[gmx-users] tpbconv for following runs

[Anton Feenstra]

Ester Chiessi wrote:
> Hi all,
> 
> I'm using tpbconv to produce an MD calculation by following runs.
[...]

> I expected that all terms were identical, not only the potential energy
> contributions.
> 
> May I consider negligible these discrepances?

Yes. Although this has been discussed at some length already a few times,
it is not a bad thing to explain it once again.

In the verlet/leapfrog integration scheme (for MD), coordinates are
available at the whole timestep (t, t+dt, and so on), and velocities
at the half (t-dt/2, t+dt/2, t+(3/2)dt, and so on). In one 'x+v' frame
in the .trr file you will find coordinates at 't' and velocities only
at 't+dt'. However, the correct velocities for calculating physical
properties (kinetic energy, temperature, pressure) should be taken
at the whole timestep. Normally during your simulation this is achieved
by taking the average of the previous (t-dt/2) and current (t+dt/2)
velocities. In the .trr file, however, the previous velocities are not
present. This causes the observed discrepancies, and will cause the
continuation simulation to be *not* the exact same as an un-interrupted
simulation would have produced. The actuall error made, however, is
very small and both simulations (interrupt+continued and uninterrupted)
can be considered 'correct'.

-- 
Groetjes,

Anton
  _____________ _______________________________________________________
|             |                                                       |
|  _   _  ___,| K. Anton Feenstra                                     |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
| \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610           |
|             | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/       |
|             | "If You See Me Getting High, Knock Me Down"           |
|             | (Red Hot Chili Peppers)                               |
|_____________|_______________________________________________________|