mVMC—Open-source software for many-variable variational Monte Carlo method

Abstract

mVMC (many-variable Variational Monte Carlo) is an open-source software based on the variational Monte Carlo method applicable for a wide range of Hamiltonians for interacting fermion systems. In mVMC, we introduce more than ten thousands variational parameters and simultaneously optimize them by using the stochastic reconfiguration (SR) method. In this paper, we explain basics and user interfaces of mVMC. By using mVMC, users can perform the calculation by preparing only one input file of about ten lines for widely studied quantum lattice models, and can also perform it for general Hamiltonians by preparing several additional input files. We show the benchmark results of mVMC for the Hubbard model, the Heisenberg model, and the Kondo-lattice model. These benchmark results demonstrate that mVMC provides ground-state and low-energy-excited-state wave functions for interacting fermion systems with high accuracy.

Single-Particle Spectral Function Formulated and Calculated by Variational Monte Carlo Method with Application to d -Wave Superconducting State

“…In the variational Monte Carlo method, we postulate an ansatz, a variational state jψi that can be used to calculate the physical quantities associated with that state. To find a good approximation to the ground state, we optimize the variational state in order to minimize the energy measured [35,38].…”

“…The computational cost of the ground-state determination (step 1) scales as N s × OðN 3 p Þ, where N p is the number of variational parameters f ij , g i , and v ij defined in Eq. (14) [35] and N s is the number of Monte Carlo samples. Some constraints can and must be used at times to reduce the degrees of freedom (N p ) of the variational ground states and thus reduce the calculation time.…”

“…Finally, in the Supplemental Material [34], the reader can access the fully functional open-source code that was used to generate the data. The code is an extension based on the open-source code mVMC [35].…”

Electronic correlation and geometrical frustration in molecular solids: A systematic ab initio study of β′−X[Pd(dmit)2]2Misawa1, Yoshimi2, Tsumuraya3 2020Phys. Rev. ResearchSelf Cite14280View full textAdd to dashboardCite

“…The analysis of the low-energy effective model in this Rapid Communication is limited to small system sizes in order to perform an exact analysis to clarify the general trend in the Pd(dmit) 2 salts. Several highly accurate wave-function methods for strongly correlated electron systems have recently been developed [52][53][54]. A clarification of the possible exotic elementary excitation of the QSL state in EtMe 3 Sb using such cutting edge methods, which is a hotly debated issue in experiments [12][13][14]16], is a significant challenge but is left for future studies.…”

Double-Precision FPUs in High-Performance Computing: An Embarrassment of Riches?

“…Our input is the wat222 example which distributes 156,240 atoms over a 16×16×16 cell domain. d) many-variable Variational Monte Carlo (mVMC) method: implemented by this mini-app is used to simulate quantum lattice models for studying the physics of condensed matter [28]. We use mVMC's included strong-scaling test, but downsize it (1/3 lattice dimensions and 1/4 of samples).…”