Skip to content

Latest commit

 

History

History
48 lines (35 loc) · 2.77 KB

readme.md

File metadata and controls

48 lines (35 loc) · 2.77 KB
Raw

Build Instructions

Building on Bessemer HPC

  1. Connect to the Bessemer HPC system and login
  2. Run the command git clone https://github.com/MILeach/TB_Model.git to copy the project files into your user area.
  3. Run cd TB_Model to move into the project directory
  4. Run sbatch build.sh to submit an HPC job which will build the project

Building locally on linux

Requirements for building locally

  • Nvidia graphics card
  • A valid CUDA installation

Build Instructions

  1. Open a terminal and navigate to the folder where you wish to set up the project
  2. Run the command git clone https://github.com/MILeach/TB_Model.git to copy the project files into your user area.
  3. Run cd TB_Model/FLAMEGPU/examples/TB_Model to move into the project directory
  4. Run make to build the project

Setting up the Model

  • Navigate to the TB_Model/FLAMEGPU/examples/TB_Model/initializer folder
  • Modify histo.csv to adjust the number of people in each sex/age category
  • run python3 preprocess.py data.in to generate the data.in file which describes the population
  • For an HPC build, copy the created data.in file to the TB_Model/input folder. For a local build, copy the file to the TB_Model/FLAMEGPU/examples/TB_Model/input folder.
  • For an HPC build, create a folder inside the TB_Model/input folder and copy the network.json file into it.
  • Parameters are controlled using the FLAMEGPU XML input files. For an HPC build, place any FLAMEGPU XML input files you wish to run together in a folder in the TB_Model/input directory. For a local build, place your FLAMEGPU XML input file in the TB_Model/FLAMEGPU/examples/TB_Model/iterations folder.

Running a Simulation

Running a simulation on Bessemer HPC

  1. Navigate to the top level TB_Model directory
  2. Ensure the input folder contains the correct data.in file and a folder containing a copy of network.json and the XML input files you wish to run as described in the Setting up the Model section
  3. Run the command sbatch run.sh input_folder_name number_of_iterations where input_folder_name is the name of the folder that contains the XML files and number_of_iterations is the number of timesteps you wish to run the simulation for
  4. When the job is completed, the output files should be available in the output folder

Running a simulation locally on linux (single XML file only)

  1. Navigate to the TB_Model/FLAMEGPU/examples/TB_Model directory
  2. Ensure the XML input file, network.json file and data.in file are present in the iterations folder
  3. Run the command ./bin/linux-x64/Release_Console/Project iterations/input_filename.xml number_of_iterations XML_output_frequency 0

Other Bessemer HPC utilities

  • sacct -v - lists your active jobs
  • scancel job_number - cancels the job with id job_number