AiiDA-qp2

Installation

This is a brief installation guide fo aiida-qp2. First one needs to setup properly aiida it should be installed when package is installed. Also RabbitMQ is importat to run aiida.

# It is better to run it in some kind of virtual environment.

$ pip install aiida-qp2

Command line code completion with TAB is very usefull, dont hesitate to activate it and use it:

eval "$(_VERDI_COMPLETE=bash_source verdi)"

AiiDA profile

First we need to setup profile. One can use postgreSQL as a db backend but for simplicity we will show you SQLite setup:

$ verdi profile setup core.sqlite_dos

You will be promted some questions but answering them is very easy. Setting up PostgreSQL more challenging, however, it is recomended for larger projects.

AiiDA computer

Now you need to specify computer where you want to run your calculations. Here is the setup for your local computer. First setup a yaml file with setting called localhost.yaml and then setup the code.

$ cat > localhost.yaml
---
label: "localhost"
hostname: "localhost"
transport: "core.local"
scheduler: "core.direct"
work_dir: "/var/tmp/aiida"
use_double_quotes: true
mpirun_command: "mpirun -np {tot_num_mpiprocs}"
mpiprocs_per_machine: "8"
prepend_text: " "
append_text: " "

$ verdi computer setup --config localhost.yaml

One can test the computer via verdi computer test <label>.

AiiDA code

Lastly we need to setup the code. Again to do the most simple setup one can pull docker image from docker hub.

$ docker pull docker://addman151/qp2-aiida:tagname

# you can use    tagname:latest - for serial qmcchem
# or you can use tagname:intel  - for parallel

Unfortunatelly, the image is litle bit large, something what has to be fix in the future. When the image is downloaded we can continue by creating new code node inside AiiDA. Again we will use an yaml file.

$ cat > qp.yaml
engine_command: "docker run -v $(pwd):/data {image_name}"
wrap_cmdline_params: false
use_double_quotes: false
image_name: 'qp2-aiida'
label: 'qp2-docker'
description: 'quantum package 2'
default_calc_job_plugin:
  - 'qp2.create'
  - 'qp2.run'
filepath_executable: '/bin/entrypoint.sh'
computer: 'localhost'
prepend_text: ' '
append_text: ' '

$ verdi code create core.code.containerized --config qp.yaml

Now run verdi status and you should not see any error messages:

 ✔ version:     AiiDA v2.5.1
 ✔ config:      /home/.../.aiida
 ✔ profile:     profile
 ✔ storage:     SqliteDosStorage[...]: open,
 ✔ rabbitmq:    Connected to RabbitMQ v3.8.9 as amqp://guest:guest@...

Running AiiDA qp2

N.B.: One can execute console script either through verdi data qp2.cli command or with a shortcut aqp

echo "2

H 0.0 0.0 0.0
H 0.7 0.0 0.0" > H2.xyz

verdi data qp2.cli create Hydrogen2 --code qp2-docker@localhost --structure H2.xyz
# Or equivalent with:
# aqp create --code qp2-docker@localhost --structure H2.xyz

verdi data qp2.cli list # Find pk of your project

verdi data qp2.cli activate <pk> # preplace by pk

verdi data qp2.cli run scf # Finally running HF calculation

verdi data qp2.cli show # Look at the results

I wrote these command from head hopefully there are no mistakes.

Running QMC=Chem

Pull QMC=Chem docker image

docker pull docker://addman151/qmcchem-aiida

You can install this code in the same way how the qp code was installed. In fact this image contain also qp so you can you this image also for previous calculations.

Recomended workflow follows:

aqp run save_for_qmcchem --trexio-bug-fix

The option --trexio-bug-fix is there because there is a bug where one has to manually put absolute path to the trexio file. This flag will fix it.

Now run QMC=Chem:

aqp run qmcchem -p "-t 1800" -p "-l 20"