A third-generation density-functional-theory-based method for calculating canonical molecular orbitals of large...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Calculating canonical molecular orbitals (CMOs) for large molecules using density-functional theory (DFT) presents significant computational challenges.
• Existing methods often struggle with scalability and efficiency for large molecular systems.

Purpose of the Study:

• To develop a grid-free, computationally efficient DFT-based method for calculating CMOs of large molecules.
• To leverage next-generation supercomputing architectures for enhanced molecular orbital calculations.

Main Methods:

• Utilized modified Cholesky decomposition (CD) for a grid-free approach.
• Employed low-rank pivoted CD and CD with adaptive metric (CDAM) for analytical downscaling of Cholesky vectors.
• Implemented a parallel computing strategy for distributing and storing Cholesky vectors.
• Calculated Coulomb, Fock exchange, and exchange-correlation terms by multiplying Cholesky vectors, avoiding explicit molecular integral evaluation.

Main Results:

• Successfully developed a DFT-based method for calculating CMOs of large molecules.
• Achieved analytical computation of exchange-correlation terms.
• Demonstrated efficient utilization of massively distributed memory parallel computers.
• Enabled calculation of CMOs without evaluating molecular integrals within self-consistent field iterations.

Conclusions:

• The developed method offers a highly efficient approach for calculating CMOs of large molecules using DFT.
• The technique effectively maximizes the capacity of modern supercomputers for quantum chemistry computations.
• This advancement facilitates larger and more complex molecular simulations in computational chemistry.