2003
DOI: 10.1063/1.1575193
|View full text |Cite
|
Sign up to set email alerts
|

The parallel implementation of a full configuration interaction program

Abstract: Both the replicated and distributed data parallel full configuration interaction (FCI) implementations are described. The implementation of the FCI algorithm is organized in a hybrid strings-integral driven approach. Redundant communication is avoided, and the network performance is further optimized by an improved distributed data interface library. Examples show linear scalability of the distributed data code on both PC and workstation clusters. The new parallel implementation greatly extends the hardware on… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
40
0
1

Year Published

2004
2004
2021
2021

Publication Types

Select...
9

Relationship

5
4

Authors

Journals

citations
Cited by 42 publications
(42 citation statements)
references
References 17 publications
1
40
0
1
Order By: Relevance
“…It can be used with single-as well as multiconfigurational zeroth-order reference functions; it systematically approximates the full CI limit while monitoring the closeness of this approximation along the way; it uses standard computational CI machinery and its efficiency is increased by the parallelization of these codes. 103,104 Application of this method using Dunning's correlation-consistent basis sets, 105 in conjunction with CBS extrapolations yielded 101 the binding energies of the diatomic molecules, C 2 , N 2 , O 2 , and F 2 , within the "chemical accuracy" 106 of 1.6 mhartree, comparing well with the recent experimentally based benchmark estimates of O'Neill and Gill. 107 In the present investigation, we demonstrate that the CEEIS method can also be successfully applied along a reaction path where the zeroth-order reference function is multiconfigurational and undergoes strong changes.…”
Section: Introductionsupporting
confidence: 54%
“…It can be used with single-as well as multiconfigurational zeroth-order reference functions; it systematically approximates the full CI limit while monitoring the closeness of this approximation along the way; it uses standard computational CI machinery and its efficiency is increased by the parallelization of these codes. 103,104 Application of this method using Dunning's correlation-consistent basis sets, 105 in conjunction with CBS extrapolations yielded 101 the binding energies of the diatomic molecules, C 2 , N 2 , O 2 , and F 2 , within the "chemical accuracy" 106 of 1.6 mhartree, comparing well with the recent experimentally based benchmark estimates of O'Neill and Gill. 107 In the present investigation, we demonstrate that the CEEIS method can also be successfully applied along a reaction path where the zeroth-order reference function is multiconfigurational and undergoes strong changes.…”
Section: Introductionsupporting
confidence: 54%
“…However, parallelization of determinant comparisons for this code has not been achieved and few parallel implementations of determinant-based full CI codes have been introduced. [43][44][45][46] Constructing A 22 uses coupling constants in the evaluation of Hamiltonian matrix elements [see eq. (25)], which is straightforward but not parallelized as mentioned earlier.…”
Section: Ci-orbital and Ci-ci Blocksmentioning
confidence: 99%
“…Jointly, programs effi ciency follows this increase due to calculation methodology improvements (AIKENS, 2004;ALEXEEV, 2002;BOLDING, 2000;CHOI, 2003;FAMULARI, 1998;FEDOROV, 2004;GLAESEMANN, 1998). In 1995, Strout and Scuseria (STROUT, 1995) presented details about effects of integrals screening on the scaling properties of HF methods in large molecular systems treated in (ALMLÖF, 1982).…”
Section: Scaling Properties Of the Hartree-fock Methodsmentioning
confidence: 99%