KERNEL TRUNCATION/LCOARSEN: Difference between revisions
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
{{TAGDEF|KERNEL_TRUNCATION/LCOARSEN| .True. {{!}} .False.|.True.}} | {{TAGDEF|KERNEL_TRUNCATION/LCOARSEN| .True. {{!}} .False.|.True.}} | ||
Description: | Description: Coarsen the charge density before padding to speed up the computation of the local potential in the Coulomb-kernel-truncation method. | ||
---- | ---- | ||
The Coulomb-kernel-truncation method ({{TAG|KERNEL_TRUNCATION/LTRUNCATE}}) modifies the electrostatic potential to remove spurious interactions between periodic images in systems with reduced dimensionality (e.g. molecules or surfaces). | |||
{{TAG|KERNEL_TRUNCATION/LCOARSEN|T}} avoids significantly increasing [[Energy cutoff and FFT meshes|FFT-grid sizes]] by using a ''coarsen-before-padding'' approach: | |||
* The full electrostatic potential is first computed under standard 3D periodic boundary conditions. | |||
* A coarse representation of the charge density with <math>1/3</math> of the [[Energy cutoff and FFT meshes|FFT-grid size]] is constructed by retaining only the long-wavelength (low-|G|) components responsible for long-range interactions. | |||
The | * The truncated Coulomb kernel is applied only to this coarse density using a padded grid. | ||
This | * The final potential is obtained by subtracting the long-range periodic contribution and replacing it with the correctly truncated one. | ||
{{NB|mind| | |||
This procedure exploits the fact that [[electrostatics |long-range electrostatics]] depends only on low-frequency components of the charge density, allowing the expensive padded FFTs to be performed on a much smaller grid with minimal loss of accuracy. | |||
{{NB|mind|If {{TAG|KERNEL_TRUNCATION/LTRUNCATE|F}}, all other KERNEL_TRUNCATION tags including {{TAG|KERNEL_TRUNCATION/LCOARSEN}} are ignored.}} | |||
== Related tags and articles == | == Related tags and articles == | ||
| Line 17: | Line 21: | ||
{{TAG|KERNEL_TRUNCATION/FACTOR}} | {{TAG|KERNEL_TRUNCATION/FACTOR}} | ||
[[Category:INCAR tag]][[Category:Electrostatics]] | [[Category:INCAR tag]][[Category:Electrostatics]][[Category:2D materials]] | ||
Revision as of 10:22, 17 March 2026
KERNEL_TRUNCATION/LCOARSEN = .True. | .False.
Default: KERNEL_TRUNCATION/LCOARSEN = .True.
Description: Coarsen the charge density before padding to speed up the computation of the local potential in the Coulomb-kernel-truncation method.
The Coulomb-kernel-truncation method (KERNEL_TRUNCATION/LTRUNCATE) modifies the electrostatic potential to remove spurious interactions between periodic images in systems with reduced dimensionality (e.g. molecules or surfaces).
KERNEL_TRUNCATION/LCOARSEN = T avoids significantly increasing FFT-grid sizes by using a coarsen-before-padding approach:
- The full electrostatic potential is first computed under standard 3D periodic boundary conditions.
- A coarse representation of the charge density with [math]\displaystyle{ 1/3 }[/math] of the FFT-grid size is constructed by retaining only the long-wavelength (low-|G|) components responsible for long-range interactions.
- The truncated Coulomb kernel is applied only to this coarse density using a padded grid.
- The final potential is obtained by subtracting the long-range periodic contribution and replacing it with the correctly truncated one.
This procedure exploits the fact that long-range electrostatics depends only on low-frequency components of the charge density, allowing the expensive padded FFTs to be performed on a much smaller grid with minimal loss of accuracy.
Mind: If KERNEL_TRUNCATION/LTRUNCATE = F, all other KERNEL_TRUNCATION tags including KERNEL_TRUNCATION/LCOARSEN are ignored.
|
Related tags and articles
KERNEL_TRUNCATION/LTRUNCATE, KERNEL_TRUNCATION/IDIMENSIONALITY, KERNEL_TRUNCATION/ISURFACE, KERNEL_TRUNCATION/IPAD, KERNEL_TRUNCATION/FACTOR