An interface package to use Torch (the open-source numerical library for machine learning) from Python. Users can easily implement deep learning based on neural networks, and can use various state-of-the-art methods. This package supports GPGPU parallel computation, and realises high-speed operation. A front-end interface for C++ is also prepared.
Software to calculate physical quantities related to phonon in solids from result calculated using first principles calculation software. Results calculated from first principles calculation software such as VASP and Wien2k are used as an input.
An open-source application for quantum chemical calculation. This package implements various methods for quantum chemical calculation such as Hartree-Fock approximation, density functional theory, coupled-cluster method, and CI (configuration interaction) method. The package is written in C++, and provides API for Python, by which users can perform for preparation of setting and execution of calculation.
An open-source application for molecular simulations. This application supports various methods such as classical and ab initio molecular dynamics, path integral simulations, replica exchange simulations, metadynamics, string method, surface hopping dynamics, QM/MM simulations, and so on. A hierarchical parallelization between molecular structures (replicas) and force fields (adiabatic potentials) enables fast and efficient computation.
Pomerol is an app for calculation one- and two-body Green’s function at finite temperatures for the Hubbard-type model based on the full exact diagonalization. Pomerol is written in C++ and supports the hybrid parallelization (MPI+openMP).
Open source library to record execution and communication time during specified regions in user’s program. C/C++ and Fortran API are provided. This can profile MPI & OpenMP hybrid parallel programs as well as serial ones.
PHYSBO is a Python library for researchers mainly in the materials science field to perform fast and scalable Bayesian optimization based on COMBO (Common Bayesian Optimization). Users can search the candidate with the largest objective function value from candidates listed in advance by using machine learning prediction. PHYSBO can handle a larger amount of data compared with standard implementations such as scikit-learn.
An application for three-dimensional visualization with the ray tracing method. This application can visualize arbitrary positions and shapes of objects such as spheres and cubes. It can visualize three-dimensional data obtained from computational fluid dynamics etc. by volume rendering. It can also be used for simple three-dimensional graphical simulator with macro functions.
A python package for the tight-binding method. PythTB supports tight-binding calculations of electronic structures and Berry phase in various kinds of systems. Users can use ab initio parameters obtained by Wannier90.
peps-torch is a python library for calculation of quantum many-body problems on two dimensional lattices. Variational principles calculation is used with an infinite PEPS (iPEPS) as the trial wave function. Therefore, the ground state is obtained in the form of the element tensor of the iPEPS. The energy of the trial state is estimated by the corner transfer matrix method (CTM), and its gradient with respect to the element tensor is computed through automatic differentiation provided by pytorch. Functions/classes for exploiting the system’s symmetry are provided for reducing the computational cost if possible. While general models and lattices are not supported, many examples of stand-alone codes would make it relatively easy for users to write their own codes to suit their needs. pytorch is required.