[gmx-users] settle issue

[Mu Yuguang (Dr)]

Dear All,
When I try to simulate TIP4Pew water box at 298K, and found the density
got from gromacs is relative low (983 g/cm3) than the amber (sander)
version 994 g/cm3. I cannot find the reason. The parameters for the
tip4pew model are identical. When I check some intra-molecular
distances, and seem to find a hint: in gromacs simulation the
intra-molecular atom-atom distance (should be fixed by settle)
fluctuation is quite large!
How can one increase the accuracy of settles in gromacs? 

Regards
Yuguang