Skip to main content
. 2023 Jun 29;19(20):6859–6890. doi: 10.1021/acs.jctc.3c00347

Table 1. Available Parallelizations of Various Modules Shown in Version 7.7a.

module functionality Fork-SMP OpenMP MPI OpenMP/MPI GPU
dscf HF/DFT energy
grad HF/DFT gradient
ridft RI-HF/RI-DFT energy X
rdgrad RI-HF/RI-DFT gradient X
aoforce HF/DFT Hessian
escf HF/DFT/GW-BSE excitation energies
egrad HF/DFT excited-state gradient X X
mpshift NMR/EPR parameters (HF/DFT/MP2) X X X
evib electron transport (HF/DFT) X
odft orbital-dependent DFT energies X X X X
mpgrad MP2 energy, gradient X X X
ricc2 RI-MP2, ADC(2), CC2 energies, gradients, spectra X X
pnoccsd PNO-MPPT and PNO–CC energies with F12 X X
ccsdf12 CCSD, CCSD(T) energies with F12 X X X X
rirpa RPA energy, gradient X X X X
riper periodic HF/DFT energy, gradient X X X X
a

Fork-SMP94 and the OpenMP version46,62,9597 are restricted to calculations on a single node. MPI5,98101 and OpenMP/MPI hybrid102 implementations allow for the use of multiple nodes. The availability of first- and second-order derivatives as well as excitation energies is also indicated.