Vectorization of 4 nested for loops

Hello everybody,

I'm trying to vectorize some code that is used in a large numeric engineering model. The original version of the code is:

 % non-vectorized code
 for i = 1:iMax
    for j = 1:round(jMax/2)
        for k = 1:iMax
            for p = 1:jMax
                wIncremental = firstMatrix(k,p)*secondMatrix(iMax-i+k,jMax-j+p);
                wIncrementSum(i,j) = wIncrementSum(i,j) + wIncremental;
            end
        end
    end
 end

Starting from that, I managed to vectorize the code as follows:

 % vectorized code
 for i = 1:iMax
    for j = 1:round(jMax/2)
        wIncrement = firstMatrix.*secondMatrix((iMax-i+1):(2*iMax-i),(jMax-j+1):(2*jMax-j));
        wIncrementSum(i,j) = sum(sum(wIncrement));
    end
 end

My question is: Is there a way of vectorizing the second version even more to achieve even shorter runtime? I'm dreaming of a code that runs without any for loops at all. So far I couldn't come up with any solution to this. If not, do you see any other room for improvement?

Thanks for your help and comments