| CASINO (program) | |
|---|---|
| Name | CASINO |
| Title | CASINO |
| Caption | Quantum Monte Carlo package |
| Developer | University of Cambridge (original), Cavendish Laboratory, contributors |
| Released | 1990s |
| Programming language | Fortran, C (bindings) |
| Operating system | Unix-like systems, Linux, macOS |
| Genre | Scientific software, Quantum Monte Carlo |
| License | Academic / research |
CASINO (program)
CASINO (program) is a computational package for performing Quantum Monte Carlo (QMC) simulations of electronic systems. It implements variational and diffusion Monte Carlo methods to compute highly accurate ground-state energies and properties for molecules, solids, and model Hamiltonians, playing an important role in benchmark studies and methodological development in quantum chemistry and condensed matter physics.
CASINO was designed to provide a robust, well-tested environment for performing many-body electronic structure calculations beyond mean-field methods such as Hartree–Fock and Density functional theory. Its purpose is to supply researchers with tools to obtain correlated electronic energies, excitation gaps, and expectation values with controlled statistical errors, supporting studies in materials science, molecular chemistry, and model systems like the homogeneous electron gas. CASINO supports research workflows that require high accuracy for cohesion of theoretical predictions with experimental measurements from facilities such as synchrotrons and quantum materials laboratories.
The program is founded on stochastic projector techniques and importance-sampled path-integral methods developed in the QMC community. Central theoretical elements include Variational Monte Carlo (VMC), which optimizes trial wave functions, and Diffusion Monte Carlo (DMC), which projects out ground states using imaginary-time evolution under the fixed-node approximation. CASINO leverages the fixed-node approximation to enforce fermionic antisymmetry and employs trial wave functions built from Slater determinants and Jastrow factors. The package interfaces with pseudopotential formalisms such as Hartree–Fock pseudopotentials and Norm-conserving pseudopotential schemes to treat core electrons, connecting to broader electronic-structure approaches like plane-wave basis and Gaussian orbitals representations.
CASINO implements a spectrum of algorithms tailored to quantum many-body problems: - Variational optimization routines for Jastrow, multi-determinant, and backflow parameters using energy and variance minimization. - Diffusion Monte Carlo with importance sampling, time-step extrapolation, and population-control measures. - Finite-size correction techniques for periodic systems, including twist averaging and model periodic Coulomb corrections. - Support for backflow transformations and multideterminant expansions to reduce fixed-node errors. - Efficient evaluation of electron–electron and electron–ion interactions using Ewald summation and optimized locality approximations for pseudopotentials. These algorithmic capabilities connect CASINO to methodological advances from investigators in groups at institutions such as the University of Cambridge, University of Oxford, Princeton University, and national laboratories.
CASINO is primarily written in Fortran with modular components and optional C bindings for I/O and analysis. The architecture separates trial-wavefunction generation, Monte Carlo driver, and utility routines for property estimations. Interfaces exist to import single-particle orbitals from external electronic-structure codes like Quantum ESPRESSO, VASP, Gaussian, CRYSTAL, and CASTEP. Parallelization is achieved through MPI and task-parallel schemes, enabling use on high-performance computing resources including national supercomputing centers and university clusters. CASINO's data formats and tools facilitate reproducible workflows and coupling with optimization packages and visualization tools common in computational physics and chemistry.
CASINO has been applied to a broad range of systems: bulk solids (semiconductors, insulators, and correlated oxides), low-dimensional materials (graphene, nanotubes), molecular binding and reaction energetics, and benchmark studies of small atoms and molecules. Notable application areas include cohesive energy calculations to validate density functional theory approximations, band-gap predictions complementary to GW approximation and Bethe–Salpeter equation approaches, and studies of electron correlation in model Hamiltonians such as the Hubbard model. Its results frequently inform experimental interpretations from techniques like photoemission spectroscopy and inelastic neutron scattering.
CASINO's accuracy has been validated against high-level quantum-chemical methods such as Coupled cluster (e.g., CCSD(T)) and against experimental thermochemical data. Extensive benchmarking on molecules and solids has established its reliability for cohesive energies, atomization energies, and quasiparticle gaps, with careful analysis of time-step, population, and finite-size errors. Performance scaling has been demonstrated on multi-node clusters using MPI parallelism and optimization for processor architectures common in national computing facilities. The code is used in intercomparison studies among packages including QWalk, QMCPACK, and emerging QMC implementations.
CASINO's development traces to efforts at the Cavendish Laboratory and collaborators in the 1990s, evolving through community contributions from academics, postdoctoral researchers, and national laboratories. The project benefits from an engaged user base that contributes tests, pseudopotentials, and validation scripts. Workshops and conferences such as the International Workshop on Quantum Monte Carlo Methods and sessions at the American Physical Society meetings have fostered collaboration. Development practices emphasize stability, reproducibility, and long-term maintenance consistent with traditions in academic scientific software stewardship.
Category:Quantum Monte Carlo Category:Computational chemistry software Category:Physics software