Defining an efficient distance function

I'm using kNN search function in matlab, but I'm calculating the distance between two objects of my own defined class, so I've written a new distance function. This is it:

           function d = allRepDistance(obj1, obj2)
           %calculates the min dist. between repr.
           % obj2 is a vector, to fit kNN function requirements

              n = size(obj2,1);
              d = zeros(n,1);
              for i=1:n
                      M =  dist(obj1.Repr, [obj2(i,:).Repr]');
                      d(i) = min(min(M));
              end

       end

The difference is that obj.Repr may be a matrix, and I want to calculate the minimal distance between all the rows of each argument. But even if obj1.Repr is just a vector, which gives essentially the normal euclidian distance between two vectors, the kNN function is slower by a factor of 200!

I've checked the performance of just the distance function (no kNN). I measured the time it takes to calculate the distance between a vector and the rows of a matrix (when they are in the object), and it work slower by a factor of 3 then the normal distance function.

Does that make any sense? Is there a solution?

[Adam]

Have you profiled it?

profile on
your code
profile off
profile viewer

I have found in previous versions (haven't really done any analysis in R2014a or b) that some class-based functionality can be very slow. In particular the overhead of accessing a class property. In your case though I don't really see any obvious excessive access of class properties.

I did spend quite a while removing quite a bit of the class-based aspect from a speed-critical algorithm last year though because of the significant increase in overhead it caused. I ended up just breaking all ideas of encapsulation and pulling properties out of a class to run algorithms on the raw data rather than in class methods.

[Roy]

Thanks. I've made some research and it seems that MATLAB oop has a major overhead issue...