I am simulating a glycerol-water mixture system in a square box of length 3.5 nm. After the NPT step my box size decreased to 2.62984 nm. However, in the beginning of the production run step I get the following error:
ERROR: The initial cell size (0.657461) is smaller than the cell size limit (0.683200)
Also, the log file says:
Initializing Domain Decomposition on 16 ranks
Dynamic load balancing: auto
Will sort the charge groups at every domain (re)decomposition
Initial maximum inter charge-group distances:
two-body bonded interactions: 0.390 nm, LJ-14, atoms 479 480
multi-body bonded interactions: 0.390 nm, Proper Dih., atoms 479 480
Minimum cell size due to bonded interactions: 0.429 nm
Maximum distance for 5 constraints, at 120 deg. angles, all-trans: 0.683 nm
Estimated maximum distance required for P-LINCS: 0.683 nm
This distance will limit the DD cell size, you can override this with -rcon
Should decrease the number of processors or just add -rcon in the mdrun command or perhaps do something else?
Thanks,
Sourav
You can even increase the cell-size by using editconf function and centering your system in a different box. Nevertheless, it might be more convenient to perform some more energy minimization and NVT-run in order to "shake" the system a bit. You also may run the whole thing on GPU, if you have access, but it still can happen that you might get an error...
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics
- Atoms