[gmx-users] free energies with tabulated potentials or PME

[Mark Abraham]

Michael Brunsteiner wrote:
> hi everybody,
> 
> Can somebody tell me whether free energy calcs (TI)
> can be done in combination with user defined tabulated
> potentials ? ... or in other words: is dU/dlamda 
> calculated numerically or analytically ?

Don't know, I expect it does work. You could try it on a simple case and 
see...

> I understand that in gmx-3.3.1 free energy + PME is
> supported, has anybody tested this implementation in
> terms of accuracy/speed ?

Reading section 3.12 there's no reason to suppose that the free energy 
calculations have any dependence on electrostatic treatment, and thus 
relative cost and accuracy will be comparable to the normal costs of 
those algorithms. Why do you think it might?

> I did ask this question before but never got an answer,
> one keeps trying ...

Perhaps people thought they were questions you could have solved for 
yourself :-)

Mark