[gmx-users] shake for water

[David Mobley]

All,

A quick question on constraints... I'm using TIP4P-Ew in gromacs 3.3.2
and am concerned with reproducing energies from another code very
precisely for several specific snapshots. I am doing a zero-step mdrun
of a setup with one small molecule and two tip4p-ew water molecules.

Anyway, I have set the shake tolerance to 1e-12 in the mdp file, but
to my surprise the internal water distances are good only to 1e-06 and
1e-07. Is this expected behavior? Note that I am running in double
precision. I assumed that, er, the distances should converge to the
shake tolerance.

Thanks,
David