c++ - Which loops should I parallelize, the outer or the inner ones -

I am writing an image processing filter, and I want to speed up computation using openmp. My pseudo code structure is as follows:

 for  (in every pixel in the image) {// some stuff here (for any combination of parameters) {// do other stuff here and Filter}}  

The code is filtering every pixel using different parameters, and selecting optimal.

My question is, what is faster: In the middle, to parallel the loop first, the processor or sequentially to reach the pixels and parallel the selection of different parameters.

I think the question can be one more general: what is fast, giving a large amount of operation to every thread, or making multiple threads with some operations.

I do not care about the details of the implementation now, and I think I can handle them with my previous expertise using OpenMP Thank you!