The GPU‐enabled divide‐expand‐consolidate RI‐MP2 method (DEC‐RI‐MP2)

We report porting of the Divide‐Expand‐Consolidate Resolution of the Identity second‐order Møller–Plesset perturbation (DEC‐RI‐MP2) method to the graphic processing units (GPUs) using OpenACC compiler directives. It is shown that the OpenACC compiler directives implementation efficiently accelerates the rate‐determining step of the DEC‐RI‐MP2 method with minor implementation effort. Moreover, the GPU acceleration results in a better load balance and thus in an overall scaling improvement of the DEC algorithm. The resulting cross‐platform hybrid MPI/OpenMP/OpenACC implementation has scalable and portable performance on heterogeneous HPC architectures. The GPU‐enabled code was benchmarked using a reduced version of the S12L test set of Stefan Grimme (Grimme, Chem. Eur. J. 2012, 18, 9955) consisting of supramolecular complexes up to 158 atoms and 4292 contracted basis functions (cc‐pVTZ). The test set results demonstrate the general applicability of the DEC‐RI‐MP2 method showing results consistent with the DEC‐RI‐MP2 introductory paper (Baudin et al., J. Chem. Phys. 2016, 144, 054102) on molecules of complicated electronic structures. © 2016 Wiley Periodicals, Inc.

[1]  Frank Neese,et al.  Sparse maps—A systematic infrastructure for reduced-scaling electronic structure methods. I. An efficient and simple linear scaling local MP2 method that uses an intermediate basis of pair natural orbitals. , 2015, The Journal of chemical physics.

[2]  Wei Li,et al.  Generalized energy-based fragmentation approach and its applications to macromolecules and molecular aggregates. , 2014, Accounts of chemical research.

[3]  Frederick R Manby,et al.  Optimization of orbital-specific virtuals in local Møller-Plesset perturbation theory. , 2012, The Journal of chemical physics.

[4]  Christof Hättig,et al.  Optimization of auxiliary basis sets for RI-MP2 and RI-CC2 calculations: Core–valence and quintuple-ζ basis sets for H to Ar and QZVPP basis sets for Li to Kr , 2005 .

[5]  C. Van Alsenoy,et al.  Ab initio calculations on large molecules: The multiplicative integral approximation , 1988 .

[6]  Roland Lindh,et al.  Atomic Cholesky decompositions: a route to unbiased auxiliary basis sets for density fitting approximation with tunable accuracy and efficiency. , 2009, The Journal of chemical physics.

[7]  S. F. Boys,et al.  Canonical Configurational Interaction Procedure , 1960 .

[8]  Frederick R. Manby,et al.  Fast linear scaling second-order Møller-Plesset perturbation theory (MP2) using local and density fitting approximations , 2003 .

[9]  Branislav Jansík,et al.  Linear scaling coupled cluster method with correlation energy based error control. , 2010, The Journal of chemical physics.

[10]  Michio Katouda,et al.  Efficient parallel algorithm of second‐order Møller–Plesset perturbation theory with resolution‐of‐identity approximation (RI‐MP2) , 2009 .

[11]  Holger Patzelt,et al.  RI-MP2: optimized auxiliary basis sets and demonstration of efficiency , 1998 .

[12]  Poul Jørgensen,et al.  The divide-expand-consolidate family of coupled cluster methods: numerical illustrations using second order Møller-Plesset perturbation theory. , 2012, The Journal of chemical physics.

[13]  Takahito Nakajima,et al.  Massively parallel algorithm and implementation of RI‐MP2 energy calculation for peta‐scale many‐core supercomputers , 2016, J. Comput. Chem..

[14]  Dmitrij Rappoport,et al.  Development of new auxiliary basis functions of the Karlsruhe segmented contracted basis sets including diffuse basis functions (def2-SVPD, def2-TZVPPD, and def2-QVPPD) for RI-MP2 and RI-CC calculations. , 2015, Physical chemistry chemical physics : PCCP.

[15]  Philippe Y. Ayala,et al.  Linear scaling second-order Moller–Plesset theory in the atomic orbital basis for large molecular systems , 1999 .

[16]  J. Almlöf,et al.  Integral approximations for LCAO-SCF calculations , 1993 .

[17]  Shridhar R. Gadre,et al.  Molecular tailoring approach in conjunction with MP2 and Ri‐MP2 codes: A comparison with fragment molecular orbital method , 2010, J. Comput. Chem..

[18]  Takeshi Ishikawa,et al.  Fragment molecular orbital calculation using the RI-MP2 method , 2009 .

[19]  Christian Ochsenfeld,et al.  A Linear-Scaling MP2 Method for Large Molecules by Rigorous Integral-Screening Criteria , 2010 .

[20]  Jan Almlöf,et al.  Elimination of energy denominators in Møller—Plesset perturbation theory by a Laplace transform approach , 1991 .

[21]  F. Weigend,et al.  Efficient use of the correlation consistent basis sets in resolution of the identity MP2 calculations , 2002 .

[22]  Shuhua Li,et al.  Improved cluster-in-molecule local correlation approach for electron correlation calculation of large systems. , 2014, The journal of physical chemistry. A.

[23]  Oreste Villa,et al.  Noniterative Multireference Coupled Cluster Methods on Heterogeneous CPU-GPU Systems. , 2013, Journal of chemical theory and computation.

[24]  Sriram Krishnamoorthy,et al.  GPU-Based Implementations of the Noniterative Regularized-CCSD(T) Corrections: Applications to Strongly Correlated Systems. , 2011, Journal of chemical theory and computation.

[25]  Guo Yang,et al.  The relative energies of polypeptide conformers predicted by linear scaling second-order Møller-Plesset perturbation theory , 2014 .

[26]  Kimihiko Hirao,et al.  An approximate second-order Møller–Plesset perturbation approach for large molecular calculations , 2006 .

[27]  Michael Dolg,et al.  Fully Automated Incremental Evaluation of MP2 and CCSD(T) Energies: Application to Water Clusters. , 2009, Journal of chemical theory and computation.

[28]  Thomas Kjærgaard,et al.  Efficient linear-scaling second-order Møller-Plesset perturbation theory: The divide-expand-consolidate RI-MP2 model. , 2016, The Journal of chemical physics.

[29]  Masato Kobayashi,et al.  Alternative linear-scaling methodology for the second-order Møller-Plesset perturbation calculation based on the divide-and-conquer method. , 2007, The Journal of chemical physics.

[30]  S. F. Boys.  Construction of Some Molecular Orbitals to Be Approximately Invariant for Changes from One Molecule to Another , 1960 .

[31]  Luca Frediani,et al.  The Dalton quantum chemistry program system , 2013, Wiley interdisciplinary reviews. Computational molecular science.

[32]  Evert Jan Baerends,et al.  Self-consistent molecular Hartree—Fock—Slater calculations I. The computational procedure , 1973 .

[33]  Thomas Kjærgaard,et al.  Molecular gradient for second-order Møller-Plesset perturbation theory using the divide-expand-consolidate (DEC) scheme. , 2012, The Journal of chemical physics.

[34]  Artur F Izmaylov,et al.  Resolution of the identity atomic orbital Laplace transformed second order Møller-Plesset theory for nonconducting periodic systems. , 2008, Physical chemistry chemical physics : PCCP.

[35]  J. L. Whitten,et al.  Coulombic potential energy integrals and approximations , 1973 .

[36]  P. Jørgensen,et al.  Orbital localization using fourth central moment minimization. , 2012, The Journal of chemical physics.

[37]  Koji Yasuda,et al.  Accelerating Density Functional Calculations with Graphics Processing Unit. , 2008, Journal of chemical theory and computation.

[38]  Jörg Kussmann,et al.  Linear-scaling atomic orbital-based second-order Møller-Plesset perturbation theory by rigorous integral screening criteria. , 2009, The Journal of chemical physics.

[39]  Ivan S Ufimtsev,et al.  Quantum Chemistry on Graphical Processing Units. 1. Strategies for Two-Electron Integral Evaluation. , 2008, Journal of chemical theory and computation.

[40]  Thomas Bondo Pedersen,et al.  Reduced scaling in electronic structure calculations using Cholesky decompositions , 2003 .

[41]  Rick A. Kendall,et al.  The impact of the resolution of the identity approximate integral method on modern ab initio algorithm development , 1997 .

[42]  Kasper Kristensen,et al.  The divide–expand–consolidate MP2 scheme goes massively parallel , 2013 .

[43]  Poul Jørgensen,et al.  Pipek–Mezey localization of occupied and virtual orbitals , 2013, J. Comput. Chem..

[44]  J. Kussmann,et al.  Efficient distance-including integral screening in linear-scaling Møller-Plesset perturbation theory. , 2013, The Journal of chemical physics.

[45]  Shuhua Li,et al.  An efficient localization procedure for large systems using a sequential transformation strategy. , 2014, The Journal of chemical physics.

[46]  Paul G. Mezey,et al.  A fast intrinsic localization procedure applicable for ab initio and semiempirical linear combination of atomic orbital wave functions , 1989 .

[47]  F. Weigend,et al.  RI-MP2: first derivatives and global consistency , 1997 .

[48]  Ivan S. Ufimtsev,et al.  An atomic orbital-based formulation of the complete active space self-consistent field method on graphical processing units. , 2015, The Journal of chemical physics.

[49]  Benjamin G. Levine,et al.  Nanoscale multireference quantum chemistry: full configuration interaction on graphical processing units. , 2015, Journal of chemical theory and computation.

[50]  Lorenzo Maschio,et al.  Local MP2 with Density Fitting for Periodic Systems: A Parallel Implementation. , 2011, Journal of chemical theory and computation.

[51]  Hans-Joachim Werner,et al.  An efficient local coupled cluster method for accurate thermochemistry of large systems. , 2011, The Journal of chemical physics.

[52]  Klaus Ruedenberg,et al.  Localized Atomic and Molecular Orbitals. II , 1965 .

[53]  Klaus Ruedenberg,et al.  Localized Atomic and Molecular Orbitals , 1963 .

[54]  P. Jørgensen,et al.  A perspective on the localizability of Hartree–Fock orbitals , 2013, Theoretical Chemistry Accounts.

[55]  Brett M. Bode,et al.  Uncontracted Rys Quadrature Implementation of up to G Functions on Graphical Processing Units. , 2010, Journal of chemical theory and computation.

[56]  D. Bernholdt,et al.  Large-scale correlated electronic structure calculations: the RI-MP2 method on parallel computers , 1996 .

[57]  P Pulay,et al.  Local Treatment of Electron Correlation , 1993 .

[58]  Georg Hetzer,et al.  Low-order scaling local electron correlation methods. I. Linear scaling local MP2 , 1999 .

[59]  Matías A Nitsche,et al.  GPU Accelerated Implementation of Density Functional Theory for Hybrid QM/MM Simulations. , 2014, Journal of chemical theory and computation.

[60]  Marcin Ziółkowski,et al.  A Locality Analysis of the Divide-Expand-Consolidate Coupled Cluster Amplitude Equations. , 2011, Journal of chemical theory and computation.

[61]  Wei Li,et al.  Generalized energy-based fragmentation approach for computing the ground-state energies and properties of large molecules. , 2007, The journal of physical chemistry. A.

[62]  Roland Lindh,et al.  Unbiased auxiliary basis sets for accurate two-electron integral approximations. , 2007, The Journal of chemical physics.

[63]  Takahito Nakajima,et al.  MPI/OpenMP Hybrid Parallel Algorithm of Resolution of Identity Second-Order Møller-Plesset Perturbation Calculation for Massively Parallel Multicore Supercomputers. , 2013, Journal of chemical theory and computation.

[64]  S. Grimme.  Supramolecular binding thermodynamics by dispersion-corrected density functional theory. , 2012, Chemistry.

[65]  Mihály Kállay,et al.  An Integral-Direct Linear-Scaling Second-Order Møller-Plesset Approach. , 2016, Journal of chemical theory and computation.

[66]  H. Werner,et al.  On the use of Abelian point group symmetry in density-fitted local MP2 using various types of virtual orbitals. , 2015, The Journal of chemical physics.

[67]  Angela K. Wilson,et al.  Gaussian basis sets for use in correlated molecular calculations. IX. The atoms gallium through krypton , 1993 .

[68]  John R. Sabin,et al.  On some approximations in applications of Xα theory , 1979 .

[69]  Yihan Shao,et al.  Accelerating resolution-of-the-identity second-order Møller-Plesset quantum chemistry calculations with graphical processing units. , 2008, The journal of physical chemistry. A.

[70]  Poul Jørgensen,et al.  Trust Region Minimization of Orbital Localization Functions. , 2012, Journal of chemical theory and computation.

[71]  Hans-Joachim Werner,et al.  Low-order scaling local electron correlation methods. IV. Linear scaling local coupled-cluster (LCCSD) , 2001 .

[72]  Dieter Cremer,et al.  Møller–Plesset perturbation theory: from small molecule methods to methods for thousands of atoms , 2011 .

[73]  Frederick R. Manby,et al.  Density fitting in second-order linear-r12 Møller–Plesset perturbation theory , 2003 .

[74]  Stinne Høst,et al.  Local orbitals by minimizing powers of the orbital variance. , 2011, The Journal of chemical physics.

[75]  Takeshi Ishikawa,et al.  RI-MP2 Gradient Calculation of Large Molecules Using the Fragment Molecular Orbital Method. , 2012, The journal of physical chemistry letters.

[76]  J. Connolly,et al.  On first‐row diatomic molecules and local density models , 1979 .

[77]  P. Jørgensen,et al.  Orbital spaces in the divide-expand-consolidate coupled cluster method. , 2016, The Journal of chemical physics.

[78]  Mark S Gordon,et al.  Quantum Chemical Calculations Using Accelerators: Migrating Matrix Operations to the NVIDIA Kepler GPU and the Intel Xeon Phi. , 2014, Journal of chemical theory and computation.

[79]  A Eugene DePrince,et al.  Coupled Cluster Theory on Graphics Processing Units I. The Coupled Cluster Doubles Method. , 2011, Journal of chemical theory and computation.

[80]  Branislav Jansík,et al.  Maximum locality in occupied and virtual orbital spaces using a least-change strategy. , 2009, The Journal of chemical physics.

[81]  Xiao He,et al.  The generalized molecular fractionation with conjugate caps/molecular mechanics method for direct calculation of protein energy. , 2006, The Journal of chemical physics.

[82]  Frank Neese,et al.  The ORCA program system , 2012 .

[83]  Martin W. Feyereisen,et al.  Use of approximate integrals in ab initio theory. An application in MP2 energy calculations , 1993 .

[84]  Ivan S. Ufimtsev,et al.  Dynamic Precision for Electron Repulsion Integral Evaluation on Graphical Processing Units (GPUs). , 2011, Journal of chemical theory and computation.

[85]  Marco Häser,et al.  Møller-Plesset (MP2) perturbation theory for large molecules , 1993 .

[86]  So Hirata,et al.  Monte Carlo MP2 on Many Graphical Processing Units. , 2016, Journal of chemical theory and computation.

[87]  Hans-Joachim Werner,et al.  Local treatment of electron correlation in coupled cluster theory , 1996 .

[88]  Thomas Kjærgaard,et al.  Linear-Scaling Coupled Cluster with Perturbative Triple Excitations: The Divide-Expand-Consolidate CCSD(T) Model. , 2015, Journal of chemical theory and computation.

[89]  Frederick R. Manby,et al.  Fast local-MP2 method with density-fitting for crystals. I. Theory and algorithms , 2007 .

[90]  Rodney J Bartlett,et al.  A natural linear scaling coupled-cluster method. , 2004, The Journal of chemical physics.

[91]  Peter Pulay,et al.  Localizability of dynamic electron correlation , 1983 .

[92]  Michio Katouda.  Application of resolution of identity approximation of second-order Møller–Plesset perturbation theory to three-body fragment molecular orbital method , 2011 .

[93]  Klaus Schulten,et al.  GPU-accelerated molecular modeling coming of age. , 2010, Journal of molecular graphics & modelling.

[94]  Michael A Collins,et al.  Approximate ab initio energies by systematic molecular fragmentation. , 2005, The Journal of chemical physics.

[95]  Thomas Kjærgaard,et al.  MP2 energy and density for large molecular systems with internal error control using the Divide-Expand-Consolidate scheme. , 2012, Physical chemistry chemical physics : PCCP.

[96]  Kaori Fukuzawa,et al.  Large scale FMO-MP2 calculations on a massively parallel-vector computer , 2008 .

[97]  F. Manby,et al.  Explicitly correlated second-order perturbation theory using density fitting and local approximations. , 2006, The Journal of chemical physics.