Download
Download the latest version (v3.0 at present) of StochasticGW. New release is coming soon! The new version of StochasticGW will feature:
- Better treatment of periodic systems
- Partially self-consistent DeltaGW0 (updating the HOMO and LUMO and therefore the Green function self consistently, in a scissors operator fashion)
- Norm-Conserving PAW, our new approach for accurate large-grid-spacings simulations
Installation
- Details of installation are provided in the compilation section of manual.
- Sample makefiles and bash shell commands:
| Server | makefile | Bash shell command |
|---|---|---|
| Cori KNL | FCMPI = ftn MPIFLG = FFTFLG = | module load cray-fftw module unload craype-haswell module load craype-mic-knl make |
| Cori Haswell | FCMPI = ftn MPIFLG = FFTFLG = | module load cray-fftw make |
| Comet | FCMPI = mpifort MPIFLG = -DMPI FFTFLG = -lfftw3 | module load fftw make |
Running the code
- Details of StochasticGW can be found in the manual.
- The StochasticGW code works with norm-conserving atomic pseudo-potentials, which can be found at either Abinit or Quantum Espresso
Contact to developers
The contact email for the stochasticGW project is info@stochasticgw.com . Bugs and errors can be reported to report@stochasticgw.com. We cannot promise that your email would be answered, but any issues reported and verified by us or any bug corrected would be detailed in periodic announcements (most likely every two months or so).
Acknowledgement
This work was supported by the Center for Computational Study of Excited-State Phenomena in Energy Materials at the Lawrence Berkeley National Laboratory, which is funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division under Contract No. DE-AC02-05CH11231, as part of the Computational Materials Sciences Program.
The work was further supported by NSF grants DMR-1611382 and CHE-1112500 of Daniel Neuhauser, CHE-1465064 of Eran Rabani, Israel Science Foundation — FIRST Program (Grant No. 1700/14) and Binational Science Foundation, Grant
2015687 of Roi Baer. The code was tested and optimized within XSEDE computational project TG-CHE170058.
The code was mostly self-written, but it has several publicly available routines, primarily the KISS random number generator of George Marsaglia; we use, and slightly modified, the implementation of Jean-Michel Brankart. The package also includes several spline routines from Netlib.