Installation and Setup

1. Setup: instructions for downloading the repository and installing the dependencies.

2. Configure Simsio Simulations: explanation of how to create and configure configuration files for simulations using the simsio library.

3. Run Simulations: instructions on how to run the simulations.

Setup

1. Download the repository from git:

   git clone --recursive git@github.com:gcataldi96/ed-lgt.git
   

2. Create the environment with all the required Python packages:

   conda env create -f ed-lgt/envs/dev.yml
   conda activate edlgt-dev
   

3. Install the library

   cd ed-lgt/
   pip install -e .
   

Exploiting SIMSIO library to manage I/O of simulations

1. Initialize the bundled simsio dependency (it should already be configured in this repository):

   git submodule update --init --recursive
   

   If you prefer a direct pip install instead of the submodule, use:

   pip install -r requirements/simsio.txt
   

This is an example of a configuration file that should be created inside the configs folder (if this folder does not exist, create it):

===:
template: |
    n$enum:
    <<<: common
    g: $g
common:
    dim: 2
    lvals: [2, 2]
    pure: false
    has_obc: false
    DeltaN: 2
    m: 1.0
n0:
    <<<: common
    g: j0
n1:
    <<<: common
    g: j1

In this file, different simulations are defined with different values of the g parameter. For example, two sets of simulations are defined with g values equal to j0 and j1.

If you want to create a larger set of simulations automatically, run a script like the following:

from simsio import gen_configs
import numpy as np

params = {"g": np.logspace(-1, 1, 10)}
gen_configs("template", params, "config_NAME_FILE")

This script will generate a series of simulations with different g values as specified in the “config_NAME_FILE.yaml” file.

Run Simulations

To run simulations, type the following command on the shell:

nohup bash -c "printf 'n%s\n' {0..N} | shuf | xargs -PA -i python SU2_model.py config_NAME_FILE.yaml {} B" &>/dev/null &

Where:

1. N is the total number of simulations specified in the configuration file config_NAME_FILE.yaml.

2. A is the number of processes in parallel.

3. B is the number of single-node threads per simulation.