danosu / vah_argonne

Run VAH model Relativistic Heavy Ion Collisions in Argonne

We will run the simulations in Argonne Bebop cluster using singularity containers.

Singularity Containers

Follow these steps to get the singularity image

1. Load the singularity module in Bebop cluster

module load singularity

2. Get the container image and make a sandbox folder with write privileges in the home directory

cd ~/
singularity build --sandbox jetscape_vah docker://dananjaya92/jetscape-bulk:v3

Now you have a folder named jetscape which has the container files. Changes you make to this folder will apear in any spawned jetscape containers. For more information on building containers with a sanbox please refer to singularity documentaion

Run Simulations

1. Clone this repo as sims_run_events folder

git clone https://github.com/danOSU/vah_argonne.git -b sims vah_run_events

2. The folowing will submit five relativistic heavy ion collision simulations to a single node only using 5 CPUs. All the simulations input parameters are set to Maximum a Posteriori(MAP) values found in this paper

sh submit.sh "My first simulation run" MAP

With each submission a history.txt file is updated with the Job_ID and with the comment made at the submision. To run the simulation with a different set of input parameters you need to generate and put the relevent input files inside a new folder in MAP. Please take a look at the MAP input files and the python script that generate these input files by going to MAP to understand what input files are needed to run an event. If you put new set of input files in a directory new_design_0 events with this input files can be run by following command.

sh submit.sh "New set of model parameter values" new_design_0

To change the resources used and to increase the number of events run for a single design point look at submit_argonne_vah file.

Simulation Results

The raw simulation results can be found in SCRATCH/results/JOB_ID folder. Results forlder contain the final particle lists as .dat files for each event. These results need to be combined for a proper analysis. This is done by check_cat_move_results.sh. The file with all the particles from all the simulated events can be found in MAP after you run >sh check_cat_move_results.sh JOB_ID MAP. Similarly in SCRATCH/logs/JOB_ID you can find the logs for each event.

sh check_cat_move_results.sh JOB_ID MAP Results from each individual run is concatenated and being put in the same folder that had the input files for the simulation runs. For the example run it would be in MAP. In the same folder there is JOB_ID.log file which has the printed standard output of the job and several .txt files with more information about the succesful completion/ failure of each individual event.