This documentation was last updated on 2023.10.26
This repository contains scripts for setting up and running GIZMO simulations. The scripts automate the process of configuring GIZMO, submitting jobs to clusters, and generating initial conditions (ICs) for a Zel'dovich pancake simulation.
I use this repository to set up and run my GIZMO simulations. All things directly related to running GIZMO that I use are in this repository. If you are interested in analysis of simulations, check out my GIZMO-analysis repository: https://github.com/Vasissualiyp/GIZMO-analysis
All but one scripts are located in the scripts folder. It is recommended that you run all scripts from current directory, i.e. ./scripts/gizmo_setup.sh.
Most of the scripts automatically detect your system (currently supported: ScieNet Niagara and CITA Sunnyvale), and execute appropriate commands.
For more information about the structure of the folders, and information about the code, check DOCUMENTATION.md.
The gizmo_setup.sh script automates the process of cloning the GIZMO repository, configuring the Makefile for different clusters (Niagara or Starq), creating a Config.sh file, and copying the TREECOOL file. Currently it also includes downloading spcool tables for GIZMO native cooling library.
This script allows you to choose which of the available repositories you would liek to clone (currently supported: public, private, STARFORGE). This script also lets you choose one of the available tools that can be useful for working with gizmo (MUSIC, Rockstar, GRACKLE).
The compile_autosub.sh script (the only script not in the scripts folder) automates the process of setting up and submitting a job to a cluster. It extracts the job name, creates a unique output directory, modifies necessary parameters based on the host system, and submits the job. Additionally, it tracks changes made to specific parameters in the ./gizmo/Config.sh and ./template/zel.params files between job submissions, logging additions, edits, or removals. These changes are archived by date, and the current versions of the files are stored for comparison with the next job submission.
The script sqc.sh allows you to quickly view and cancel running jobs on a cluster using either the squeue or qstat commands, depending on the host system. It simplifies job management by providing a list of your current jobs with their IDs and runtimes, and lets you easily specify which jobs you want to cancel by their indices. The script updates the job progress in real time.
The script job_tracker.sh works like sqc.sh, except it tracks the current timestep of the latest simulation.
The script day_clear.sh clears all jobs from today and resets the directory as if no jobs were ran today.
The script performance_report.sh writes current timestep of the simulation and real time since the script was started into a csv file in the directory where the simulation is running. This script is useful if you want to track how fast does the simulation progress with time. You should run this script at the same time as the job is running. What I would suggest is running this script at the same time with sqc.sh to make sure that you are not tracking inactive jobs. For this script, it is also very important to set its maximum time of running if you are going to make it run automatically with nohup (in the background, with possibility of logging out)
heatmap_generator.py and runtime_check.ccp are legacy scripts that were used to check the dependence of short run speed (max timestep) on a pair of parameters. They create a heatmap that should in theory show which jobs perform the best. Should be used in combination with generate_configs.py script for generation of 'grid' of parameters for the heatmap and with compile_autosub.sh script for batch job submission. Currently the axes for heatmap have to be set manually in the script.
The snapshottimes_generator.py script generates a geometric sequence of snapshot times between a start time and an end time, and writes these times to a snapshot_times.txt file.
The zeldovich_ics_gen/hdf5zelgenglass.py script generates 3D initial conditions (ICs) for a Zel'dovich pancake simulation, and writes these ICs to an HDF5 file.
To run these scripts, navigate to the directory containing the scripts and execute them. Note that the scripts may need to be modified to suit your specific simulation setup and computing environment.