Is it possible to accelerate using multiple threads directly over the algorithm interface provided by PCL

I’m new to multithreaded parallel processing. I was recently working on a 3d reconstruction of a box. I used FPCS algorithm for point cloud coarse registration, but the program needed a relatively long time to run. In the header file of the FPCS algorithm, a function is described as follows.
enter image description here

OpenMP is mentioned in the function description above. So, I try to use multithreading for program acceleration.
But how can I do that?
Do I have to compile PCL source code and then make changes to it?

The following is the original call to the FPCS algorithm. Can I directly modify this program to achieve the purpose of parallel acceleration?
enter image description here

Any suggestions will be apprecaited!

Source: Windows Questions