This project has been a cooperation between Carl Petter Duedahl, Daniel Hass and me. The code structure is based on this template (thank you).
In this project we have tried to simulate a bosonic system of particles in a magnetic trap, and find the lowest energy state for this system. To do this, we have used tools such as Variational Monte Carlo, Gradient Decent and the Blocking Method. This file will show how to navigate our repository and use our programs. The main calculations are written in C++14, with analysis, plotting and statistical analysis performed using Python3.
To run and compile the C++, a version of openMP is required. The compilation and linking is performed using CMake. A handy script compile_project is provided for easier compilation. Run
./compile_projectOr perform manually, using
# Create build-directory
mkdir build
# Move into the build-directory
cd build
# Run CMake to create a Makefile
cmake ../
# Make the Makefile using two threads
make -j2The linking with openMP might differ based on your system. Currently, it is configured to compile for Mac (M1 chip). Configure the CMakeLists.txt for your desired system if problems are encountered.
After compilation, five executables should be presented in the build directory. With some overlapping functionality, they briefly do:
vmc: Perform a single non-interactive vmc calculation.timing: Perform timing on the non-interactive vmc calculation, using either analytical or numerical double derivatives.gradient: Performs gradient decent on either a non-interactive or interactive system. .interact: Perform a single interactive vmc calculation.parallelinteract: Perform multiple interactive vmc calculations in parallel.
Usage can be seen by running the programs with no arguments.
The directory Analysis contains Python scripts to create plots and perform statistical analysis. To run (all) of these, the following packages are required.
cycler==0.11.0matplotlib==3.5.1numpy==1.21.5pandas==1.4.4seaborn==0.12.2