Default Changes And Printing Changes Log for Q-Chem 6.3
This log is intended to serve as a resource for upgrading users who use scripts and parsers to submit jobs and collect data from Q-Chem output files, for example in the case of high-throughput calculation workflows. It provides more details on printing changes and default changes that could potentially affect these users. This log is still experimental, so improvements and suggestions are welcome.
Did we miss something? Do you have suggestions? Contact our support team here!
Q-Chem 6.3.0 Release
May 23, 2025
Changes to keywords:
Change ALPHA in BW-s2 jobs to n/1000 rather than n/100
The ALPHA keyword in the variable section for BW-s2 jobs is now divided by a scaling factor of 1000 rather than 100. This was changed to increase the level of precision available to the user. To restore the original behavior, adjust the ALPHA in the input files for Q-Chem versions prior to 6.3 by a factor of 10.
Changes to default behavior:
Make O-V projection the default for BASIS2 calculations (BASISPROJTYPE = OVPROJECTION)
This was changed to increase robustness in SCF and DFT calculations using basis set projection (BASIS2). The default is now to use O-V projection, rather than Fock projection. To restore the original behavior, set BASISPROJTYPE = FOPPROJECTION.
Use MDQ-uFERF as default for EDA2 jobs (EDA_UFERF = TRUE)
EDA2 now uses the new MDQ-uFERF procedure, rather than DQ-FERF. This was done to improve computational efficiency, as the new MDQ-uFERF offers similar accuracy with 5-10 times less computational cost. To restore the original behavior, set EDA_UFERF = FALSE.
All NEO jobs will now use a tight initial guess by default
All NEO jobs now use the tight initial guess by default (see the Q-Chem manual section "NEO-SCF Specific Keywords" for details on the different guess types). To restore the original behavior, modify NEO_SCF_GUESS_N to the desired value.
Use nuclear core guess by default for NEO jobs with ghost atoms
A nuclear core guess is now used by default for NEO jobs with ghost atoms. To restore the original behavior, modify NEO_SCF_GUESS_N to the desired value.
Use simultaneous SCF by default for NEO calculations
All NEO calculations now use simultaneous SCF by default, which provides significant improvements to performance. In order to restore the original behavior, set NEO_SIMULTANEOUS_SCF = FALSE.
Seminumerical SMD frequency
A new seminumerical SMD is used by default to speed up frequency calculations. The printing in this section may slightly differ from the previous version.
Changes to printing behavior:
RIBW-S2 printing changes
There have been modifications to the printing changes in some RIBW-s2 calculations. The section before the BW-s2 section, which previously was formatted as follows:
Normal F+W unitary transformation algorithm will be used. Initial guess from unregularized MP2 amplitudes. Initial guess (MP2) correlation energy = -1.1067175586 Amplitudes converge when Delta < 1.0e-08
has become formatted in the following way, to provide more information:
Allocated Buffer size: 635882 kB Normal F+W unitary transformation algorithm will be used. Initial MP2 energy: -1.1067175586 Initial guess from MP2 amplitudes (Correlation energy: -1.1067175586). Amplitudes converge when: |Delta| < 1.0000e-08 |Res| < 1.0000e-08
Additionally, a new section will print non-iterative contributions to the correlation energy, just after convergence of the amplitudes:
Non-iterative contributions to the correlation energy:
aaaa contribution = -0.0000000000 a.u.
bbbb contribution = -0.0000000000 a.u.
Parsers that use "contribution =" as one of their parsing keywords should be adjusted accordingly.
Changes in PCM printing
The following line was removed from PCM printing:
Please cite: CITATION
The following line was also added, for PCM calculations involving Hessians:
Compute PCM Hessian
Changes in Boys printing
There has been significant cleanup in the Boys localization print section. timing data line for the Boys localization procedure has been changed from "In BOYS_JACOBI SWEEPS for 10s" to "In Boys Jacobi sweeps for 10s". The following print lines have been added when writing MOs to file:
Write Boys-localized MOs to FILE_MO_COEFS (both spins)
A similar line has been added for writing alpha MOs only:
Write alpha-spin Boys-localized MOs to FILE_MO_COEFS
The following lines have also been added for the different localization methods:
==> Boys localization via Jacobi sweeps ==> Boys localization via conjugate gradient ==> Boys localization via DIIS
A print section that includes information about thresholds and other values:
in get_boys.C: p = [NUMBER] guessthresh [NUMBER] numsofar [NUMBER] MinD [NUMBER]
has been updated to:
Get Boys: p = [NUMBER] guessthresh [NUMBER] numsofar [NUMBER] MinD [NUMBER]
The line that previously read:
Orbital 1 Variance 20.2023450682 Orbital 2 Variance 20.2023450683
Is now printed in a nicely-formatted section:
Orbital variances (bohr^2):
1 20.202
2 20.202
Additionally, upon successful completion, the code previously printed:
Success!!! No Negative Eigenvalues: Ending Get Boys in C++ ...Bye Bye
and now prints:
Success! No Negative Eigenvalues.
There are also several minor adjustments to spacing, and the Davidson solver now has a useful header describing each of the values printed. Other changes may be present. It is strongly suggested that anyone using parsers that rely on the formatting of the Boys localization section run a sample job, review the changes, and adjust their parser accordingly.
Change in NTO printing
The line "excitation amplitudes in the NTO basis (Diabatic)" in NTO calculations that go through the older NTO module has now been changed to "excitation amplitudes in the NTO basis (diabatic)" (lowercase "d").
Additionally, some useful debug printing has been added.
DFT XC integrated density error printing
A new section is printed under the SCF "Total energy" line:
Integrated density by DFT XC: Number of electrons = 10 Numerical integral = 10.000001941423 Relative error = 0.0000194142 %
Dispersion correction printing for NEO
Dispersion correction for NEO is now printed at the beginning the "Simultaneous Nuc./Elec. Block" section, e.g.:
Adding empirical dispersion correction ---------------------------------------------------------------- Doing -D3M(BJ) with s6 = 1.00000 s8 = 1.46668 a1 = 0.27867 a2 = 4.60631 ----------------------------------------------------------------
Simultaneous SCF for NEO calculations
The simultaneous DIIS SCF procedure for NEO calculations (now the default) removes the separate printing of "Nuclear Block" and "Electronic Block" sections within the SCF convergence section, since these are optimized simultaneously, and only prints the "Simultaneous Nuc./Elec. Block" section. Scripts which parse NEO files assuming the presence of these blocks should be modified.