- Code repository migrated to GitHub.
- Code updated for compatibility with Quantum ESPRESSO 7.3. QE must be compiled without CMake.
- Release of TDDFT analytical forces (Jin et al., J. Chem. Theory Comput. 19, 8689 (2023)).
- Improved the performance of TDDFT and BSE on GPU.
- In spin-polarized QDET calculations, the two spin channels can have different bands in the active space.
- Updated CI/CD. Added tests to cover the new functionalities.
- Updated documentation. Updated build instructions for NERSC/Perlmutter. Added more tutorials.
- Code updated for compatibility with Quantum ESPRESSO 7.2. QE must be compiled without CMake.
- Added support to TDDFT: hybrid functionals, and spin-flip.
- Changed default
macropol_calculationfrom N to C. If the system has vacuum, one should setmacropol_calculationto N. - Enabled the use of the adaptively compressed exchange (ACE) method when evaluating the exact exchange in GW, QDET, BSE, and TDDFT calculations on top of hybrid functionals.
- Ported
westppto GPU. - Improved the performance of TDDFT and BSE on GPU.
- Improved the performance of Wannier localization for large systems.
- Reduced memory usage for large-scale QDET runs.
- Bug fix. Fixed QDET for spin-polarized systems.
- Updated CI/CD. Added tests to cover the new functionalities.
- Updated documentation. Updated build instructions for ALCF/Polaris, ALCF/Theta and NERSC/Perlmutter. Added more tutorials.
- Release of the Bethe-Salpeter solver using density matrix perturbation theory (DMPT, D. Rocca et al., J. Chem. Phys. 133, 164109 (2010) and D. Rocca et al., Phys. Rev. B 85, 045116 (2012)).
- Added truncation of lowest occupied bands in BSE.
- Added an option to include fewer PDEPs when computing off-diagonal terms of the self-energy, reducing file system usage in large-scale QDET runs.
- Added support for Wannier localization in BSE.
- Added Wannier localization to
westpp. - Updated CI/CD. Added tests to cover the new functionalities.
- Updated documentation. Updated build instructions for ALCF/Theta and NERSC/Perlmutter. Added build instructions for ALCF/Polaris. Added more tutorials.
- Release of the Bethe-Salpeter solver in finite field (BSE-FF, N. L. Nguyen et al., Phys. Rev. Lett. 122, 237402 (2019)).
- Added spin-polarization to QDET.
- Added calculation of localization factor with integration using a sphere (supports periodic boundary conditions).
- Added support for cuSOLVER 64-bit API.
- Reduced memory usage in
wfreq. - Reduced memory usage in
westppwhen using hybrid functionals. - Added memory estimate for
wfreq. - Bug fix. Fixed a bug and improved the performance of QDET on GPU.
- Updated documentation. Updated build instructions for ANL/LCRC/Bebop, ANL/LCRC/Swing, NERSC/Cori, and NERSC/Perlmutter. Added more tutorials.
- Release of Quantum Defect Embedding Theory (QDET, N. Sheng et al., J. Chem. Theory Comput. 18, 3512 (2022)), enabling the use of WEST to study strongly correlated states of defects in solids.
- Improved the performance of the k-point case of
wfreq. - Bug fix. Fixed the k-point case of
wfreqwhenqp_bandsis used or whenqp_bandrangedoes not start from 1. - Updated CI/CD. Added tests to cover the new functionalities.
- Updated documentation. Added more tutorials.
- Full release of GPU-accelerated full-frequency GW implementation (V. Yu and M. Govoni, J. Chem. Theory Comput. 18, 4690 (2022)), e.g., enabling the use of WEST at OLCF/Summit, NERSC/Perlmutter, ALCF/Polaris. GPU acceleration of
wstatandwfreqis restricted to NVIDIA GPUs. - Introduced infrastructure changes to prepare for QDET release. Enabled calculation with fractional occupation, added the possibility to specify a set of bands, instead of a range in
wfreq, enabled the computation of the off-diagonal matrix elements in G0W0. - Code updated for compatibility with Quantum ESPRESSO 7.1. QE must be compiled without CMake.
- Added the calculation of inverse participation ratio and localization factor in
westpp. - Bug fix. Fixed calculation of dipole matrix elements in
westpp. - Updated CI/CD to use the pytest framework.
- Updated documentation. Updated build instructions for ALCF/Theta, NERSC/Cori, and UChicago/RCC/Midway3. Added build instructions for NERSC/Perlmutter, OLCF/Summit, ANL/LCRC/Bebop, ANL/LCRC/Swing, and NVIDIA DGX A100. Added more tutorials.
- Updated WEST to be compatible with QE 7.0 (before it was compatible with QE 6.1.0). QE must be compiled without CMake.
- Activated pool parallelization in
wstatandwfreqto distribute spin polarization (for systems with nspin=2). - Activated and optimized band group parallelization in
wstatandwfreq(was disabled in QE 6.1.0). - Enabled the calculation of dipole matrix elements in
westpp. - Updated build system. Automatically refresh the
make.dependfiles when runningmake conf. - Added interface to the ELPA eigensolver.
- Added support to read the HDF5 output of pwscf.
- Removed large direct access I/O, resolving issues with the NVIDIA nvfortran compiler.
- Removed obsolete or non-standard Fortran code, reducing the number of warnings during compilation.
- Updated library dependency to Json-Fortran 8.3.0.
- Bug fix. Print an error message when the code fails to read the restart files.
- Bug fix. Fixed the
XwgQrestart mode ofwfreq, i.e., computing Q from previously completed W and G. - Bug fix. Avoid overflow in
IO_kernel/wfreqio.f90in large-scale runs. - Bug fix. Check that the mandatory logicals
nosymandnoinvare set to true for systems with k-points. - To avoid overwriting JSON files when files with the same name already exist, a suffix is appended to the name of the new file.
- Updated CI/CD. Adapted nightly tests to cover OpenMP and ScaLAPACK.
- Updated CI/CD. Added tests of hybrid functionals.
- Updated documentation.
- Introduced new data layout. Parallelization over bands allows to distribute data in a more flexible way in
wstat. This feature also helps reduce memory per image. Band parallelization is enabled by specifying-nb xxxfrom the command line. - Introduced checkpointing in
wstat. With the new keywordn_steps_write_restartone can control how often the code produces restarts (default value is1). - Improved I/O in
wfreq. The number of I/O operations is reduced insolve_wfreq, and in the gamma case ofsolve_gfreq. - Updated library dependency to Json-Fortran 8.2.1, resolving compilation issues with PGI 19.10.
- Updated the initialization of Forpy, such that in case of a module import error, the code has a better chance to print a clear error message.
- Updated build. Now the code builds with the NVIDIA/PGI Fortran compiler.
- Bug fix. Fixed an undefined variable in
Tools/set_npwq.f90. Added `IMPLICIT NONE to all program units to let the compiler catch such errors in the future. - Bug fix. Avoid overflow in
Wstat/wstat_memory_report.f90in large-scale runs. - Bug fix. Reset permissions to all source files in Westpp. All files appeared to be executable before.
- Bug fix. Reset
make.dependfiles, added support tomake -j. - Updated CI/CD. Use updated Docker images.
- Updated CI/CD. Check numerical results and fail the CI if results don't match (see
check.py). DFT checks error in total energy, WSTAT checks maximum error in PDEP eigenvalues, WFREQ checks maximum error in QP energies. - Updated CI/CD. Added tests of images and OpenMP threads to nightly tests.
- Updated documentation. Fixed doc build with sphinx 3.5.0+.
- Updated documentation. Updated build instructions for ALCF/Theta, NERSC/Cori, UChicago/RCC/Midway3, macOS.
- Updated documentation. Added more tutorials.
- Updated documentation. Updated manual.
- Added support for python 3.8 (--embed)
- Updated scripts for RCC-Midway and MacOSX
- Solved bugs in reporting conf layer in Makefile
- Updated documentation
- Introduced automatic installation of missing python packages
- Introduced the conf layer in the Makefile to ease installation
- Updated manual
- Improved usability of client/server mode with
server_control - Reduced execution time of
wfreq(W) for solids - Added build instructions for RCC-Midway and MacOSX
- Updated build instructions for ALCF-Theta
- Updated manual
- Added client/server mode
- Added coupling to Qbox code (http://qboxcode.org)
- Added python3 interface
- Simplified the input format (now accepting both JSON and YAML formats)
- Expanded documentation
- Python suite for pre- and post process WEST calculations
- Introduction of k-points sampling
- Porting to Intel KNL
- Migration of all developments to a private GitLab server, and master branch mirrored to GitHub
- Test suite and continuous integration in place to automatically test the integrity of the code at every addition
- Expansion of documentation, and streamlining of its generation using markup language (Sphinx)
- Restructuring of I/O in JSON (JavaScript Object Notation) format, thus enabling seamless integration with pre-/postprocessing tools and compatibility with Jupyter electronic notebooks
- Implementation of spin-orbit coupling
- Implementation of novel hybrid functionals derived from GW self-energy
- Addition of postprocessing routines, forming the seed for WESTpy
- Efficiency improvements and bug fixes to the contour deformation technique
- Efficiency improvements and bug fixes to the PDEP algorithm
- Initial beta release of WEST: GW without empty states