SU2xSU2 1.2

Hybrid Monte Carlo with Fourier Acceleration simulation package for the N=2 Principal Chiral model.

SU(2) x SU(2)

This python package offers efficient simulation and data analysis routines for the SU(2) x SU(2) Principal Chiral model. The key feature offered is the integration of Fourier Acceleration into the Hybrid Monte Carlo algorithm which leads to a significant reduction in the degree of critical slowing down.

Currently the simulation is only supported for a two dimensional cubic lattice.

Installation

To install SU2xSU2 using pip run:

pip install SU2xSU2

Its is recommended to work in a virtual environment.

Documentation

Read the docs here.

Example

A basic example showing how to set up a simulation using Fourier accelerated HMC to measure the wall-to-wall correlation function. Further examples can be found here.

from SU2xSU2.SU2xSU2 import SU2xSU2

# define model and lattice parameters 
model_paras = {'L':40, 'a':1, 'ell':5, 'eps':1/5, 'beta':0.6}
model = SU2xSU2(**model_paras)
# define simulation parameters and measurements
sim_paras = {'M':500, 'thin_freq':1, 'burnin_frac':0.5, 'accel':True, 'measurements':[model.ww_correlation_func], 'chain_paths':['corfunc_chain.npy']}
model.run_HMC(**sim_paras) 

Licence

SU2xSU2 is free software made available under the MIT License. For details see the LICENSE file.

To DO

• add stylesheet to distribution and uncomment its use in files
• change name of function that integrates the beta function (state that 1/beta expansion of integrand is used) and use numerical integration instead of expanding the integrand
• Runtime warning in correlations l.64
• consider restructuring: analysis module only contains simulation code and plots are made using routines in plotting module. Would allow to make plots without running simulations again
• add tests
• plotting
  • get latex error when plotting within analysis.py
  • no apparent option to add errorbar format '.' in style sheet
  • include mplstyle file in stylelib/ to be used globally. Currently, the file needs to be copied manually into the directory. Possible approaches:
    • https://github.com/garrettj403/SciencePlots/blob/master/scienceplots/__init__.py using https://github.com/matplotlib/matplotlib/blob/main/lib/matplotlib/style/core.py
    • https://matplotlib.org/stable/tutorials/introductory/customizing.html#distributing-styles
    • https://stackoverflow.com/a/52997575
    • https://stackoverflow.com/questions/35851201/how-can-i-share-matplotlib-style
• generalize simulation and data analysis to d-dimensional cubic lattice