Gaussian 16 Revision C.01 ⭐ Newest

1. What is Gaussian 16 Rev C.01?

• Major release: Gaussian 16 (G16)
• Revision: C.01
• Significance: One of the later stable revisions of G16. It includes bug fixes and performance improvements over Rev A.xx and B.01.
• Key improvements over earlier revisions:
  • Better parallel scaling (shared-memory and distributed-memory).
  • Updated DFT functionals and dispersion corrections (e.g., DFT-D3 with Becke-Johnson damping).
  • Enhanced ECP/ basis set handling.
  • Support for new hardware and compilers.

Gaussian 16 Revision C.01: A Comprehensive Guide to Features, Performance, and Practical Applications

7. Transitioning from Rev B.01 to Rev C.01: A User’s Checklist

If your group runs hundreds of jobs weekly, migrating from an older revision requires validation. Follow this checklist:

• [ ] Test reproducibility – Run a small set of molecules (e.g., thermochemistry of alkanes) and compare total energies. Differences should be < 1e-6 hartree unless a bug is fixed.
• [ ] Update PBS/Slurm scripts – Use %NProcShared instead of %NProc for pure OpenMP jobs; Rev C.01 handles hybrid parallel differently.
• [ ] Check dispersion corrections – Rev C.01 defaults to Grimme’s D3 with zero damping; earlier revisions may have used D2.
• [ ] Revalidate excited state results – TD-DFT oscillator strengths are corrected; re-calculate any published spectra.
• [ ] Set up HDF5 scratch – For large jobs, add %UseHDF5=yes to avoid file system bottlenecks.

Many HPC centers (e.g., NERSC, ARCHER2, CSCS) have switched default Gaussian symlink to Rev C.01 because of its robustness. gaussian 16 revision c.01

Best practices when using Gaussian 16 Rev C.01

• Always validate key results with multiple methods/basis sets: run single-point energies with a higher-level method or larger basis set, and verify critical barriers or spectroscopic assignments.
• Monitor SCF and geometry convergence: enable tight convergence for energetics-critical runs and consider DIIS/SCF convergence aids or different initial guesses if problems occur.
• Use appropriate basis sets and ECPs for heavy elements: match ECPs and basis compatibly; check relativistic needs for heavy‑atom chemistry.
• Exploit frozen-core and local approximations judiciously for large systems to balance cost and accuracy.
• Keep work directories and checkpoint files: checkpoint (.chk) files allow restart, orbital analysis, and smoother continuation between jobs.
• Reproduce and document input: preserve input decks and software revision in publications; cite Gaussian properly and note the revision.

5.4. Common Installation Issues

• "Cannot open checkpoint file": Ensure $GAUSS_SCRDIR is writable and not in NFS (use local SSD).
• Linda workers not spawning: Set %LindaWorkers correctly and ensure passwordless SSH between nodes.
• GPU not detected: Run g16 -gpu-info; if empty, set CUDA_VISIBLE_DEVICES.

2.3. Solvation and Molecular Mechanics

• SMD solvation model: Corrected atomic radii for halogens, leading to more accurate hydration free energies.
• ONIOM with XTB: Experimental support for xTB as a low-level method (requires separate installation, but Rev C.01 provides the interface).

2.1. Density Functional Theory (DFT)

• New functionals added: Rev C.01 includes up-to-date Minnesota functionals (M06-2X, M06-L) with corrected parameters for thermochemistry. Also, the ωB97X-V range-separated functional (with VV10 nonlocal correlation) is fully supported.
• Improved SCF convergence: New default convergence criteria for DIIS (Direct Inversion in the Iterative Subspace) reduce oscillations in highly correlated systems like transition metal complexes.
• Self-consistent GW (scGW) : A beta implementation of single-shot GW (G0W0) and self-consistent GW for band gaps in molecular systems—unique to Rev C.01.