SAFIREStochastic
Auxiliary-
Fields for
InteRacting
Electrons
is an efficient implementation of the auxiliary-field quantum Monte Carlo (AFQMC) method for high-precision simulations of molecules, materials, and simplified models. It is developed at the Center for Computational Quantum Physics at Flatiron Institute.

A set of random walks progressing from left to right.

Auxiliary-field quantum Monte Carlo

The AFQMC method expresses the many-body wave function of a quantum system as the integral over many noninteracting wave functions. This integral is then efficiently evaluated using Monte Carlo sampling.

For most realistic systems, the Monte Carlo sampling is complicated by a sign or phase problem, leading to intractable statistical errors. To circumvent this problem, the AFQMC sampling is constrained using a trial wave function. This introduces a bias, which can be improved by increasing the accuracy of the trial wave function.

One code – many applications

Molecules

SAFIRE interfaces with a multitude of quantum chemistry codes to generate trial wave functions and Hamiltonians to do ab initio calculations of molecules.

Materials

Efficient tensor decompositions implemented in SAFIRE enable realistic material calculations with long-range interactions.

Lattice models

An integrated toolbox allows building and simulating the Hubbard model and a breadth of its extensions on arbitrary lattices.

Getting started

To get started with an experimental version of the code,

git clone https://github.com/SFQMC/SAFIRE.git

and follow the Installation section of the documentation.

How to Cite

If you use SAFIRE in a publication, please cite our upcoming paper (check back here for more information).

Companion projects

CoQuí is an efficient implementation of the GW method. SAFIRE interfaces with to generate downfolded Hamiltonians for materials calculations.

AutoHF is a GPU-accelerated Hartree-Fock code optimized for challenging energy landscapes. SAFIRE interfaces with it to generate trial wavefunctions for lattice model Hamiltonians.