[gmx-users] Bug in g_energy for calculating heat capacity?

[Zhang Zhigang]

Hi, all,
     Actually I think others have also noticed this "bug": g_energy often
generate incorrect heat capacity.  When I checked the source code of
gmx_energy.c, I found the calculation of heat capacity is only related with
the fluctuation of temperature, which is actually the algorithm for NVE
ensemble only (refer to Allen & Tildesley, Computer simulation of liquids,
pp. 53, eqn. 2.82). For other ensembles, e.g. NVT ensemble, the heat
capacity should be related with the fluctuation of energies.
     Another problem: even with the energy fluctuation algorithm for the
heat capacity, I found it is not easy to get an convergent value. The most
important factor is that the fluctuations are sensitively related with the
temperature (pressure) coupling time constants for NVT (NPT) ensembles.
I wonder if anyone can help me resolving this problem? Thanks a lot!

Regards,

zgzhang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20081221/374225ee/attachment.html>