A5024994202- Advanced Chemical Physics Studies
- Parallel Computing and Optimization Techniques
- Spectroscopy and Quantum Chemical Studies
- Advanced NMR Techniques and Applications
- Machine Learning in Materials Science
- Genomics and Phylogenetic Studies
- Microbial Inactivation Methods
- Distributed and Parallel Computing Systems
- Quantum and electron transport phenomena
- Physics of Superconductivity and Magnetism
- Ultrasound and Cavitation Phenomena
- Evolution and Paleontology Studies
- Mass Spectrometry Techniques and Applications
- Protein Structure and Dynamics
- Advanced Data Storage Technologies
- Industrial Vision Systems and Defect Detection
- Mathematical functions and polynomials
- Chemical Thermodynamics and Molecular Structure
- Quantum Computing Algorithms and Architecture
- Scientific Computing and Data Management
- Cloud Computing and Resource Management
- Nanocluster Synthesis and Applications
- Quantum, superfluid, helium dynamics
- Catalysis and Oxidation Reactions
- Enzyme Structure and Function
The University of Melbourne
2024-2025
Australian National University
2014-2024
University of Salerno
2013-2015
Ospedale Annunziata di Cosenza
2003
A discussion of many the recently implemented features GAMESS (General Atomic and Molecular Electronic Structure System) LibCChem (the C++ CPU/GPU library associated with GAMESS) is presented. These include fragmentation methods such as fragment molecular orbital, effective potential orbital methods, hybrid MPI/OpenMP approaches to Hartree-Fock, resolution identity second order perturbation theory. Many new coupled cluster theory have been in GAMESS, multiple levels density functional/tight...
We present a single-determinant approach to three challenging topics in the chemistry of excited states: double excitations, charge-transfer states, and conical intersections. The results are obtained by using Initial Maximum Overlap Method (IMOM) which is modified version (MOM). new algorithm converges better than original, especially for these difficult problems. By considering several case studies, we show that framework provides simple accurate alternative modeling states cases where...
The understanding of recombination photogenerated electron/hole pairs at defect sites is a key enabler to develop bismuth vanadate (BiVO4) photoanodes scale and low cost for photoelectrochemical water splitting. Here, we report systematic investigation the impact vanadium vacancies on efficiency BiVO4 photooxidation. X-ray photoelectron photoluminescence spectroscopies reveal that surfaces nanostructured obtained by high-temperature synthesis, here used as model system, suffer from...
The primary focus of GAMESS over the last 5 years has been development new high-performance codes that are able to take effective and efficient advantage most advanced computer architectures, both CPU accelerators. These efforts include employing density fitting fragmentation methods reduce high scaling well-correlated (e.g., coupled-cluster) as well developing novel can optimal graphical processing units other modern Because accurate wave functions be very complex, an important...
This article presents a novel algorithm for the calculation of analytic energy gradients from second-order Møller–Plesset perturbation theory within Resolution-of-the-Identity approximation (RI-MP2), which is designed to achieve high performance on clusters with multiple graphical processing units (GPUs). The uses GPUs all major steps calculation, including integral generation, formation required intermediate tensors, solution Z-vector equation and gradient accumulation. implementation in...
A novel implementation of the self-consistent field (SCF) procedure specifically designed for high-performance execution on multiple graphics processing units (GPUs) is presented. The algorithm offloads to GPUs three major computational stages SCF, namely, calculation one-electron integrals, and digestion electron repulsion diagonalization Fock matrix, including SCF acceleration via DIIS. Performance results a variety test molecules basis sets show remarkable speedups with respect...
Electronic structure calculations have the potential to predict key matter transformations for applications of strategic technological importance, from drug discovery material science and catalysis. However, a predictive physicochemical characterization these processes often requires accurate quantum chemical modeling complex molecular systems with hundreds thousands atoms. Due computationally demanding nature electronic complexity modern high-performance computing hardware, chemistry...
This Article presents two optimized multi-GPU algorithms for Fock matrix construction, building on the work of Ufimtsev and Martinez [
Hartree-Fock (HF) theory is most often applied to study the electronic ground states of molecular systems. However, with advent numerical techniques for locating higher solutions self-consistent field equations, it now possible examine extent which such mean-field are useful approximations excited states. In this Communication, we use maximum overlap method locate 11 low-energy HF equation H2 molecule and find that, only one exception, these yield surprisingly accurate models low-lying...
How many electrons are excited in an electronic transition? In this Letter, we introduce the excitation number η to answer question when initial and final states each modeled by a single-determinant wave function. We show that calculated values lie close positive integers, leading unambiguous assignments of electrons. This contrasts with previous definitions quantities which can lead mis-assignments. consider several examples where provides improved excited-state characterizations.
Electronic structure theory (especially quantum chemistry) has thrived and become increasingly relevant to a broad spectrum of scientific endeavors as the sophistication both computer architectures software engineering advanced. This article provides brief history advances in hardware software, from early days IBM mainframes current emphasis on accelerators modern programming practices.
We present a high-performance, GPU (graphics processing unit)-accelerated algorithm for building the Fock matrix. The is designed efficient calculations on large molecular systems and uses novel dynamic load balancing scheme that maximizes throughput avoids thread divergence could occur due to integral screening. Additionally, code adopts ERI digestion exploits all forms of permutational symmetry, combines efficiently evaluation both Coulomb exchange terms together, eliminates explicit...
In Self-Consistent Field (SCF) calculations, the choice of initial guess plays a key role in determining time-to-solution by influencing number iterations required for convergence. However, focusing solely on reducing may overlook computational cost associated with improving accuracy guesses. This study critically evaluates effectiveness two methods─basis set projection (BSP) and many-body expansion (MBE) Hartree-Fock hybrid Density Functional Theory (B3LYP MN15) methods. We also introduce...
A novel methodology for the evaluation of two electron integrals up to f functions using Graphics Processing Units (GPUs) is presented. The Head-Gordon-Pople recursion relationships are solved via a simple heuristic minimize number evaluated intermediates in trees. Automatic code generation used generate highly optimized CUDA kernels. approach presented which integral classes split into smaller subclasses register pressure and exploit additional parallelism at cost recomputing small...
As computer systems dedicated to scientific calculations become massively parallel, the poor parallel performance of Fock matrix diagonalization becomes a major impediment achieving larger molecular sizes in self-consistent field (SCF) calculations. In this Article, novel, highly and diagonalization-free algorithm for accelerated convergence SCF procedure is presented. The algorithm, called Q-Next, draws on second-order SCF, quadratically convergent direct inversion iterative subspace (DIIS)...
We report the three main ingredients to calculate three- and four-electron integrals over Gaussian basis functions involving geminal operators: fundamental integrals, upper bounds, recurrence relations. In particular, we consider that may arise in explicitly correlated F12 methods. A straightforward method obtain is given. derive vertical, transfer, horizontal relations build up angular momentum centers. Strong, simple, scaling-consistent bounds are also reported. This latest ingredient...
The evaluation of contracted two-electron integrals over a Gaussian geminal operator is pivotal to diverse quantum chemistry methods. In this article, using the unique factorization properties and sparsity these integrals, novel, near-optimal computation algorithm presented. Our method employs combination recently developed upper bounds, recurrence relations in spirit Head-Gordon-Pople approach, late- early-contraction paths PRISM style. A detailed study FLOP (floating-point operations) cost...
Usage of Graphics Processing Units (GPU) has become strategic for simulating the chemistry large molecular systems, with majority top supercomputers utilizing GPUs as their main source computational horsepower. In this paper, a new fragmentation-based Hartree-Fock matrix build algorithm designed scaling on many-GPU architectures is presented. The uses novel dynamic load balancing scheme based binned shell-pair container to distribute batches significant shell quartets same code path...
Correlated electronic structure calculations enable an accurate prediction of the physicochemical properties complex molecular systems; however, scale these is limited by their extremely high computational cost. The Fragment Molecular Orbital (FMO) method arguably one most effective ways to lower this cost while retaining predictive accuracy. In paper, a novel distributed many-GPU algorithm and implementation FMO are presented. When applied in tandem with Hartree-Fock RI-MP2 methods, new...
We present a quadrature-based algorithm for computing the opposite-spin component of MP2 correlation energy which scales quadratically with basis set size and is well-suited to large-scale parallelization. The key ideas, are rooted in earlier work Hirata co-workers, abandon all two-electron integrals, recast as seven-dimensional integral, approximate that integral by quadrature, employ cutoff strategy minimize number intermediate quantities. discuss our implementation detail show it...
Second-order Møller-Plesset perturbation theory using the Resolution-of-the-Identity approximation (RI-MP2) is a state-of-the-art approach to accurately estimate many-body electronic correlation effects. This critical for predicting physicochemical properties of complex molecular systems; however, scale these calculations limited by their extremely high computational cost. In this paper, novel many-GPU algorithm and implementation molecular-fragmentation-based RI-MP2 method are presented...
Explicitly-correlated F12 methods are becoming the first choice for high-accuracy molecular orbital calculations, and can often achieve chemical accuracy with relatively small gaussian basis sets. In most many three- four-electron integrals that formally appear in theory avoided through judicious use of resolutions identity (RI). However, order not to jeopardize intrinsic wave function, associated RI auxiliary set must be large. Here, inspired by Head-Gordon-Pople (HGP) PRISM algorithms...