How can I be sure 'parfor' loop iterations are completely independent of each other

I am a newbie to parallel computing, however I want to speed up my code using 'parfor' in Matlab's parallel computing. I have nested 'for' loops and want to be sure of a correct result. How can I be sure the inner 'parfor' loop iterations is completely independent of each other loops.

for j = 1:length(w)
    j
    Kp2 = zeros(length(Kp1),1);
    for n = 1:length(Kp1)

          Kf2 = zeros(length(Kf1),1);
          eta = zeros(length(Kf1),1);
          parfor m = 1:length(Kf1)
              Kp2(m) = Kp1(n);
              Kf2(m) = Kf1(m);

              Z_Imp = double(subs(Z,[s Kp Kf Td], [s1(j) Kp2(m) Kf2(m) 0.2]));
              eta(m) = -real(Z_Imp(1,2)*Z_Imp(2,1))/abs(Z_Imp(1,2)*Z_Imp(2,1)) + 2*(real(Z_Imp(1,1))*real(Z_Imp(2,2)))/abs(Z_Imp(1,2)*Z_Imp(2,1));
          end

          Kp_Kf_eta(:,:,n) = [Kp2 Kf2 eta];

      end
      %%reshape data from [40,3,40] to [1600,3] by stacking [40,3] pages upon each other
      temp1 = permute(Kp_Kf_eta, [1,3,2]);
      Kp_Kf_eta_w(:,:,j) = reshape(temp1,1600,3);

end