Programs in Physics & Physical Chemistry
|[Licence| Download | New Version Template] aelg_v1_0.tar.gz(107956 Kbytes)|
|Manuscript Title: BerkeleyGW: A Massively Parallel Computer Package for the Calculation of the Quasiparticle and Optical Properties of Materials and Nanostructures|
|Authors: Jack Deslippe, Georgy Samsonidze, David A. Strubbe, Manish Jain, Marvin L. Cohen, Steven G. Louie|
|Program title: BerkeleyGW|
|Catalogue identifier: AELG_v1_0|
Distribution format: tar.gz
|Journal reference: Comput. Phys. Commun. 183(2012)1269|
|Programming language: Fortran 90, C, C++, Python, Perl, BASH.|
|Computer: Linux/UNIX workstations or clusters.|
|Operating system: Tested on a variety of Linux distributions in parallel and serial as well as AIX and Mac OSX.|
|RAM: (50-2000) MB per CPU (Highly dependent on system size)|
|Keywords: Many Body Physics, GW, Bethe-Salpeter Equation, Quasiparticle, Optics, Exciton.|
|PACS: 73.22.-f, 71.15.-m, 71.35.-y, 71.35.Cc.|
|Classification: 7.2, 7.3, 16.2, 18.|
External routines: BLAS, LAPACK, FFTW, ScaLAPACK (optional), MPI (optional). All available under open-source licenses.
Nature of problem:
The excited state properties of materials involve the addition or subtraction of electrons as well as the optical excitations of electron-hole pairs. The excited particles interact strongly with other electrons in a material system. This interaction affects the electronic energies, wavefunctions and lifetimes. It is well known that ground-state theories, such as standard methods based on density-functional theory, fail to correctly capture this physics.
We construct and solve the Dyson's equation for the quasiparticle energies and wavefunctions within the GW approximation for the electron self energy. We additionally construct and solve the Bethe- Salpeter equation for the correlated electron-hole (exciton) wavefunctions and excitation energies.
The material size is limited in practice by the computational resources available. Materials with up to 500 atoms per periodic cell can be studied on large HPCs.
1-1000 minutes (depending greatly on system size and processor number)
|Disclaimer | ScienceDirect | CPC Journal | CPC | QUB|