mirror of https://gitlab.com/QEF/q-e.git
6148f5e5c5
See https://stackoverflow.com/questions/50148175/what-does-cd-echo-0-sed-s-1-do-in-bash-script Note: some trailing blanks have been removed as well by the script I used. Use "git diff -b" to see only the true changes. |
||
|---|---|---|
| .. | ||
| reference | ||
| README | ||
| run_example | ||
README
This example illustrates how to use pw.x to compute the equilibrium geometry of a simple molecule, CO, and of an Al (001) slab. The calculation proceeds as follows (for the meaning of the cited input variables see the appropriate INPUT_* file) 1) make a geometry relaxation for CO molecule performing a series of self-consistent calculations and computing the forces on atoms (input=co.rx.in, output=co.rx.out). The molecule is put in a cubic box of side 12 Bohr. Note that ibrav=0 therefore the Bravais lattice fundamental vectors are read after cards 'CELL_PARAMETERS' (where we also specify the type of symmetry, cubic or hexagonal). The cell parameter is not specified in celldm(1), but deduced from Bravais lattice vectors. Calculation is set to 'relax', so specifying that a structural relaxation is performed. While approaching the minimum, the scf threshold (initially conv_thr=1.0d-8) will automatically become smaller (stricter convergence) because of the need to evaluate correctly forces and the tiny energy differences involved in the relaxation. This tightening of the scf threshold is however limited by the upscale=10 statement that specifies that conv_thr ccannot become smaller than its starting value / upscale**2 (=1.0d-10 in the present example). 2) make a geometry relaxation for a Al (001) slab performing a series of self-consistent calculations and computing the forces on atoms (input=al001.rx.in, output=al001.rx.out). This is a 7-atomic-layer slab separated by about 4 vacuum layers. The unit cell in tetragonal (ibrav=6) with celldm(1)=alat_fcc/sqrt(2). Calculation is set to 'relax'. While approaching the minimum, the scf threshold (initially conv_thr=1.0d-6) will automatically become smaller (stricter convergence) because of the need to evaluate correctly forces and the tiny energy differences involved in the relaxation. This tightening of the scf threshold is however limited by the upscale=10 statement that specifies that conv_thr ccannot become smaller than its starting value / upscale**2 (=1.0d-8 in the present example). 3) make a geometry relaxation for the same Al (001) slab used in step 2 performing a series of self-consistent calculations, computing the forces on atoms and evolving the atomic positions according to Newton equation. Whenever a velocity component is opposite to the corresponding force component, the velocity is stopped. (input=al001.mm.in, output=al001.mm.out).