Gaussian Inc. provides a range of support and resources for users of Gaussian 16 Revision C.01, including:
Faster methods for calculating excited states of larger systems.
: Legacy cluster systems operating on Linda 9.1 or older are fully incompatible with Revision C.01. gaussian 16 revision c.01
| Revision | Release Date | Key Features | |---|---|---| | A.03 | January 2017 | Initial Gaussian 16 release; GPU support introduced | | B.01 | February 2018 | Additional DFT functionals and methods | | C.01 | July 2019 | NBO7 support, RESP charges, expanded GPU support | | C.02 | March 2022 | NVIDIA A100 GPU support; no other functional changes |
Are you planning to install this revision on a or an HPC cluster ? Gaussian Inc
What or theoretical methods (e.g., DFT, ONIOM, excited states) do you primarily work with? Share public link
Perhaps the most prominent addition in Rev. C.01 is full support for Natural Bond Orbital (NBO) version 7, an external program for analyzing molecular electronic structure. The Population keyword gained several new options: Pop=NPA7 for Natural Population Analysis, Pop=NBO7 for full Natural Bond Orbital Analysis, Pop=NBO7Read for custom NBO input, and Pop=NBO7Delete for analyzing interaction deletions using NBO7. These options allow researchers to perform more sophisticated analyses of charge distributions, bonding patterns, and orbital interactions. Importantly, deletion analyses and optimizations with deletions are now compatible with both NBO6 and NBO7, ensuring backward compatibility while offering the enhanced capabilities of NBO7. | Revision | Release Date | Key Features | |---|---|---| | A
Opt Freq : Command route telling the software to optimize the structure, then calculate the vibrational frequencies.
Given that many journals (e.g., J. Chem. Phys. , J. Phys. Chem. A ) require reporting the exact Gaussian revision used, citing is still perfectly acceptable and recommended for work done through 2023–2025.
If your SCF fails to converge, add SCF=XQC to your route section. This switches to a quadratically convergent algorithm if the default EDIIS/CDIIS approach fails.
Optimized allocation routines prevent segmentation faults when running large-scale calculations on nodes with high core densities.