Skip to content

Latest commit

 

History

History
126 lines (88 loc) · 3.84 KB

README.md

File metadata and controls

126 lines (88 loc) · 3.84 KB
Raw

CI

PlanckTon-Flow

PlanckTon-flow is a lightweight dataspace manager that leverages the Signac framework to submit molecular dynamics simulations of organic photovoltaics using PlanckTon. PlanckTon-flow works with Singularity and is designed for use on supercomputing clusters.

Install

PlanckTon-flow uses the conda package manager. Before installing PlanckTon-flow, please install Miniconda.

  1. First download PlanckTon-flow:

    git clone [email protected]:cmelab/planckton-flow.git
cd planckton-flow

  2. Then install its requirements:

    conda env create -f environment.yml
conda activate planckton-flow

    PlanckTon-flow is not a python package, so it does not need to be installed.

  3. In order to use PlanckTon-flow, the PlanckTon container must be pulled to your machine and its location assigned the environment variable $PLANCKTON_SIMG.

    The following example shows the container pulled to a directory called ~/images:

    cd ~/images
singularity pull docker://cmelab/planckton_gpu:latest
export PLANCKTON_SIMG=$(pwd)/planckton_gpu_latest.sif

    Or you can run this command (while still in the directory where you pulled the image) to add the image location to your bashrc file so you never have to run this step again

    echo "export PLANCKTON_SIMG=$(pwd)/planckton_gpu_latest.sif" >> ~/.bashrc

And that's it--you are ready to run simulations!

Run

Pre-run steps (Click to expand):

(These commands can be added to your .bashrc to save time.)

  1. Make sure singularity is available,

    Fry:

    module load singularity

    Bridges2: singularity is loaded by default

  2. CUDA libraries are on your path,

    Fry:

    module load cuda

    Bridges2:

    module load cuda/10

  3. The conda environment is active,

    conda activate planckton-flow

  4. And the PLANCKTON_SIMG variable is set,

The basic workflow is something like this:

  1. Edit the init file to define state point space

    vim src/init.py

  2. Run the init script to create a workspace

    python src/init.py

  3. Check to make sure your jobs look correct

    python src/project.py submit --pretend

  4. Submit the project script to run your simulations

    python src/project.py submit

    src/project.py contains all of the job operations.

Cluster support

Beyond the officially supported flow environments we support:

  • R2
  • Borah
  • Fry

Included shell scripts

clean.sh clean.sh is a shell script file that will remove the workspace folder, including the files inside of it, along with any files matching status.txt, signac*, *.log, *.gsd, or *.out within the present planckton-flow clone you are using. To run, type the following in your terminal: 
bash clean.sh

THIS WILL DELETE ANY JOBS YOU HAVE RUN SO FAR! MAKE SURE YOU ARE IN THE CLONE OF PLANCKTON-FLOW THAT YOU WANT THE WORKSPACE REMOVED FROM!

print_status.sh print_status.sh is a shell script file that prints out the status of your recently submitted jobs and saves this information to a file called status.txt. To run, type the following in your terminal: 
 bash print_status.sh