A repository for config and job script files for the NEW study of Z without S1

Run the following command in the ~/packages/ directory on the harvard machine
git clone https://github.com/kvjmistry/NEWDiffusion.git

The job config folder contains the nexus and IC configuration files. You should not need to edit these files. Preferably use these as templates and edit them directly from the jon submission script.

To launch jobs, go into the job folder. You will see the following files:

NEW_MC208_NN_job.sh
launch_jobs.sh
cleanup_jobs.sh

The NEW_MC208_NN_job.sh is the template bash script which executes a set of commands on each of the computers on each job.

The launch_jobs.sh uses NEW_MC208_NN_job.sh as a template, it creates a separate log directory and replaces the diffusion value we want to simulate in the file using the array: declare -a StringArray=("0.1" "0.2" "0.3" "0.4" "0.5"). You can configure this array depending on the diffusion values you want to run.

To launch the jobs, you can do source launch_jobs.sh. The default number of jobs are set to run 10 jobs 1 - 10 in the sbatch command at the end. You can increse the number 10 to 100 if say you want to run 100 jobs etc. Each job generates 2 x 50000 events in nexus. You can also increase/decrease the number of events in each job in the NEW_MC208_NN_job.sh to as you need. Note that by the time the events get to esmeralda the final job output is about 50 events.

The cleanup_jobs.sh can be sourced to clean up the everything if you want to start a fresh. Make sure the array configured is the same one as in launch_jobs.sh.

The output files will be located in the following area: $SCRATCH/guenette_lab/Users/$USER/NEW_MC208/$Diff/"

The final output files are called:
NEW_Tl208_ACTIVE_esmeralda_jobid_<jobid>_[1..5]_Diff<diffusion value>.next.h5