Parallel coding in matlab

Hello everyone,

I am writting a code that In this code I have 8 independant equations which each equation is a second order equation that should be solved using Runge-Kutta 4th order method. But I need to solve them paralerly.

Do you know how can I do this work?

[Bjorn Gustavsson]

Why do you need to solve them parallely if they are independent? If they are independent shouldn't their solutions be possible to calculate one after the other, and then you can put them together without any-one knowing the difference?

[Mitra Taghizadeh]

Becuse the computation time ia too much if I solve each equation after the athor. In fact this solution is in a for-loop, and this 8 equation should be solved in each time.

In a paper, it was mentioned that if you use the parallel algorithem it will have a less computational time.

Answer 1

[Walter Roberson]

In your r2019b release you would need the Parallel Computing Toolbox and use parfor or parfeval or spmd

With sufficiently new MATLAB versions there is a new built-in background pool that can execute parfeval without needing any additional toolbox.

Note that depending on the details of what is being calculated and on the memory patterns, there are cases where calculations in parallel will be slower than calculating sequentially.