Release log for Q-Chem 5.2




Download Q-Chem 5.2

Release log for prior release, Q-Chem 5.1


5.2.2 Release

Improvements and Bugfixes

  • Value of MEM_STATIC has been capped at MEM_STATIC=8191.
  • Added clarification in the output regarding ECP fits.
  • Tightened threshold for printing normal modes in vibrational analysis to about 5 cm-1.
  • Improved stability of the P-RFO algorithm for geometry optimizations.
  • Added multiple performance improvements for large TD-DFT calculations.
  • Enabled the evaluation of angular momentum properties for EOM-CC transitions (Pavel Pokhilko).
  • Enabled analysis of relaxed and transition TD-DFT densities (Felix Plasser).
  • Enabled high spin multiplicity (nonets and higher) in spin-flip TD-DFT calculations.
  • Enabled automatically generated SAD SCF guess (SCF_GUESS=AUTOSAD) for restricted open-shell SCF.
  • Enabled the use of the new SCF code (gen_scfman) in AIMD calculations.
  • Enabled the use of MOM and IMOM with restricted open-shell SCF.
  • Implemented SPADE orbital partitioning for projection-based embedding theories (Yuezhi Mao).
  • Added capability to visualize frozen and polarized ALMOs (Yuezhi Mao).
  • Resolved an issue that caused CAS-SCF calculations to crash.
  • Resolved an issue with the orthogonalization of the multipole basis in the EDA2 procedure.
  • Resolved an issue with generating noncovalent interaction (NCI) plots for unrestricted SCF.
  • Resolved an issue with invoking EPR and polarizability computations via JOBTYPE=POLARIZABILITY and MAGNET=TRUE.
  • Resolved an issue with incorrectly evaluated MM force in specific cases of custom force fields $force_field_params (Bhaskar Rana).
  • Resolved an issue with generating excited state density plots for TD-DFT calculations.
  • ALMO-based analysis for charge-transfer effects on energetics and molecular properties (Yuezhi Mao, Matthias Loipersberger, and Martin Head-Gordon).
  • Resolved an issue with using automated SAD SCF guess with effective core potentials (Yuezhi Mao).
  • Resolved an issue with plotting open-shell natural transition orbitals with EOM-CC (Felix Plasser).
  • Resolved an issue with handling SMx solvation models with hybrid SCF algorithms such as DIIS/GDM.
  • Resolved an issue that caused not-a-number errors when using Fock extrapolation with AIMD.
  • Resolved an issue that prevented Hirshfeld analysis from being performed in systems with ghost atoms (Yuezhi Mao).
  • Resolved multiple issues with plotting molecular orbitals, attachment/detachment and spin densities (Yuezhi Mao).
  • Resolved an issue with setting up Voronoi complex absorbing potential for molecules with ghost atoms (James Gayvert).


5.2.1 Release

New Features

  • Evaluation of Dyson orbitals for EOM-DIP and CVS-EOM-IP methods (Wojciech Skomorowski, Anna Krylov).
  • Self-consistent state-specific method based on constrained equilibrium manipulation (Fan Wang).


Improvements and Bugfixes

  • Speed up of numerical quadrature in DFT and TD-DFT calculations.
  • Range operator (:) is available for specifying atom lists in constrained optimization input.
  • Removed limits on the number of shells per atom in basis sets (Susi Lehtola).
  • Atomic masses in AIMD can be customized using the $mass input section.
  • Point group symmetry automatically disabled for MBD-rdW calculations.
  • Improved input validation in the $molecule input section.
  • Improved memory footprint of RI-J frequency calculations.
  • Resolved a memory issue in unrestricted TD-DFT frequency calculations.
  • Resolved a numerical instability in some IRC calculations.
  • Resolved an issue that prevented the use of occ-RI-K with range-separated hybrid DFT.
  • Resolved an issue with geometry optimization using the SM12 solvation model.
  • Resolved an issue with projected complex absorbing potential in CAP-EOM.
  • Resolved an issue that caused RI-K calculations to crash in certain cases.
  • Resolved an issue that caused RI-ADC calculations to crash due to insufficient memory.


5.2.0 Release

Changes in Default Settings

  • OpenMP shared-memory parallelism used by default, add -mpi argument for MPI based distributed memory parallelism.
  • Pure basis functions are used by default with BASIS=GEN.
  • Default number of grid points in Lebedev grids in solvent models changed from 302 to 194 points (non-Hydrogen) and 110 points (Hydrogen) atoms.
  • Use of SWIG charges for SMx models.
  • Input format for XPol, SAPT and XSAPT, and MBE jobs has changed.
  • Use EDA2 as the default driver for ALMO-EDA.
  • Frozen core approximation no longer applied by default in RAS-CI calculations.


New Features and Methods

  • General improvements:
    • Increased availability of basis sets: High angular momentum basis functions (up to k-functions) supported for most SCF, RI-MP2, CC/EOM-CC/ADC calculations.
    • Streamlined input format for RI-SCF calculations.
    • Added the def2- family of density fitted (RI) basis sets for SCF and post-SCF calculations (Courtesy of Dr. Florian Weigend).
    • On-the-fly generation for the superposition of atomic densities guess for SCF (Kevin Fenk, John Herbert).
    • Reintroduction of legacy ECPs without fitting.
    • Performance improvements in the computation of partial hessians with DFT.
    • Easy specification of basis sets on fragments, reading of basis sets from an external file (Zheng Pei and Yihan Shao).


  • Improvements in DFT capabilities:
    • Support for analytic frequency calculations using meta-GGA density functionals.
    • Support for analytic frequency calculations using resolution-of-the-identity (density-fitted) Coulomb.
    • Improved performance of partial hessian calculations using DFT.
    • New density functionals: revM06, revM11 (Pierpaolo Morgante and Roberto Peverati).


  • Improvements in implicit solvation models:
    • Revised PCM tessellation grids for improved performance (John Herbert).
    • Change default tessellation method to SWIG in SMD to ensure smooth potential energy surfaces (Yuezhi Mao).
    • Improved performance of the general SCF program with SMx solvation models (Yuezhi Mao).


  • Improvements in the MP2 codes:
    • Regularized kappa-OOMP2 method (Joonho Lee and Martin Head-Gordon).
    • Resolved issue with incorrect computation of the non-Brillouin singles contribution in RO-RIMP2.


  • Improvements in the coupled clusters codes:
    • Suite of mixed-precision CC and EOM-CC capabilities for up to 2x speedup (Pavel Pokhilko, Anna Krylov, Ilya Kaliman).
    • Damped response, dynamic polarizabilities for two-electron absorption using EOM-CC (Kaushik Nanda and Anna Krylov).
    • Improved evaluation of spin-orbit coupling constants across EOM-CC states (Pavel Pokhilko and Anna Krylov).
    • Better handling of linear point groups in ADC and CC methods.
    • Improved performance of disk-based ADC/CC algorithms.


  • Improvements in energy decomposition analysis methods:
    • Switch to EDA2 as the default driver for ALMO-EDA calculations.
    • Added electron density difference (EDD) plots and the ETS-NOCV analysis (Yuezhi Mao).
    • Added support for PCM and SMD solvation models in ALMO-EDA (Yuezhi Mao).
    • Resolved several issues that caused instabilities in MP2-EDA calculations (Yuezhi Mao).


  • New methods and capabilities:
    • Polarizable embedding model (Maximilian Scheurer).
    • Incremental FCI method (Paul Zimmerman).
    • Transition potential DFT for core-valence excitations.
    • Analytic evaluation of Raman intensities (Zheng Pei and Yihan Shao).
    • Pilot implementation of NEO-RHF method (Fabijan Pavosevic and Sharon Hammes-Schiffer).


  • New features of the BrianQC GPU module:
    • GPU accelerated 1-electron integrals.
    • GPU accelerated DFT exchange-correlation with support for LDA, GGA, and meta-GGA functionals.
    • Support for new GPUs using NVIDIA's Turing architecture.
    • Optimized memory consumption.


  • Other fixes and improvements:
    • Resolved issue with evaluating and printing Dyson orbitals in RI/CD-CC calculations.
    • Resolved issue with plotting EOM attachment/detachment densities.
    • Resolved issues with reading user input in open-shell EOM-EE calculations.
    • Resolved issue with estimating optimal subspace size in SF-TD-DFT calculations.
    • Resolved issue that prevented dual-basis DFT calculations from utilizing shared-memory parallelism.
    • Resolved an issue with geometry optimizations using RO-SCF that caused non-convergence of SCF in subsequent geometry optimization steps (Yuezhi Mao).


Other Release Logs: