wesleybowman / qgsolver

another Quasi-Geostrophic solver:

Overview

This library implements a 3D z-level QG solver in regular grid or in lon/lat grid. It relies on petsc4py for parallelized PV inversions but should also work in serial in the future

Install

qgsolver

Download and install with:

git clone https://github.com/apatlpo/qgsolver.git
cd qgsolver

Install dependencies with conda

The principal dependcy is petsc4py

Download Miniconda3 from the conda website

bash Miniconda3-latest-Linux-x86_64.sh
(specify .miniconda3 and not miniconda3 as target dir for conda)

conda update conda

Create the conda environment from the enviroment.yml file, and activate the new environment for use.

conda env create -f environment.yml

source activate petsc

Run on datarmor

Generate input files if necessary on your workstation

cd qgsolver/input/
python create_input_roms.py -o roms_in/

On datarmor, copy input files if necessary:

cp -r /home/slyne/aponte/natl60/qgsolver/input/roms_in /home1/datawork/aponte/qgsolver/

Run qgsolver on Datarmor:

bash
source activate petsc
cd /home/slyne/aponte/natl60/qgsolver/run/
python run_datarmor.py qgsolver/roms_out roms.py /home1/datawork/aponte/qgsolver/roms_in/

API

See doc