What CASINO is

CASINO is a program for quantum Monte Carlo (QMC) calculations. It computes the properties of interacting electrons in atoms, molecules, crystals, surfaces and model systems by solving the many-electron Schrödinger equation with stochastic methods, working with the many-electron wave function itself rather than through an approximate functional of the density.

Two methods do most of the work. Variational Monte Carlo evaluates the energy and other properties of an explicitly correlated trial wave function and optimizes its parameters. Diffusion Monte Carlo then improves on it, projecting out the lowest-energy state that has the same nodal surface as the trial function. CASINO also implements reptation Monte Carlo, which samples pure rather than mixed expectation values.

QMC is not a black box. Its results carry statistical error bars, and a small number of systematic errors that can be named: the fixed-node approximation, by which the nodes of the trial function are imposed on the answer; the finite time step; the finite population of walkers; finite-size effects in periodic systems; and, where they are used, pseudopotentials. Each can be studied and reduced, and running QMC well is largely a matter of doing so. The method is expensive. The cost of a given statistical precision rises roughly as the third or fourth power of the number of electrons, and steeply with atomic number in all-electron work. In exchange, the calculations parallelize almost perfectly, and CASINO has run efficiently on several hundred thousand processor cores.

What it is good for

QMC is most useful where electron correlation decides the answer and cheaper methods give a spread of answers depending on the approximation chosen: binding, cohesive and adsorption energies, the relative stability of structures and defects, weakly bound systems, and reference values against which faster methods can be tested. It also suits systems that other methods find awkward: the electron gas in one, two and three dimensions, excitons and their complexes, positrons, and particles of arbitrary mass and charge.

CASINO treats finite systems and systems periodic in one, two or three dimensions. Its trial wave functions are of Slater–Jastrow form, with optional multiple determinants, backflow and pairing functions, built from orbitals expanded in Gaussians, plane waves, blip functions, Slater-type functions or numerical grids. The orbitals come from more than twenty other programs, among them Quantum ESPRESSO/PWSCF, CASTEP, ABINIT, QBOX, CRYSTAL, GAUSSIAN, GAMESS, TURBOMOLE, PSI4, DALTON, MOLPRO, CFOUR, NWCHEM, OPENMOLCAS, Q-CHEM, PySCF, ORCA, Quantum Package, and AMS/ADF.

Applications, algorithms and theory

CASINO is meant to be used, for problems whose answers matter beyond QMC itself. Quantum Monte Carlo has so far promised more than it has delivered in practical science, and narrowing that gap is the main aim of the present work. The methods are not finished either, and the code is also a place to study them: where the systematic errors of diffusion Monte Carlo come from and how to control them, how long a calculation must run to reach a stated precision, what the nodal surface of a trial wave function does to the energy, and how the random walks of QMC relate to the trajectories of the de Broglie–Bohm pilot-wave theory. The distribution includes LOUIS, a small program for calculating such trajectories.

History

The research programme that led to CASINO was begun in the early 1990s by Richard Needs in the Theory of Condensed Matter group at the Cavendish Laboratory, Cambridge. Mike Towler, who joined the group in 1996, generalized its test codes into a single program that could treat any system, and the first public version of CASINO appeared in 1999; the name was suggested by Paul Kent. Neil Drummond, from 2002, and Pablo López Ríos, from 2004, made major improvements and extensions to the code, and many others contributed over the years. Through the 2000s and early 2010s CASINO was developed actively and used by groups around the world.

After that it was maintained rather than developed. There was no official release between 2012 and 2026, and the documentation and the links to other programs fell behind the code.

CASINO in 2026

CASINO v3.0, released in August 2026, ended that gap, and this autumn Mike Towler began a thorough overhaul of the whole program which is still under way:

• Speed. The code is profiled on realistic calculations and its costly parts rewritten, each change measured.
• Correctness. Every change is run against the full test suite and reviewed before release, and the faults this has brought to light, some of them long-standing, are being fixed.
• Methods. A DMC run can now stop by itself when a requested statistical error is reached.
• Interfaces. The converters that read orbitals from other programs are being checked one by one.
• Documentation. The manual is being rewritten to explain the methods as well as list the options.
• Building. The build has been brought up to date for current compilers and machines.

More than six hundred changes have been made since v3.0, each described in the public DIARY.

People and contact

CASINO is looked after by Neil Drummond at Lancaster University, Pablo López Ríos in Stuttgart, and Mike Towler, who works from his hideaway in the Apuan Alps at the Vallico Academy in Tuscany.

CASINO is free for academic and other non-profit use: request a copy, and please cite it in work that uses it. Anyone interested in using CASINO, in its methods, or in contributing to it is welcome to get in touch, or to join the discussion on the forum.