Talk:Ewald summation

In the final section, it says that direct calculation has O(N^2) scaling. With proper choice of the cutoff-potential, I believe that direct calculation has O(N ln N) scaling. —Preceding unsigned comment added by 129.105.92.178 (talk) 22:47, 1 June 2010 (UTC)[reply]

The first paragraphs & formulas in the PME are useless and (thus) misleading, as they are similar to the classic ewald approach. Relevant info starts when FFT is mentionned. — Preceding unsigned comment added by 130.88.113.186 (talk) 10:47, 8 January 2013 (UTC)[reply]

I think that some sentences in the "Dipole term section" can be misleading, even if maybe they aren't strictly wrong. In fact, I calculated the electric field at the center of a spherical shell of identical (same strength, same direction) point-like electrical dipoles and I found it to be zero. Thus, if my result is correct, in the summation of spherical shells of dipoles the convergence issue does not even arise. I did not try to perform the analytical calculation on a cubic shell, but some numerical calculations make me suspect that it is zero too. I think that someone should challenge and eventually confirm my results. Data 1 (talk) 08:41, 9 February 2017 (UTC)[reply]