QAssemble: A Pure Python Package for Quantum Many-Body Theory
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917432735563776 |
|---|---|
| author | Mo, Seongjun Li, Dongming Han, Mancheon Jönsson, Johan Kang, Byungkyun Lee, Hoonkyung Kotliar, Gabriel Choi, Sangkook |
| author_facet | Mo, Seongjun Li, Dongming Han, Mancheon Jönsson, Johan Kang, Byungkyun Lee, Hoonkyung Kotliar, Gabriel Choi, Sangkook |
| contents | QAssemble is a pure-Python package for the quantum many-body problem. It implements various functional approaches, such as tight-binding, Hartree-Fock, and GW approximations within a unified object-oriented architecture. Each physical concept--crystal structure, Hamiltonian, Green's function, self-energy, polarizability, screened Coulomb interaction--is represented as a distinct class. The modular design prioritizes code clarity and extensibility, leveraging NumPy, SciPy, and libdlr for numerical operations. Performance-critical kernels, including the polarizability bubble, Dyson equation inversion, and lattice Fourier transforms, are systematically vectorized and combined with the discrete Lehmann representation to achieve practical efficiency within a pure-Python environment. We validate QAssemble on the electronic structure of graphene with local and non-local interactions. Furthermore, benchmarks on a five-orbital extended Hund-Hubbard model demonstrate that this strategy delivers up to a 60x speedup over traditional loop-based Matsubara implementations. QAssemble supports both batch execution for production calculations and interactive workflows for method development. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_22223 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | QAssemble: A Pure Python Package for Quantum Many-Body Theory Mo, Seongjun Li, Dongming Han, Mancheon Jönsson, Johan Kang, Byungkyun Lee, Hoonkyung Kotliar, Gabriel Choi, Sangkook Strongly Correlated Electrons Materials Science QAssemble is a pure-Python package for the quantum many-body problem. It implements various functional approaches, such as tight-binding, Hartree-Fock, and GW approximations within a unified object-oriented architecture. Each physical concept--crystal structure, Hamiltonian, Green's function, self-energy, polarizability, screened Coulomb interaction--is represented as a distinct class. The modular design prioritizes code clarity and extensibility, leveraging NumPy, SciPy, and libdlr for numerical operations. Performance-critical kernels, including the polarizability bubble, Dyson equation inversion, and lattice Fourier transforms, are systematically vectorized and combined with the discrete Lehmann representation to achieve practical efficiency within a pure-Python environment. We validate QAssemble on the electronic structure of graphene with local and non-local interactions. Furthermore, benchmarks on a five-orbital extended Hund-Hubbard model demonstrate that this strategy delivers up to a 60x speedup over traditional loop-based Matsubara implementations. QAssemble supports both batch execution for production calculations and interactive workflows for method development. |
| title | QAssemble: A Pure Python Package for Quantum Many-Body Theory |
| topic | Strongly Correlated Electrons Materials Science |
| url | https://arxiv.org/abs/2604.22223 |