Q-MP2-OS: Møller-Plesset Correlation Energy by Quadrature.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Calculating electron correlation energy is crucial for accurate molecular modeling.
• Traditional methods for MP2 correlation energy face computational challenges with large systems.

Purpose of the Study:

• To develop a computationally efficient algorithm for the opposite-spin component of MP2 correlation energy.
• To enable accurate calculations for large molecules through parallelization and reduced scaling.

Main Methods:

• A quadrature-based algorithm is employed, recasting the energy as a seven-dimensional integral.
• The algorithm abandons two-electron integrals and uses a cutoff strategy to minimize intermediate quantities.
• Implementation is discussed with a focus on large-scale parallelization.

Main Results:

• The algorithm exhibits quadratic scaling with basis set size.
• Near-perfect parallelization was achieved on 840 cores for cyclosporine (approx. 200 atoms).
• A scaling behavior was demonstrated for polyglycine sequences.

Conclusions:

• The developed algorithm is highly efficient and scalable for large molecular systems.
• Quadrature accuracy is the primary limitation, suggesting avenues for future method development.
• This approach significantly advances the feasibility of high-accuracy electronic structure calculations.