PMASS: Difference between revisions
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
{{TAGDEF|PMASS|[Real]|1000}} | {{TAGDEF|PMASS|[Real]|1000}} | ||
Description: {{TAG|PMASS}} assigns a fictitious mass (in amu) to the lattice degrees-of-freedom in case of Parrinello-Rahman dynamics (in case VASP was compiled with [[ | Description: {{TAG|PMASS}} assigns a fictitious mass (in amu) to the lattice degrees-of-freedom in case of Parrinello-Rahman dynamics (in case VASP was compiled with [[Precompiler options#-Dtbdyn|-Dtbdyn]]). | ||
---- | ---- | ||
When running ''NpT'' simulations with a [[MDALGO#LangevinEOM|Langevin thermostat]]<ref name="Allen91"/> ({{TAG|MDALGO}}=3), using the method of [[MDALGO#ParrinelloRahman|Parrinello and Rahman]],<ref name="Parrinello80"/><ref name="Parrinello81"/> a fictitious mass (in amu) for the lattice degrees-of-freedom has to be assigned using the {{TAG|PMASS}}-tag. | When running ''NpT'' simulations with a [[MDALGO#LangevinEOM|Langevin thermostat]]<ref name="Allen91"/> ({{TAG|MDALGO}}=3), using the method of [[MDALGO#ParrinelloRahman|Parrinello and Rahman]],<ref name="Parrinello80"/><ref name="Parrinello81"/> a fictitious mass (in amu) for the lattice degrees-of-freedom has to be assigned using the {{TAG|PMASS}}-tag. | ||
| Line 10: | Line 10: | ||
The optimal setting for {{TAG|PMASS}} depends very much on the particular system at hand and can be considered as a compromise between two opposing factors: too large values lead to very slow variation of lattice degrees of freedom (and hence the sampling becomes inefficient) while too small value can lead to too large geometric changes in a MD step and hence may cause numerical problems. We strongly recommend to make careful tests with various settings before performing the production run. | The optimal setting for {{TAG|PMASS}} depends very much on the particular system at hand and can be considered as a compromise between two opposing factors: too large values lead to very slow variation of lattice degrees of freedom (and hence the sampling becomes inefficient) while too small value can lead to too large geometric changes in a MD step and hence may cause numerical problems. We strongly recommend to make careful tests with various settings before performing the production run. | ||
== Related | == Related tags and articles == | ||
{{TAG|LANGEVIN_GAMMA_L}}, | {{TAG|LANGEVIN_GAMMA_L}}, | ||
{{TAG|LANGEVIN_GAMMA}}, | {{TAG|LANGEVIN_GAMMA}}, | ||
| Line 25: | Line 25: | ||
---- | ---- | ||
[[Category:INCAR]][[Category:Molecular | [[Category:INCAR tag]][[Category:Molecular dynamics]] | ||
Revision as of 14:21, 7 April 2022
PMASS = [Real]
Default: PMASS = 1000
Description: PMASS assigns a fictitious mass (in amu) to the lattice degrees-of-freedom in case of Parrinello-Rahman dynamics (in case VASP was compiled with -Dtbdyn).
When running NpT simulations with a Langevin thermostat[1] (MDALGO=3), using the method of Parrinello and Rahman,[2][3] a fictitious mass (in amu) for the lattice degrees-of-freedom has to be assigned using the PMASS-tag. The friction coefficient for lattice degrees-of-freedom have to be specified (in ps-1) by means of the LANGEVIN_GAMMA_L-tag.
The friction coefficients γ for the atomic degrees-of-freedom are specified using the LANGEVIN_GAMMA-tag.
The optimal setting for PMASS depends very much on the particular system at hand and can be considered as a compromise between two opposing factors: too large values lead to very slow variation of lattice degrees of freedom (and hence the sampling becomes inefficient) while too small value can lead to too large geometric changes in a MD step and hence may cause numerical problems. We strongly recommend to make careful tests with various settings before performing the production run.
Related tags and articles
LANGEVIN_GAMMA_L, LANGEVIN_GAMMA, MDALGO
References
- ↑ M. P. Allen and D. J. Tildesley, Computer simulation of liquids, Oxford university press: New York, 1991.
- ↑ M. Parrinello and A. Rahman, Phys. Rev. Lett. 45, 1196 (1980).
- ↑ M. Parrinello and A. Rahman, J. Appl. Phys. 52, 7182 (1981).