Large Scale Quantum Chemistry with Tensor Processing Units

Pederson, Ryan; Kozlowski, John; Song, Ruyi; Beall, Jackson; Ganahl, Martin; Hauru, Markus; Lewis, Adam G. M.; Yao, Yi; Mallick, Shrestha Basu; Blüm, Volker; Vidal, Guifré

doi:10.1021/acs.jctc.2c00876

Cited by 17 publications

(13 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to a need for the development of novel programming models, compiler technologies, and optimized libraries to target these platforms, the move to accelerators often requires the re-evaluation of algorithmic design due to fundamental differences in execution strategies being appropriate only for particular classes of workloads (e.g., vectorized, low precision, and high arithmetic-intensity). AI-hardware’s low mixed-precision floating-point operations, in particular, add new challenges to the numerical accuracy, algorithm stability, and convergence estimates for quantum chemical methodologies. − …”

Section: Programming Models and Software Integrationmentioning

confidence: 99%

A Perspective on Sustainable Computational Chemistry Software Development and Integration

Di Felice,

Mayes,

Richard

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

The power of quantum chemistry to predict the ground and excited state properties of complex chemical systems has driven the development of computational quantum chemistry software, integrating advances in theory, applied mathematics, and computer science. The emergence of new computational paradigms associated with exascale technologies also poses significant challenges that require a flexible forward strategy to take full advantage of existing and forthcoming computational resources. In this context, the sustainability and interoperability of computational chemistry software development are among the most pressing issues. In this perspective, we discuss software infrastructure needs and investments with an eye to fully utilize exascale resources and provide unique computational tools for next-generation science problems and scientific discoveries.

show abstract

Section: Programming Models and Software Integrationmentioning

confidence: 99%

A Perspective on Sustainable Computational Chemistry Software Development and Integration

Di Felice,

Mayes,

Richard

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

show abstract

“…The current TPU architecture integrates 4,096 chips into a so-called TPU Pod, which achieves 1.1 exaflops in aggregate at half precision. TPUs are publicly available for cloud computing and can be leveraged for fluids simulations (Wang et al, 2022) and other scientific computing tasks (Belletti et al, 2019;Lu et al, 2020;Pederson et al, 2022), with remarkable computational throughput and scalability. Large, high-bandwidth memory and fast chip-to-chip interconnects (currently 1.1 PB/s) contribute to the performance of TPUs and alleviate bottlenecks that computational fluid dynamics (CFD) applications typically face on accelerator platforms (Balaji, 2021).…”

Section: Introductionmentioning

confidence: 99%

Accelerating large-eddy simulations of clouds with Tensor Processing Units

Chammas

Wang

Schneider

et al. 2023

Preprint

View full text Add to dashboard Cite

Clouds, especially low clouds, are crucial for regulating Earth’s energy balance and mediating the response of the climate system to changes in greenhouse gas concentrations. Despite their importance for climate, they remain relatively poorly understood and are inaccurately represented in climate models. A principal reason is that the high computational expense of simulating them with large-eddy simulations (LES) has inhibited broad and systematic numerical experimentation and the generation of large datasets for training parametrization schemes for climate models. Here we demonstrate LES of low clouds on Tensor Processing Units (TPUs), application-specific integrated circuits that were originally developed for machine learning applications. We show that TPUs in conjunction with tailored software implementations can be used to simulate computationally challenging stratocumulus clouds in conditions observed during the Dynamics and Chemistry of Marine Stratocumulus (DYCOMS) field study. The TPU-based LES code successfully reproduces clouds during DYCOMS and opens up the large computational resources available on TPUs to cloud simulations. The code enables unprecedented weak and strong scaling of LES, making it possible, for example, to simulate stratocumulus with $10\times$ speedup over real-time evolution in domains with a $34.7 \mathrm{km} \times 53.8 \mathrm{km}$ horizontal cross section. The results open up new avenues for computational experiments and for substantially enlarging the sample of LES available to train parameterizations of low clouds.

show abstract

“…First-principles methods such as density functional theory (DFT) , can provide detailed insight into the structural and electronic properties of supported metal atoms, ,, and how they are affected by the atomic-scale structure of the substrate surface. However, periodic surface slab models often exhibit poor computational scaling behavior that limits the application of more accurate higher-rung density functional approximations (DFAs) when studying large, periodic models . Due to the exhaustive computational requirements, the choice of DFA is often limited in large-scale studies to generalized gradient approximations (GGAs) or meta-GGAs (MGGAs) when calculating the Kohn–Sham ground-state energy. , These DFAs typically estimate either the adsorption energy or the reaction barriers correctly, but rarely both. , GGAs also often lack inclusion of long-range dispersion interactions, which are crucial for an accurate description of hybrid organic–inorganic interfaces. , Long-range dispersion correction methods, such as the Grimme series of methods or many-body dispersion (MBD) approaches, − are well-established strategies to address this shortcoming.…”

Section: Introductionmentioning

confidence: 99%

Stability of Single Gold Atoms on Defective and Doped Diamond Surfaces

2023

View full text Add to dashboard Cite

Polycrystalline boron-doped diamond (BDD) is widely used as a working electrode material in electrochemistry, and its properties, such as its stability, make it an appealing support material for nanostructures in electrocatalytic applications. Recent experiments have shown that electrodeposition can lead to the creation of stable small nanoclusters and even single gold adatoms on the BDD surfaces. We investigate the adsorption energy and kinetic stability of single gold atoms adsorbed onto an atomistic model of BDD surfaces by using density functional theory. The surface model is constructed using hybrid quantum mechanics/ molecular mechanics embedding techniques and is based on an oxygen-terminated diamond (110) surface. We use the hybrid quantum mechanics/molecular mechanics method to assess the ability of different density functional approximations to predict the adsorption structure, energy, and barrier for diffusion on pristine and defective surfaces. We find that surface defects (vacancies and surface dopants) strongly anchor adatoms on vacancy sites. We further investigated the thermal stability of gold adatoms, which reveals high barriers associated with lateral diffusion away from the vacancy site. The result provides an explanation for the high stability of experimentally imaged single gold adatoms on BDD and a starting point to investigate the early stages of nucleation during metal surface deposition.

show abstract

Large Scale Quantum Chemistry with Tensor Processing Units

Cited by 17 publications

References 56 publications

A Perspective on Sustainable Computational Chemistry Software Development and Integration

A Perspective on Sustainable Computational Chemistry Software Development and Integration

Accelerating large-eddy simulations of clouds with Tensor Processing Units

Stability of Single Gold Atoms on Defective and Doped Diamond Surfaces

Contact Info

Product

Resources

About