use MATLAB Coder to convert matlab code to mex without using multi-threading

My code is as follows

function all_Spins = simulate_particles(num_particles, num_steps, all_Spins, SpinLocOri, Gamma, magnetic_field, GdAmp, StepT, spherePositions, sphereRadii, Lx_intra, Ly_intra, Lz_intra, penetration, Lx_extra, Ly_extra, Lz_extra)
parfor mun = 1:num_particles
    particle = all_Spins(mun, :);
    particle = simulate_particle_motion(num_steps, particle, SpinLocOri, Gamma, magnetic_field, GdAmp, StepT, spherePositions, sphereRadii, mun, Lx_intra, Ly_intra, Lz_intra, penetration, Lx_extra, Ly_extra, Lz_extra);
    all_Spins(mun, :) = particle;
end
end

After using MATLAB Coder to convert to mex, I found that multi-threading was not used and my cpu was idle.

Answer 1

[Infinite_king]

Hi 梅花,

By default, MATLAB Coder generates code in which the parfor loop body runs in parallel using OpenMP.

In the code generation configuration object, you can enable 'OpenMP' as follows,

cfg = coder.config('mex'); 
cfg.EnableOpenMP = true; 
% generate code
codegen myFunction -config cfg

For more information, refer the following resources,

Enabling/Disabling OpenMP - https://www.mathworks.com/help/coder/ug/control-compilation-of-parfor-loops.html

MATLAB Answer related to OpenMP support - https://www.mathworks.com/matlabcentral/answers/237411-can-i-make-use-of-openmp-in-my-matlab-mex-files?#answer_190619

Parfor - https://www.mathworks.com/help/coder/ref/parfor.html

Hope this is helpful.

Answer 2

[Bruno Luong]

https://www.mathworks.com/help/coder/ug/use-parallel-for-loops.html