Skip to content

GitLab

Last edited by Andrew Wildman
Page history

CC

Table of contents

Details

The CC section handles parameters for coupled cluster (CC) calculations. This is a required section for coupled cluster jobs. Currently, only two-component wavefunctions (GHF/X2CHF) with the singles and doubles approximation are implemented.

To run a GHF/X2C-CCSD calculation, Job = CC must be specified in the [QM] section and ALG = INCORE must be set in the [INTS] section.

Compilation notes

To use CC code, ChronusQ must be compiled with TiledArray.1 Currently, only the GCC compiler with MPICH is tested. There is a known issue trying to use the Intel compilers - work is underway to resolve this issue. If you have all the prerequisites, the following commands show how to compile ChronusQ with TiledArray.

git clone https://urania.chem.washington.edu/chronusq/chronusq_public.git
mkdir chronusq_public/build && cd chronusq_public/build
cmake \
      -DCMAKE_CXX_FLAGS="-O3" \
      -DCQ_ENABLE_MPI=ON \
      -DCQ_ENABLE_TA=ON \
      -DCMAKE_CXX_COMPILER=mpicxx \
      -DCMAKE_C_COMPILER=mpicc \
      -DCMAKE_Fortran_COMPILER=mpif90 \
      -DCMAKE_BUILD_TYPE=Release \
      ..
make -j 12

If you need to point to a particular ScaLAPACK library, use the CMake variable CQ_SCALAPACK_LIBRARIES to point to the particular library file you wish to use. (e.g. ${HOME}/local/lib/libscalapack.a)

Keywords

Keyword Type Description Default Required?
TYPE String A specific Coupled-cluster method CCSD Yes
USEDIIS Bool Use DIIS in CC iterations or not TRUE No
NDIIS Integer Number of DIIS vectors kept 8 No
ETOL Double precision float Criterion of energy convergence in CC iterations 10^{-8} No
TTOL Double precision float Criterion of amplitude convergence in CC iterations 10^{-6} No
MAXITER Integer Maximum number of CC iterations 1000 No

The TYPE Keyword

ChronusQ currently only supports coupled cluster singles and doubles (CCSD).

Examples

GHF CCSD

[QM]
Reference = GHF
Job = CC

[CC]
Type = CCSD

[INTS]
Alg = INCORE

X2C CCSD

[QM]
Reference = X2CHF
Job = CC

[CC]
Type = CCSD

[INTS]
Alg = INCORE

Citations

  1. "TiledArray: A general-purpose scalable block-sparse tensor framework", Justus A. Calvin and Edward F. Valeev, https://github.com/valeevgroup/tiledarray ↩