Hi,

I have a very strange behavior when performing replica exchange with the ACE implicit solvent model, and I think it's a bug.

Here is the input:

read rtf card name @toppar/toph19.inp
read param card name @toppar/param19.inp
read ACEParameters card name @toppar/ace/acepar19.inp

read psf card name conf.psf
read coor card name conf.cor

! Activate ace2
energy atom ace2 ieps 1.0 seps 80.0 alpha 1.3 sigma 2.5 switch -
vdis vswi cutnb 15.0 ctonnb 10.0 ctofnb 14.0

repd nrep 2 exlm unit 33 freq 1000

energy atom ace2 ieps 1.0 seps 80.0 alpha 1.3 sigma 2.5 switch -
vdis vswi cutnb 15.0 ctonnb 10.0 ctofnb 14.0

stop

The problem is: the second energy statement gives a different value than the first one. And all the difference comes from the terms calculated by ACE.

I did some tests without any implicit solvent, or with the FACTS model, and the energy was constant, as expected.
I also tested the BLOCK module with ACE, and it worked perfectly (as the doc says).

The complete output is attached.

Attached Images
output.txt (11.01 KB, 523 downloads)