TURBOMOLE Release Notes
TURBOMOLE V7.2.1 (November 2017)
Turbomole version 7.2 shows (on some CPU types with AVX2 and newer) unexpected behavior when running the SMP parallel version (job seems to run, but does not proceed). It also refuses to run on older Linux distributions and stops with a GLIBC error.
In addition, we have been asked about an own version for the new AMD CPU types (Ryzen and EPYC). And we had a couple of user requests, see list below.
We thus release a minor update, version 7.2.1, which includes a couple of changes:
- parallel SMP version runs stable also on new CPU types
- new optimized binaries for the latest AMD CPU family
- minimum requirement for GLIBC version reduced to GLIBC_2.4 and should now run on older Linux versions too
- TTEST script failed due to security changes of Perl on newest Debian (and alike) distributions, fixed
- new X2C basis set family included 
- XCFun functionals (M06-L, ...) enabled in MPI version
- MPI server task can now be sent to sleep during the calculation, reduces the CPU usage of this task distributing process.
export TM_SERVERSLEEP=on (see documentation)
- NumForce uses all CPUs from the $PARNODES variable if PARA_ARCH=SMP is set
- fix problem with RECL error when using aoforce
- fix of point charge gradients for MPI version (7.2 gives NaN in some cases)
- improved TDDFT restart
- EDA keyword recognition modified
- changes which lead to a more stable behavior if small molecules are calculated with a large number of CPUs
- visualization of periodic geometry optimizations in 1D and 2D corrected
- improved import of old TmoleX projects
- spectra plot, corrected plot range
- assignment of memory for TmoleX itself under Windows corrected
 Patrik Pollak and Florian Weigend
Segmented Contracted Error-Consistent Basis Sets of Double- and Triple-ζ Valence Quality for One- and Two-Component
Relativistic All-Electron Calculations
Journal of Chemical Theory and Computation 2017 13 ( 8 ), 3696-3705
TURBOMOLE V7.2 (July 2017)
- VCD spectra (Hartree-Fock and DFT) 
- Vibrational frequency calculations with COSMO
- Periodic boundary conditions: 
- optimization of unit cells
- density of state plots
- plots of crystal orbitals
- improved OpenMP parallelization
- in core storage of RI integrals
- Bethe-Salpeter excitation energies 
- two-photon transition moments at CC2 level 
- local hybrid functionals 
- phosphorescense lifetimes using (SOC-PT-)CC2 
- AIM critical points (attractors, bonds, rings, others)
- Fukui functions
- PBE and PBE0 functionals for NMR chemical shieldings
- new functionals: B97-3c and SCAN 
- Nonadiabatic molecular dynamics using spin-unrestricted TDDFT 
- new and more stable TDDFT solver 
- faster and more stable RI-RPA
- GlobalArray library removed, new shared-memory implementation for MPI jobs (ridft/rdgrad)
- new OpenMP/MPI hybrid parallelization of dscf, grad, aoforce, ricc2, pnoccsd
- internal full 64bit data model as default (former 'huge' version)
- reduced output of parallel versions (no slave.* files any more)
- usage of internal 64bit data models lead to virtually no limitations for size of the molecules or memory usage
- TmoleX 4.3:
- AIM (atoms in molecules) visualization of critical points
- Project/Job management: change order of jobs, move jobs from on project to another project
- better support for periodic boundary condition calculations (cif file import, optimization of the unit cell, weight derivatives,...)
- builder with stereochemical description (R/S) and modifications
- density of states
- new job template editor
- SMILES support (also in batch jobs)
- easier and interactive generation of population properties (Mulliken, NBO, Loewdin, Wiberg, Paboon)
- improved combinatorial library tool
 K. Reiter, M. Kuehn, and F. Weigend,
Vibrational circular dichroism spectra for large molecules and molecules with heavy elements
The Journal of Chemical Physics 146, 054102 (2017)
 R. Lazarski, A. M. Burow, L. Grajciar, M. Sierka,
Density Functional Theory for Molecular and Periodic Systems Using Density Fitting and Continuous Fast Multipole Method: Analytical Gradients
J. Comput. Chem. 2016, 37, 2518-2526
 K. Krause, W. Klopper
Implementation of the Bethe-Salpeter equation in the TURBOMOLE program
J. Comput. Chem. 2017, 38, 383-388
 D. H. Friese, C. Haettig, K. Ruud
Calculation of two-photon absorption strengths with the approximate coupled cluster singles and doubles model CC2 using the resolution-of-identity approximation
Phys. Chem. Chem. Phys. 2012, 14, 1175-1184
 H. Bahmann, M. Kaupp
Efficient self-consistent implementation of local hybrid functionals
J. Chem. Theory Comput., 11, 1540-1548, (2015)
S. Klawohn, H. Bahmann, M. Kaupp
Implementation of molecular gradients for local hybrid density functionals using seminumerical integration techniques
J. Chem. Theory Comput. 12, 4254-4262, (2016)
T. M. Maier, H. Bahmann, M. Kaupp
Efficient semi-numerical implementation of global and local hybrid functionals for time-dependent density functional theory.
J. Chem. Theory Comput. 11, 4226-4237, (2015)
 B. Helmich-Paris, C. Haettig, C. van Wuellen
Spin-Free CC2 Implementation of Induced Transitions between Singlet Ground and Triplet Excited States
J Chem Theory Comput. 2016, 12(4), 1892-904
 SCAN: https://arxiv.org/ftp/arxiv/papers/1511/1511.01089.pdf
B97-3c: S. Grimme et al, in preparation
 J. C. Vincent, M. Muuronen, K. C. Pearce, L. N. Mohanam, E. Tapavicza, F. Furche
That Little Extra Kick: Nonadiabatic Effects in Acetaldehyde Photodissociation
J. Phys. Chem. Lett., 7, (2016), 4185 - 4190
 F. Furche, B. Krull, B. D. Nguyen, J. Kwon
Accelerating molecular property calculations with nonorthonormal Krylov space methods
J. Chem. Phys., 144, (2016), 174105
TURBOMOLE V7.1 (July 2016)
- DFT calculations for periodic systems (riper) 
- Analytic gradients now allow geometry optimizations
- ECPs are available for energy and gradient calculations
- fractional occupation and smearing for metals
- band structure plots and visualization of the density
- DFT-D3 dispersion correction for energies and gradients
- RI-RPA gradients for unrestricted calculations (rirpa) 
- genetic algorithm for global structure optimization (see DoDo section in the manual) 
- COSMO isorad analytic gradients (rdgrad/grad)
- faster Davidson algorithm for TDDFT 
Nonorthonormal Krylov space methods in combination with RI methods yielding 2-5 fold speedups in (TD)HF and hybrid (TD)DFT response calculations.
- OpenMP/MPI parallel version of 2nd analytic derivatives (aoforce)
- pob-TZVP basis set added for periodic boundary condition RI-DFT. 
- New scripts and tools:
- MD postprocessing tools for internal coordinates and rotational constants (log2int, log2rc)
- tm2ezspec - create input for ezspectrum (Franck-Condon factors, Krylov group: iopenshell.usc.edu/downloads/
- panama - generate input for visualization of unrelaxed difference
densities from escf output
- IBO (intrinsic bond orbital) analysis (proper)
- TmoleX 4.2:
- support of periodic boundary condition RI-DFT calculations
- visualization of band structures
- easier generation of user-definded building blocks in molecular builder
- generation of combinatorial libraries by automatic permutation of substitutives
 For details of the periodic boundary condition implementation, see
and the list of publications on those sites
 Analytical First Order Molecular Properties and Forces Within The Adiabatic Connection Random Phase Approximation A. M. Burow, J. E. Bates, F. Furche, H. Eshuis J. Chem. Theory Comput., 2014, 10 (1), 180-194 http://pubs.acs.org/doi/abs/10.1021/ct4008553
 Synergy between theory and experiment in structure resolution of low-dimensional oxides. M. Sierka Prog. Surf. Sci. 2010, 85, 398-434 http://dx.doi.org/10.1016/j.progsurf.2010.07.004
 Accelerating molecular property calculations with nonorthonormal Krylov space methods, F. Furche, B. T. Krull, B. D. Nguyen, J. Kwon J. Chem. Phys. 2016, 144, 174105 http://dx.doi.org/10.1063/1.4947245
 Consistent Gaussian basis sets of triple-zeta valence with polarization quality for solid-state calculations. M.F. Peintinger, D. V. Oliveira, T. Bredow J Comput Chem. 2013, 34(6), 451-9 http://dx.doi.org/10.1002/jcc.23153
TURBOMOLE V7.0 (June 2015)
- DFT calculations for periodic systems (new module: riper)
riper is a module for the calculation of ground state energies
for molecular and periodic systems at the RI-DFT level 
- Low-scaling MP2 and MP2-F12 (new module: pnoccsd)
The pnoccsd program contains pair natural orbital-based
correlation methods (hybrid OSV-PNO approximation) with low-cost
scaling for large systems. (currently restricted to MP2) [2,3]
- Two-component of CCS, ADC(2), CIS(inf), and CC2 excitation energies
and excited state transtition moments for ADC(2) and CC2 (ricc2) 
- Self-defined and several pre-defined density functionals (ridft, rdgrad, dscf, grad, aoforce, escf)
now available through the XCFun library version 1.99 (https://repo.ctcc.no/projects/xcfun/wiki)
- RI-RPA is now parallelized with OpenMP (rirpa)
- No program limitation for the number of atoms and the number of basis functions
- New scripts and tools:
- Interactive property program (new module: proper)
proper is an interactive program for first-order properties
(expectation values) from one-electron densities, population
analysis, and generation of interface files for visualization
- Energy decomposition analysis (EDA) (ridft and new module promowa)
- Grimme PBEh-3c functional, including DFT-D3 and gCP 
- Interactive property program (new module: proper)
- TmoleX 4.1:
- Coupling to the 2D structure editor JChempaint
- 2-c ECPs
- excited state potential energy scans
- re-engineered the Orbital/Density Plot
- remote machine administration simplified and extended
- Automatic clean-up of memory by the SMP/GA version of ridft/rdgrad
 Roman Łazarski, Asbjörn M. Burow, and Marek Sierka, Density Functional Theory for Molecular and Periodic Systems Using Density Fitting and Continuous Fast Multipole Methods, Journal of Chemical Theory and Computation, Article ASAP, DOI: 10.1021/acs.jctc.5b00252 (http://pubs.acs.org/doi/abs/10.1021/acs.jctc.5b00252)
 G. Schmitz, B. Helmich, C. Haettig. A O(N 3) scaling PNO-MP2 method using a hybrid OSV-PNO approach with an iterative direct generation of OSVs. Mol. Phys. 111, 2463-2476 (2013)
 G. Schmitz, C. Haettig, D.P. Tew. Explicitly correlated PNO-MP2 and PNO-CCSD and its application to the S66 set and large molecular systems. Phys. Chem. Chem. Phys. 16, 22167-33178 (2014)
 K. Krause and W. Klopper, Description of spin-orbit coupling in excited states with two-component methods based on approximate coupled-cluster theory, J. Chem. Phys., 142 (2015), 104109
 S. Grimme, J. G. Brandenburg, C. Bannwarth, A. Hansen JCP, submitted
TURBOMOLE V6.6 (June 2014)
- Reaction path optimization (new module: woelfling) The optimization starts with a linear synchronous transit (LST) followed by a chain-of-states method that optimizes reaction paths under the sole constraint of equally spaced structures. In contrast to Nudged Elastic Band or Growing String Method, it requires no spring forces, interpolation algorithms,or other heuristics to control structure distribution. 
- First order electron vibration coupling (new module: evib)
- Many body perturbation theory in the GW approximation (escf)
- Solvation effects on excited states ADC(2) with COSMO (ricc2)
- The M06 and M06-2X functionals (ridft, rdgrad, dscf, grad, aoforce, escf)
The XCFun library is now used for easier interchange of new functioals.
- CCSD(T) energy with interference-corrected MP2-F12 (ricc2)
- RI-RPA gradients available (rirpa)
- Two-component RI-RPA energies (rirpa)
- Two-component TD-DFT for spin-orbit effects (escf)
- VV10 like density based dispersion correction (ridft, rdgrad) 
- MP2/COSMO is included in ricc2. The rimp2 module is removed since all functionalities are now available in ricc2.
- New def2 basis sets for the Lanthanides 
- New scripts and tools:
- scanprep. Prepares a scan along frozen internal coordinates
- log2rog. Calculate the radius of gyration from MD log files
- past. Turns coord in principal axis system and prints rotational constants
- NumGrad, NumHess. Computes numerical gradient and second derivatives
- TmoleX 4.0:
- A new look and feel and some code reorganisation
- The builder is improved to more chemical intuition
- New option in Transition State Search: Reaction Path Sampling
- First version of an online update possibility
- The SMP/GA version of ridft/rdgrad can now be run with more one job on one machine
- Problem of define for small molecules in symmetry creating start orbitals fixed.
 P. Plessow. Reaction Path Optimization without NEB Springs or Interpolation Algorithms. J. Chem. Theory Comput., 9(3), 1305-1310, (2013).
 W. Hujo, S. Grimme. Performance of the van der Waals Density Functional VV10 and (hybrid)GGA Variants for Thermochemistry and Noncovalent Interactions. J. Chem. Theory Comput. 7(12), 3866, (2011).
 R. Gulde, P. Pollak, and F. Weigend. Error-Balanced Segmented Contracted Basis Sets of Double- to Quadruple-Zeta Valence Quality for the Lanthanides. J. Chem. Theory Comput., 8(11), 4062-4068 (2012).
TURBOMOLE V6.5 (May 2013)
- CCSD vertical excitation energies
- MP2 & CC2 polarizabilities 
- MP3(F12) and MP4(F12) energies
- new odft module, for all orbital-dependent Kohn-Sham DFT methods. Includes LHF and a new implementation of the exchange-only optimized effective potential (OEP-EXX) 
- Linux version: overall faster code on newer Intel or AMD CPUs
- parallel SMP version for RI-K gradient
- enhanced parallel SMP version of ridft and rdgrad
- new SMP parallel versions of dscf, grad, ridft and rdgrad available as alternative to default parallelization
- again, IBM Platform MPI (8.2) is included in the Turbomole distribution, for details about Platform MPI see:
http://www-03.ibm.com/systems/technicalcomputing/platformcomputing/products/mpi Turbomole users can run parallel calculations 'out-of-the-box' without the need to install MPI itself. Uses shared memory on one node and supports Infiniband or most other high-speed interconnects between nodes.
- new scripts MECPprep and MECPopt for preparing and performing minimum-energy crossing point optimizations.
- new auxiliary basis sets for RI-MP2 and RI-CC (cbasen) for
from H-Rn (without Lanthanides).
- TmoleX 3.4
- new feature: batch jobs (graphical version of the 'calculate' script). It is now possible to define a job template with one or several different jobs (basis set, method, charge, kind of job, etc.) and apply it to a LIST of molecules. Runs either on the local system or on a remote machine.
- new feature: simple build-in queuing system for local and remote jobs. If you tell TmoleX how many CPUs you want to use on your local or on remote systems, it is possible to start many different jobs at a time.
TmoleX will only run a limited number of jobs, automatically starting jobs as soon as a CPU is free for computation.
- new features: TDA (Tamm-Dancoff) is now supported, ESP fits and EPN (electrostatic potential at nuclei) added.
- export of MOs to WFN, molden, and AOMix enabled.
easier plotting of electrostatic potential on density
check work load on remote system from within TmoleX
Note: parts of the new SMP parallel features are only available under Linux 64bit.
 D. H. Friese, N. O. C. Winter, P. Balzerowski, Raffael S., and C. Haettig, J. Chem. Phys. 136, 174106 (2012)
 A. Hesselmann, A. W. Goetz F. Della Sala, and Andreas Goerling, J. Chem. Phys. 127, 054102 (2007)
TURBOMOLE V6.4 (April 2012)
- non-adiabatic TDDFT surface hopping
- spin-flip TDDFT
- semi-numerical HF exchange for one- and two-component calculations
- gradients for 2c-DFT energies
- FDE, frozen density embedding (dscf module only)
- O(N^4) implementation of SOS-CC2 excitation energies
- parallel SMP version for CCSD(F12)
- parallel MPI versions of dscf and grad can skip time demanding pre-job step which determines the task distribution for large molecules by setting environment variable $SKIP_PARASTAT=yes NOTE: use with care, it assumes that you are not using more CPUs than symmetry non-redundant shells.
- max. number of CPUs for parallel MPI jobs increased to > 2000
- new option 'nocheck' for point charges skips check for overlapping point charges and correct symmetry. Significantly faster if many point charges are given.
- new version of Platform MPI (8.2) is included in the Turbomole distribution, for details see: features-benefits
- new script 'DRC' for performing dynamic/intrinsic reaction coordiante calculations
- interface to the hotfcht program from R. Berger hotFCHT from the Berger group
- full 64bit version with 'huge' parameter set (up to 1400 atoms) for Linux systems are available upon request. Those binaries will use more memory and run a bit slower, but should be able to use a more or less unlimited amount of memory and disk space.
- TmoleX 3.3
- molecular builder with new simple draw-tool
- job templates can handle user-defined multi-jobs, for example
- geometry optimization with subsequent vibrational frequencies calculation
- optimize with small basis set, then with larger basis set, then run single-point MP2 or CCSD(T) energy calculation
- run jobs using different density functionals at a time
- job templates available for COSMO files as needed for the COSMO-RS TZVPD-FINE level
- combine results from different jobs, including export to Excel file and (for simple organic molecules) 2D graphics
- New viewer for spectra, including broadening of lines using Lorentzian or Gaussian
- finite temperature contributions from vibrational frequencies
- POV-Ray export of molecular orbitals
- support of DFT-D3 with BJ damping
- diffuse basis functions def2-SVPD, def2-TZVPD, def2-QZVPD supported
- new FINE cavity for COSMO single-point energies can be chosen
TURBOMOLE V6.3.1 (June 2011)
New features and bug fixes:
- DFT-D3 in version 6.3 had problems when using ECPs, in such cases the original code of the Grimme group has not been invoked correctly and wrong dispersion parameters have been used. With default Turbomole/Karlsruhe basis sets this did occur for all elements beyond Kr
- RI-CC2 excited state gradient problems in some rare cases where the program stopped in 6.3
- COSMO + dscf in parallel jobs gave I/O problems in some cases at the end of the job
- TDDFT + COSMO need more balanced thresholds for integral screening, for very small symmetric molecules and many excited states TDDFT did not converge in 6.3
- ScaLAPACK turned out to be very slow on clusters with a slow network, here the Turbomole-own parallel linear algebra routines are faster. Now users can add $parallel_platform cluster to the control file to switch off the usage of ScaLAPACK
- TmoleX 3.2
- + includes Turbomole 6.3.1
- + new possibility to control the assignment of ECPs to the basis sets
- + problems with potential energy scans running on remote systems removed
- + various minor changes
TURBOMOLE V6.3 (March 2011)
- parallel SMP version for
- + 2nd analytic derivatives (module aoforce)
- + TDDFT excited state energies and gradients (modules escf and egrad)
- + CCSD and CCSD(T) (module ricc2)
- Symmetry in CCSD and CCSD(T) for D2h and its subgroups
- vibrational frequency calculations are now restartable (module aoforce)
- TDDFT vertical excitation energies with full COSMO solvation treatment (module escf)
- CCSD(F12*) as a more cost-efficient alternative to CCSD(F12)
- two-component MP2-F12 energy calculations for spin-orbit coupling (module ricc2)
- Property-optimized Gaussian basis sets addition of diffuse functions to the Karlsruhe basis sets: def2-SVPD, def2-TZVPD, def2-QZVPD
- New segmented contracted basis sets for one- and two-component Dirac-Fock effective core potentials
- support for DFT-D3 dispersion correction (original code from Grimme group), including 2nd derivatives with module aoforce (new keyword $disp3)
- Platform MPI 7.1 is included in the Turbomole distribution, for details see: features-benefits
- new parallel environment setting for SMP/multi-core systems
- new script 'evalgrad' which monitors the values of bond length, bond angle or dihedral angle of all steps of a geometry optimization or an MD run
- new toolkit 'thermocalc' to calculate batch-wise atomization energies and heats of formation with customizable computational protocols
- NumForce transfers resulting Hessian and vibrational modes to the original control file for a subsequent transition state search
- TmoleX 3.1
- + improved molecular builder
- + more features of Turbomole supported
TURBOMOLE V6.2 (June 2010)
- initial implementation of CCSD, CCSD(F12) variants, and CCSD(T) (module ricc2)
- LHF is now parallelized with OpenMP (module dscf)
- TmoleX 3.0:
- + the molecule builder has been re-engineered
- + MOPAC7 is included as a pre-optimizer
- + the job type selection has been laid out results oriented
- + a visualization of spectra has been implemented
- + new way of how to define and run PES scan jobs
- + more features of TURBOMOLE can be used now
- LHF supports up to h-functions and heavy elements with ECP
- '$pointval xc' option, to plot exchange-correlation potential on an user define grid
- new MoleControl version 2.1 with enhanced features and an easier installation procedure
- the sign of electronic rotatory dispersion has been corrected (module escf/egrad)
- various bugfixes for the the gobal array (GA) version (module ridft_ga/rdgad_ga)
TURBOMOLE V6.1 (October 2009)
- N^4 spin scaling SOS-RI-MP2 (module ricc2)
- faster MP2-F12 (module ricc2)
- (One-electron) transition moments between excited states in CC2
- Douglas-Kroll-Hess Energies in C1
- OpenMP versions of ricc2 and dscf
- new parallel version of ridft and rdgrad, especially for SMP systems (BETA version - passes all tests, but needs to be watched in production runs !) including:
- parallel linear scaling exact HF-exchange for DFT hybrid functionals (energy and gradients)
- parallel RI-K routines for energy calculations, faster pre-step for parallel RI-MP2 and RI-CC2 calculations
- no master process needed any more
and what to do if an error occurs: Go to faq
- up to 30% faster binaries for current AMD CPUs
- better speed up of new parallel versions compared to the standard MPI version, especially on SMP systems (ridft, rdgrad, and ricc2)
- standard MPI version with reduced CPU usage for the master process
- TmoleX 2.2 with several bug fixes for running Turbomole jobs on remote systems
- DFT-D aoforce calculations (6.0 gave wrong frequencies in some cases)
- external E-Field crash on AMD CPUs fixed (ridft)
- wfn output files containing f functions fixed with $wfn option
- initial start velocity of molecular dynamics set ups now oriented to what the user has entered as temperature (mdprep)
- fixed crashes on some CPU types of Intel and AMD when using highly optimized linear algebra routines by using the latest Intel MKL library (Intel: spin-orbit calculations, AMD: aoforce)
- NumForce and TmoleX fixes for parallel runs or remote jobs when having tcsh as default shell on target machines
- Ir basis set in Turbomole 6.0 had a missing s function in def-TZVP
- new keyword ($lastdiag) increases the accuracy of the output of orbital energies for very small HOMO-LUMO gaps - helpful for TDDFT calculations which complain about the violation of the Aufbau principle
TURBOMOLE V6.0 (January 2009)
- SCS-CC2 for ground and excited states:
Energies, excitation energies, gradients, transitions moments, and first order properties (module ricc2).
- Support of meta-GGA functionals (TPSS and TPSSh) in second analytical derivative calculations (module aoforce).
- MP2-F12 (module ricc2, preliminary version. H-Ar only, except Na, Mg).
- Gradients for point charges in ricc2 (for QM/MM).
- MP2-COSMO gradients (module rimp2).
- MoleControl: A workflow management script kindly provided by BASF SE (see usability).
- Usage of difference density for DFT quadrature (see efficiency, ridft).
- Difference density speeds up DFT ground state energy calculations with ridft (serial version only).
- Change of the default optimizer from relax to statpt:
More stable performance in critical cases, handles fixed coordinates more flexible.
- Improved efficiency in second analytical derivative calculations.
- Parallel gradients and response properties on ADC(2) level fully working.
- MoleControl runs and controls Turbomole jobs using Python scripts.
Many user have written their own scripts and programs to organize their calculations.
The idea of this project is to provide a standardized interface for such workflow developments.
It is intended that the scripts are expanded by the users and exchanged in the community.
Input generation with define is done by MoleControl, users have to provide coordinates and the type of job only.
- Analysis of 2-component wavefunctions.
- The vibration script produces a gnupolt input file, which displays IR spectra (vibration -spectrum).
- Extended export of orbitals to the AOMix program (script t2aomix).
- COSMO calculations with a scaled isodensity cavity for post-SCF thermodynamic properties with the COSMOtherm program.
- Simplified setup of COSMO calculations (just add $cosmo to the control file), making cosmoprep obsolete.
- Simplified properties and wavefunctions analysis (-proper flag).
- Added more basis sets to the library like:
d-aug-cc-pVXZ, 6-311G-type, 6-31G-type
- Default settings in parallel version changed:
$numprocs and $parallel_platform keywords obsolete,
fully direct task distribution now reduces I/O in dscf,
improved interoperability of serial and parallel version
- aoforce with hybrid functionals and RI-J
- spin flip
- parallel PEECM
- 2-component DFT with RI-JK
- parallel NumForce
TURBOMOLE V5.10 (January 2008)
- Vibrational Raman intensities
using analytical derivatives of frequency-dependent TDHF and (hybrid) TDDFT polarizabilities
- Spin-Orbit Coupling
Two-component RI-Hartree-Fock and RI-DFT calculations with Spin-Orbit-ECPs (spin-orbit coupling)
- SCS-RI-MP2 (module ricc2)
energy and gradients (for closed-shell HF and UHF reference states, sequential and parallel implementation)
- Periodic Point Charges
for HF and DFT, energy and gradient calculations
DFT with dispersion, method of S. Grimme (reference: see manual)
- COSMO for RI-MP2 calculations (using rimp2 module)
- Properties: Merz-Kollman ESP fit
- Spin flipping for broken symmetry treatment (module define)
- RI-MP2/RI-CC2 (module ricc2)
gradients/properties for RI-HF-based RI-MP2 and RI-CC2 calculations (i.e. RI-JK approximation for Z-vector equation and gradients)
- NumForce handles frozen cartesian coordinates (option -frznuclei)
- NumForce and jobex:
possibility to use RI-JK together with RI-MP2/RI-CC2 (option -rijk)
- molecular orbital files can be kept in binary format
- tm2molden supports g-functions
- cube format for 3D grid files
- convert vibrational frequency output to G98 format
TURBOMOLE V5.9.1 (April 2007)
- Linux/PC and Itanium2/Linux parallel binares are now using HP-MPI.
- HP-MPI comes with Turbomole, there is no need to install it
- no license key is needed for HP MPI, Turbomole will run 'out of the box'
- HP-MPI is able to run on several different interconnects like
TCP/IP, SMP, Infiniband, Myrinet, Quadrix, ...
- statpt is able to fix Cartesian coordinates
- more stable internal coordinates
- Bug with reducible E representations has been fixed
- RI-R12-MP2 bug fixed
- EM64T binaries for large input files are more stable now
- ricc2 is able to plot properties on grids like potential, electric field, el. field gradients, etc.
- NumForce bug when having fixed Cartesian coordinates fixed
- define bugs (basis set change, delete atom with ECPs, ...) fixed
- jobbsse bug with ECPs fixed
- tmole bug when scanning torsions fixed
- several minor bugs fixed
New binary version:
- Windows version of Turbomole available now - binaries are delivered together with the new
graphical user interface TmoleX.
TURBOMOLE V5.9 (December 2006)
New features include:
- RI-J for Hartree-Fock and DFT hybrid functionals: Efficiency!
- RI-JK gradients implemented and DFT hybrid functional calculations enabled
- aoforce greatly enhanced and extended - module for analytical harmonic vibrational frequencies available for basis sets up to g functions, ECPs up to g projectors enabled
- RI-CC2/RI-MP2: parallel version for ground and excited state gradients and gradients for excited state ADC(2) and CIS(D_infinity) implemented
- MD: Finite temperature canonical Born-Oppenheimer molecular dynamics using Nosé-Hoover thermostat
- NPA - Weinhold's Natural Population Analysis
- New, improved, and automated optimization method for minima and transition structures, new GDIIS method, improved overall stability and generation of internal redundant coordinates
- RI-MP2-R12: Single-point explicitly-correlated RI-MP2-R12 energies for closed- and open-shell systems.
- BSSE: Geometry optimizations can be carried out within the framework of the full function-counterpoise method to avoid basis-set superposition errors
- RI-J in combination with Hartee-Fock exchange speeds up large Hartree-Fock or DFT hybrid-functional calculations
- MARI-J (Multipole Accelerated RI-J): Increased accuracy with reduced overall computational cost
- New parallel diagonalizer with better speed-up for parallel calculations
- aoforce: increased efficiency
Basis sets and auxiliary basis sets:
- Basis sets for Lanthanides and Actinides up to Lr
- RI-J auxiliary basis sets (jbasen) for def2- basis sets H-Rn and def- basis sets for Lanthanides and Actinides
- RI-JK auxiliary basis sets (jkbasen) for all triple and quadruple zeta valence basis sets, including the Lanthanides and Actinides
- New auxiliary basis sets for RI-MP2 and RI-CC2 (cbasen) for:
- def2-QZVPP for Rb-Rn (without Lanthanides)
- def2-SVP, def2-TZVP, def2-TZVPP for H-Rn (without Lanthanides)
- (aug-)cc-pVXZ-PP and (aug-)cc-pwCVXZ-PP with X=D,T,Q,5 for Cu, Ag, Au and Zn, Cd and Hg
- Possibility to use auxiliary basis sets up to l functions
- New grids for Lanthanide and Actinides up to Lr
- Grid optimization: no 'unoptimized grid' for DFT any more
- DFT grids extended for diffuse cases
- Fixed numerical problems with quadrature
TURBOMOLE V5.8 (November 2005)
New features include:
- Implementation of new basis sets for the elements H - Rn, except lanthanides, which guarantee consistent accuracy across the periodic table.
- Small core ECPs for 5p and 6p elements and corresponding basis sets.
- ECP routines including g-projectors (energy and gradient only).
- Parallel multipole accelerated RI-J (MARI-J).
- Faster integral routines, especially for RI treatments.
- Improved functionality and simplified handling of tools for molecular properties, wavefunction analysis, and interfaces to visualization tools.
- Analytical excited state gradients for RI-CC2.
- Parallel RI-MP2 and RI-CC2 ground state and RI-CIS(D), RI-CC2, and RI-ADC(2) excitation energies (ricc2 module).
- Full support of the ricc2 program by the input module define.
- Additional auxiliary basis sets for RI-MP2/RI-CC2 calculations for the QZVPP (H-Kr) and cc-pwCVXZ (B-Ne, Al-Ar) orbital basis sets.
- COSMO with symmetry and more stable A-matrix setup.
- Virtual cavities for condensed phases and bond distance constraints added to molecular dynamics.
- A new and user-friendly input generator TMOLE with extended functionality (e.g. Z-matrix input, potential curve calculations).
- New DIIS methods and damping defaults for accelerated SCF convergence.
- Automatic switching from internal coordinates to Cartesians and back in case of internal coordinate failure for minimization and transition state optimization.
- Maximum number of atoms and basis functions increased to 700 and 10000 resp.
- better interfaces to graphics programs
available Turbomole add-ons:
- SNF provided by Carsten Kind, Markus Reiher, J. Neugebauer, C. Herrmann, and Bernd A. Hess
The program package SNF is a parallel (PVM or MPI) program for the calculation of vibrational frequencies, IR and Raman intensities.
- qmpot provided by Marek Sierka.
The program qmpot is able to perform transition state optimizations (and more...) using Turbomole.
TURBOMOLE V5.7 (July 2004)
New features include:
- Effective core potentials supporting basis sets with high l-quantum numbers
- Improved analytic second derivative calculations (for basis sets with up to d functions)
- Excitation energies for the algebraic diagrammatic construction through second order ADC (2)
- Visualization of ground and excited state densities and differential densities
- RI-J approximation for excited state properties, geometry optimizations, and (numerical) force constante calculations using TDDFT
- Analytical gradients for multipole accelerated RI-J (MARIJ)
- TPSS and TPSSH meta-GGA functionals (energies and gradients), including RI-J
- Gradients for the optimization of auxiliary basis functions for RI-MP2 and RI-CC2 calculations
- Basis sets of quadruple zeta valence quality for atoms H-Kr
- Improved transition state optimization; new optimizer for ground states (alternative to relax)
- Spin-unrestricted exact exchange methods using the LHF approximation
- Various other improvements, including a better documentation (user's guide)
TURBOMOLE V5.6 (December 2002)
- Excited state properties, geometry optimization, and (numerical) force constant calculations using TDDFT, CIS, and RPA
- RI-CC2 one-electron properties for ground and excited states, ground geometry optimization and force constant calculations (using numerical a hessian)
- Improved analytic second derivative calculations (including partial use of the RI approximation)
- Direct iterative methods for computing the lowest vibrational modes of a molecule
- Efficient calculation of optical rotations using TDDFT
- "O(N)-RIDFT" using the multipole accelerated RI-J approximation
- Occupation number optimization using (pseudo-Fermi) thermal smearing
- RI-JK approximation for Hartree-Fock including highly accurate optimized auxiliary basis sets
TURBOMOLE V5.5 (January 2002)
- calculation of electronic excitation energies at the CIS, CIS(D) and CC2 level using either a closed shell RHF or a UHF SCF reference function. Employs the RI technique to approximate two-electron integrals
- Universal Force Field (UFF) for geometry optimizations at a force field level and calculation of the analytical cartesian hessian
- Gaussian version of B3-LYP (using the simplified VWN-III functional)
- analytic closed shell hessian calculation at DFT level
- efficient Block-Davidson algorithm for CIS, RPA and DFT hybrid functional excitation energies, including spin-unrestricted cases
- RI-TDDFT method for dynamic polarizabilities and rotatory dispersions
- new COSMO routines for closed and open shell calculations of the energy and the gradients
- transition states search algorithm using the reduced gradient following method
- aug-cc-pVDZ, aug-cc-pVTZ and aug-cc-pVQZ Dunning basis sets
- online documentation, FAQ and Turbomole user mailing list available
- new DFT routines with better scaling
- new and very fast linear algebra routines with low memory requirements
- better scaling of RI integral routines
- new CPHF routines for more stable force calculations
- 64Bit binaries on HP (HP-UX 11)
- no 2GB file size limit on Linux machines (Kernel 2.4 required)
TURBOMOLE V5.3 (April 2000)
- New documentation (more detailed)
- COSMO code parallelized
- Point charges: number limitation removed, contribution to the gradient implemented
TURBOMOLE V5.2 (November 1999)
- Automatic definition of redundant internal coordinates
- Chemical shifts with DFT
- All-electron basis sets for Rb to Xe
- Standard calculations of molecules with up to 7000 basis functions (minimum storage requirements: 250 MB)
TURBOMOLE V5-1 (February 1999)
- FORTRAN-90 dynamic memory allocation
- ESCF: performance improved by simulataneous vector iteration (Block Davidson Algorithm). Calculation of chiroptical properties.
- Better grid-point distribution in DFT quadrature (up to 30% acceleration)
TURBOMOLE V4-9 (September 1998)
- RIMP2 with optimized auxiliary basis sets