[gmx-users] dipole corrections to PME

[David van der Spoel]

Hi,

For your information, I have tested the dipole corrections to PME in gmx 
3.1 (I admit I should have done this before release but anyway). The 
absolute fluctations in the total system dipole go down considerably, in 
particular when using small epsilon_surface. However you can compensate 
for that when computing the dielectric constant epsilon_0 (by an option to 
g_dipoles)

System: 820 SPC/E, cubic box, 2 ns, nstlist = 5, average and stddev over 
last 500 ps as measured with xmgrace
epsilon_surface	epsilon_0 (std.dev.)
1		84 (15)		==> vacuum boundary
30		74 (1.4)
78.5		65 (0.6)
200		71 (0.8)
inf		73 (0.9)	==> conducting boundary

Apart from eps_surf = 1, which is very noisy, the systems seem to behave 
well. It is a bit strange that at 78.5, the real dielectric constant, 
there is a minimum in the measured one, although the errors are larger 
than what's given here (the stddev of the running average, which 
basically says how well it converges).

-- 
Groeten, David.
________________________________________________________________________
Dr. David van der Spoel, 	Biomedical center, Dept. of Biochemistry
Husargatan 3, Box 576,  	75123 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://zorn.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++