Project

General

Profile

Bug #620

Total dipole in energy files is incorrect

Added by David van der Spoel almost 9 years ago. Updated over 6 years ago.

Status:
Closed
Priority:
Normal
Assignee:
-
Category:
mdrun
Target version:
Affected version - extra info:
Affected version:
Difficulty:
uncategorized
Close

Description

Due to the fact that gromacs internally breaks molecules during a simulation the total dipole is incorrect. If this total dipole is stored in the energy file it can be used for computing the dielectric constant in g_dipoles, resulting in values that are way too large. I have put a fatal error in g_dipoles whenever someone tries to use the energy file for computing the dielectric constant. This is unfortunate since that would be much quicker.

We can choose to either
- fix the computation of the dipole in mdrun
- not store the total dipole in the energy file anymore

History

#1 Updated by Rossen Apostolov about 8 years ago

  • Description updated (diff)
  • Assignee deleted (Erik Lindahl)
  • Target version deleted (4.5.1)

#2 Updated by David van der Spoel about 8 years ago

Let's go for the option not to store the dipole anymore.

#3 Updated by David van der Spoel about 8 years ago

  • Status changed from New to 3

git commit -m "Augmented redmin issue 620, by not writing the total dipole anymore to the energy file, since it is incorrect. The total dipole is still computed though, and some overhead may be reduced if that computation is removed from the code as well." -a
[release-4-5-patches 6a25aa5] Augmented redmin issue 620, by not writing the total dipole anymore to the energy file, since it is incorrect. The total dipole is still computed though, and some overhead may be reduced if that computation is removed from the code as well.

#4 Updated by Rossen Apostolov about 8 years ago

  • Status changed from 3 to Closed

#5 Updated by David van der Spoel about 8 years ago

  • Status changed from Closed to New
  • Target version set to 4.6

The change was abandoned in gerrit and postponed to 4.6 because it would create incompatible energy files. Therefore the issue needs to remain open until later.

#6 Updated by Berk Hess over 7 years ago

  • Status changed from New to In Progress

The dipole can not (easily) be calculated correctly with broken molecules.
Thus I have removed it from the energy file, unless it is used for Ewald
surface correction or dipole correction for 2D geometry (which you should
anyhow only use for system where molecules do not move across periodic
boundaries).
This fix is currently in the nbnxn_hybrid_acc branch and will be merged into 4.6.

#7 Updated by Erik Lindahl over 6 years ago

  • Status changed from In Progress to Closed

Now merged in release-4.6.

Also available in: Atom PDF