Quantum mechanical calculations typically involve some trade-off between the accuracy of the method (i.e. how well the real-world chemistry and physics is described) and the cost of the calculation (i.e. how much of time is required on high performance supercomputers).
Many methods in widespread use suffer from either insufficient and unpredictable accuracy, or poor scaling with both system size and computational resources. The latter restriction is expected to become more important in future, as current trends in computational architectures continue to favor increased CPU core counts, with a small amount of memory per core.
In contrast quantum Monte Carlo (QMC) methods display high accuracy, scale as a small power of the system size (typically N3), require little memory, and can scale almost perfectly across many thousands of CPU cores. In addition, QMC methods are applicable to systems that exhibit strong non-dynamical correlation effects, and to excited states. Moving toward scalable, fault tolerant and low-limited methods, that preserve our needs for accuracy and transferability, paves the way for cloud-based quantum mechanical research.
Our CSIRO Molecular/Materials Quantum Monte Carlo (CMQMC) package is making these methods easier to use, and applicable to a wider range of systems and properties. It’s the user-friendly QMC code you have been looking for.
For more information, contact the Principal Developer, Dr Manolo Per.