Skip to content
Snippets Groups Projects
Commit f1ba9870 authored by Ruocheng Han's avatar Ruocheng Han
Browse files

Update README.md

parent 1258123a
No related branches found
No related tags found
No related merge requests found
......@@ -25,28 +25,93 @@ Legendre quadrature (200 radial points) and Lebedev quadrature (noted as 031, 35
```shell
# prepare wavefunction and FCIDUMP files, here we use Psi4 package
# one can also use other packages, e.g. PySCF, OpenMolcas
cd examples/ch4_200_031/calcu/
psi4 -n 8 -i input.dat -o output.dat
$ cd examples/ch4_200_031/calcu/
$ psi4 -n 8 -i input.dat -o output.dat
# convert molden format of wavefunction and txt format of FCIDUMP (in Psi 4 is INPDUMP)
# into .npy files
# Note that external package is needed to convert molden to .wfn first, we use molden2aim
# (https://github.com/zorkzou/Molden2AIM) for example.
cd ../prep/
chmod +x mlcorr_prep.sh
./mlcorr_prep.sh ch4
$ cd ../prep/
$ chmod +x mlcorr_prep.sh
$ ./mlcorr_prep.sh ch4
# calculate MP2 MP3 MP4 energy density
cd ../eneden/
chmod +x eneden.sh
./eneden.sh ch4 # Here STDOUT will give the values of summing up of the energy density,
$ cd ../eneden/
$ chmod +x eneden.sh
$ ./eneden.sh ch4 # Here STDOUT will give the values of summing up of the energy density,
# which should equals to MPn correlation energy
# calculate reference value:
cd ../ref/
psi4 -n 8 -i input.dat -o output.dat # you can find reference MP2, MP3, and MP4
$ cd ../ref/
$ psi4 -n 8 -i input.dat -o output.dat # you can find reference MP2, MP3, and MP4
# correlation energy in output.dat
```
Results: contribution for each atom (Atomic Contribution)
```bash
PT2:
C: -0.0713488682277866
H: -0.0072268498366822
PT3:
C: -0.0056150246921240
H: -0.0019143657456038
PT4:
C: -0.0031286420616285
H: -0.0005634905197682
Combine:
MP2: -0.1002562675745154
MP3: -0.1135287552490546
MP4(SDTQ): -0.1189113593897559
Compare with reference correlation energy:
MP2: -0.099901794206
MP3: -0.113086192916
MP4(SDTQ): -0.118437326291
```
~0.5 mH deviation is due to the integration issue. One can decrease by further increase
the density of quadratures, or redistribute the deviation to each atom.
### ch4_50_011
This is to calculate energy density on Carbon and Hydrogen atom of CH4@3-21g with grid settings:
Legendre quadrature (50 radial points) and Lebedev quadrature (noted as 011, 50 angular points).
**Note that this example settings is not recommended for calculation**
```shell
# First you need to modify:
codes/prep/Becke_grid.py:
qd.line_segment.gauss_legendre(200) -> qd.line_segment.gauss_legendre(50)
qd.sphere.lebedev_031() -> qd.sphere.lebedev_011()
# Others same with ch4_200_031
```
Results: contribution for each atom (Atomic Contribution)
```bash
PT2:
C: -0.0731792208782222
H: -0.0073744070379437
PT3:
C: -0.0059725794473153
H: -0.0019610653069297
PT4:
C: -0.0032755188689802
H: -0.0005762340521139
Combine:
MP2: -0.102676849029997
MP3: -0.116493689705031
MP4(SDTQ): -0.1220741447824669
Compare with reference correlation energy:
MP2: -0.099901794206
MP3: -0.113086192916
MP4(SDTQ): -0.118437326291
```
~3.5 mH deviation is too large, so this setting is **not recommended**.
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment