Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel optimization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,585 records · Page 88
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

The continuous readout stream of the MicroBooNE liquid argon time projection chamber for detection of supernova burst neutrinos

The MicroBooNE continuous readout stream is a parallel readout of the MicroBooNE liquid argon time projection chamber (LArTPC) which enables detection of non-beam events such as those from a supernova neutrino burst. The low energies of the supernova neutrinos and the intense cosmic-ray background flux due to the near-surface detector location makes triggering on these events very challenging. Instead, MicroBooNE relies on a delayed trigger generated by SNEWS (the Supernova Early Warning System) for detecting supernova neutrinos. The continuous readout of the LArTPC generates large data volumes, and requires the use of real-time compression algorithms (zero suppression and Huffman compression) implemented in an FPGA (field-programmable gate array) in the readout electronics. In this paper we present the results of the optimization of the data reduction algorithms, and their operational performance. To demonstrate the capability of the continuous stream to detect low-energy electrons, a sample of Michel electrons from stopping cosmic-ray muons is reconstructed and compared to a similar sample from the lossless triggered readout stream.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
NASA NTRS1999

Parallelization of Rocket Engine Simulator Software (P.R.E.S.S.)

Parallelization of Rocket Engine System Software (PRESS) project is part of a collaborative effort with Southern University at Baton Rouge (SUBR), University of West Florida (UWF), and Jackson State University (JSU). The project has started on October 19, 1995, and after a three-year period corresponding to project phases and fiscal-year funding by NASA Lewis Research Center (now Glenn Research Center), has ended on October 18, 1998. The one-year no-cost extension period was granted on June 7, 1998, until October 19, 1999. The aim of this one year no-cost extension period was to carry out further research to complete the work and lay the groundwork for subsequent research in the area of aerospace engine design optimization software tools. The previous progress for the research has been reported in great detail in respective interim and final research progress reports, seven of them, in all. While the purpose of this report is to be a final summary and an valuative view of the entire work since the first year funding, the following is a quick recap of the most important sections of the interim report dated April 30, 1999.

Cezzar, Ruknet↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗
NASA NTRS1962

Studies in the Fields of Space Flight and Guidance Theory: Progress Report - No. 2

Compiled in this paper are 11 progress papers from 7 of the agencies working under contract to MSFC in the areas of guidance theory and space flight theory. This is the second paper of the "Progress Reports" and covers the period from December 1, 1961 to June 15, 1962. Extensive references are made to "Progress Report No. 1." This second progress report is referred to as "report" and "Progress Report No. 1" will be referred to as the "first report" in this introduction. Information given in the first report is not repeated herein. The reports of the various contractors will be referred to by index number as papers. There are two parallel series of publications covering the over-all activities at MSFC in the areas of guidance theory and space flight theory. One is the series of progress reports of which this paper is the second in the series. The other is the series of "Status Reports on Theory of Space Flight and Adaptive Guidance." These series along with a few other special reports, give a complete picture of the immediate objectives, accomplishments, and final goals of Aeroballistics Division and associated contr

FLIGHT OPTIMIZATION↗
DOE OSTI2023

CSPlib: A performance portable parallel software toolkit for analyzing complex kinetic mechanisms

Computational singular perturbation (CSP) is a method to analyze dynamical systems. It targets the decoupling of fast and slow dynamics using an alternate linear expansion of the right-hand side of the governing equations based on eigenanalysis of the associated Jacobian matrix. This representation facilitates diagnostic analysis, detection and control of stiffness, and the development of simplified models. For this work, we have implemented CSP in a C++ open-source library CSPlib using the Kokkos parallel programming model to address portability across diverse heterogeneous computing platforms, i.e., multi/many-core CPUs and GPUs. We describe the CSPlib implementation and present its computational performance across different computing platforms using several test problems. Specifically, we test the CSPlib performance for a constant pressure ignition reactor model on different architectures, including IBM Power 9, Intel Xeon Skylake, and NVIDIA V100 GPU. The size of the chemical kinetic mechanism is varied in these tests. As expected, the Jacobian matrix evaluation, the eigensolution of the Jacobian matrix, and matrix inversion are the most expensive computational tasks. When considering the higher throughput characteristic of GPUs, GPUs performs better for small matrices with higher occupancy rate. CPUs gain more advantages from the higher performance of well-tuned and optimized linear algebra libraries such as OpenBLAS.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1988

Laminate behavior for SiC fiber-reinforced reaction-bonded silicon nitride matrix composites

The room temperature mechanical properties of SiC fiber reinforced reaction-bonded silicon nitride matrix composite laminates (SiC/RBSN) have been measured. The laminates contained approx 30 volume fraction of aligned 142-micron diameter SiC fiber in a porous RBSN matrix. Three types of laminate studied were unidirectional: (1) (0) sub 8, (2) (10) sub 8, and (3) (45) sub 8, and (90) sub 8; cross plied laminates (0 sub 2/90 sub 2); and angle plied laminates: (+45 sub 2/-45 sub 2). Each laminate contained eight fiber plies. Results of the unidirectionally reinforced composites tested at various angles to the reinforcement direction indicate large anisotropy in in-plane properties. In addition, strength properties of these composites along the fiber direction were independent of specimen gage length and were unaffected by notches normal to the fiber direction. Splitting parallel to the fiber at the notch tip appears to be the dominant crack blunting mechanism responsible for notch insensitive behavior of these composites. In-plane properties of the composites can be improved by 2-D laminate construction. Mechanical property results for (0 sub 2/90 sub 2)sub s and (+45/-45 sub 2) sub s laminates showed that their matrix failure strains were similar to that for (0) sub 8 laminates, but their primary elastic moduli, matrix cracking strengths, and ultimate composite strengths were lower. The elastic properties of unidirectional, cross-ply, and angle-ply composites can be predicted from modified constitutive equations and laminate theory. Further improvements in laminate properties may be achieved by reducing the matrix porosity and by optimizing the bond strength between the SiC fiber and RBSN matrix.

Rhatt, R. T.↗
NASA NTRS1990

Laminate behavior for SiC fiber-reinforced reaction-bonded silicon nitride matrix composites

The room temperature mechanical properties of SiC fiber reinforced reaction-bonded silicon nitride matrix composite laminates (SiC/RBSN) have been measured. The laminates contained approx 30 volume fraction of aligned 142-micron diameter SiC fiber in a porous RBSN matrix. Three types of laminate studied were unidirectional: (1) (0) sub 8, (2) (10) sub 8, and (3) (45) sub 8, and (90) sub 8; cross plied laminates (0 sub 2/90 sub 2); and angle plied laminates: (+45 sub 2/-45 sub 2). Each laminate contained eight fiber plies. Results of the unidirectionally reinforced composites tested at various angles to the reinforcement direction indicate large anisotropy in in-plane properties. In addition, strength properties of these composites along the fiber direction were independent of specimen gage length and were unaffected by notches normal to the fiber direction. Splitting parallel to the fiber at the notch tip appears to be the dominant crack blunting mechanism responsible for notch insensitive behavior of these composites. In-plane properties of the composites can be improved by 2-D laminate construction. Mechanical property results for (0 sub 2/90 sub 2) sub s and (+45/-45 sub 2) sub s laminates showed that their matrix failure strains were similar to that for (0) sub 8 laminates, but their primary elastic moduli, matrix cracking strengths, and ultimate composite strengths were lower. The elastic properties of unidirectional, cross-ply, and angle-ply composites can be predicted from modified constitutive equations and laminate theory. Further improvements in laminate properties may be achieved by reducing the matrix porosity and by optimizing the bond strength between the SiC fiber and RBSN matrix.

Bhatt, Ramakrishna T.↗
NASA NTRS1991

Method for calculating and observing microwave absorption by a sphere in a single mode rectangular cavity

A new theory of microwave absorption by a lossy dielectric sphere in a single mode rectangular cavity has been recently developed. The absorption was treated in the framework of an electromagnetic scattering problem. That theory is summarized here and calculated results that bear on optimizing the processing of materials are illustrated. Methods for observing power absorption and other results predicted by the scattering model are discussed. Cavity perturbation theory provides a bridge between theoretical calculations and experimental observations, and a special problem that arises when an established version of cavity perturbation theory is applied to spheres is identified, analyzed, and resolved. The direct problem of predicting shifts in frequency and Q from model calculations is discussed for a sphere in a cavity when the sphere's complex dielectric constant is known. Also, the inverse problem of determining the complex dielectric constants from measured values of those shifts is considered. The small sphere limit, where an electrostatic or quasistatic model is valid, is treated in detail, and planned work on parallel problems for larger spheres is described.

Jackson, H. W.↗
NASA NTRS1999

NCC: A Multidisciplinary Design/Analysis Tool for Combustion Systems

A multi-disciplinary design/analysis tool for combustion systems is critical for optimizing the low-emission, high-performance combustor design process. Based on discussions between NASA Lewis Research Center and the jet engine companies, an industry-government team was formed in early 1995 to develop the National Combustion Code (NCC), which is an integrated system of computer codes for the design and analysis of combustion systems. NCC has advanced features that address the need to meet designer's requirements such as "assured accuracy", "fast turnaround", and "acceptable cost". The NCC development team is comprised of Allison Engine Company (Allison), CFD Research Corporation (CFDRC), GE Aircraft Engines (GEAE), NASA Lewis Research Center (LeRC), and Pratt & Whitney (P&W). This development team operates under the guidance of the NCC steering committee. The "unstructured mesh" capability and "parallel computing" are fundamental features of NCC from its inception. The NCC system is composed of a set of "elements" which includes grid generator, main flow solver, turbulence module, turbulence and chemistry interaction module, chemistry module, spray module, radiation heat transfer module, data visualization module, and a post-processor for evaluating engine performance parameters. Each element may have contributions from several team members. Such a multi-source multi-element system needs to be integrated in a way that facilitates inter-module data communication, flexibility in module selection, and ease of integration.

Liu, Nan-Suey↗
NASA NTRS1996

The Structure and Stability of B(n)H(+) Clusters

The geometries of the B(n)H(+) clusters for n=1-13 have been optimized at the B3LYP level of theory. Excluding B2H(+) all clusters are low-spin coupled. The structures appear to be related to the bare B(+)(n) clusters in two ways: 1) an H atom is added to the B(+)(n) cluster or 2) a BH is added to the B(+)(n-1) cluster. The B(+)(n)-H binding energies are computed at the B3LYP level of theory, and calibrated n using the CCSD(T) level of theory. The computed results tend to fall between the experimental best estimates and the experimental lower bounds. The B(+)(n)-H binding energies show an inverse correspondence with the B+(n-1)-B values) while the B(+)(n-1)-B and B(+)(n-1)-BH results tend to parallel each other.

Ricca, Alessandra↗
DOE OSTI2025

Distributed Stochastic Optimization of a Neural Representation Network for Time-Space Tomography Reconstruction

4D time-space reconstruction of dynamic events or deforming objects using X-ray computed tomography (CT) is an important inverse problem in non-destructive evaluation. Conventional back-projection based reconstruction methods assume that the object remains static for the duration of several tens or hundreds of X-ray projection measurement images (reconstruction of consecutive limited-angle CT scans). However, this is an unrealistic assumption for many in-situ experiments that causes spurious artifacts and inaccurate morphological reconstructions of the object. To solve this problem, we propose to perform a 4D time-space reconstruction using a distributed implicit neural representation (DINR) network that is trained using a novel distributed stochastic training algorithm. Our DINR network learns to reconstruct the object at its output by iterative optimization of its network parameters such that the measured projection images best match the output of the CT forward measurement model. Here, we use a forward measurement model that is a function of the DINR outputs at a sparsely sampled set of continuous valued 4D object coordinates. Unlike previous neural representation architectures that forward and back propagate through dense voxel grids that sample the object's entire time-space coordinates, we only propagate through the DINR at a small subset of object coordinates in each iteration resulting in an order-of-magnitude reduction in memory and compute for training. DINR leverages distributed computation across several compute nodes and GPUs to produce high-fidelity 4D time-space reconstructions. We use both simulated parallel-beam and experimental cone-beam X-ray CT datasets to demonstrate the superior performance of our approach.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Porting Classical Approaches for Quantum Simulations to Quantum Computers

Simulating quantum many-body systems is one of the most promising problems in which we might anticipate that quantum computers should show quantum advantage. Unfortunately, there is still a gap between this promise and actual practice. New quantum algorithms need to be developed and the current quantum algorithms have various difficulties - e.g efficient state preparation - which must be overcome and improved upon. In many cases, classical approaches need to be ported over to quantum devices. In this project we have developed a suite of new quantum algorithms which makes progress in this regard. We developed a new optimization scheme for variational quantum eigensolvers, UBOS, which mitigates problems with local minimas and barren plateaus while improving convergence to the ground state by an order of magnitude. We developed a new way to utilize qubitization to find ground states of nearly frustration-free Hamiltonians faster than all previous methods. We developed a series of state preparation techniques which helps initialize parameterized quantum circuits into reasonable starting points on which quantum algorithms are then applied. In addition to the development of novel algorithms, it is critical to have classical simulation techniques for approximately simulating quantum circuits which can be used to benchmark and understand quantum algorithms. Toward that end, we developed a novel POVM formalism to simulate quantum circuits as well as exemplify the massive parallelization of tensor network methodologies. Finally, we developed physical understanding of entanglement phase transitions such as many-body localization and random tensor networks.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
NASA NTRS2011

Parallel Eclipse Project Checkout

Parallel Eclipse Project Checkout (PEPC) is a program written to leverage parallelism and to automate the checkout process of plug-ins created in Eclipse RCP (Rich Client Platform). Eclipse plug-ins can be aggregated in a feature project. This innovation digests a feature description (xml file) and automatically checks out all of the plug-ins listed in the feature. This resolves the issue of manually checking out each plug-in required to work on the project. To minimize the amount of time necessary to checkout the plug-ins, this program makes the plug-in checkouts parallel. After parsing the feature, a request to checkout for each plug-in in the feature has been inserted. These requests are handled by a thread pool with a configurable number of threads. By checking out the plug-ins in parallel, the checkout process is streamlined before getting started on the project. For instance, projects that took 30 minutes to checkout now take less than 5 minutes. The effect is especially clear on a Mac, which has a network monitor displaying the bandwidth use. When running the client from a developer s home, the checkout process now saturates the bandwidth in order to get all the plug-ins checked out as fast as possible. For comparison, a checkout process that ranged from 8-200 Kbps from a developer s home is now able to saturate a pipe of 1.3 Mbps, resulting in significantly faster checkouts. Eclipse IDE (integrated development environment) tries to build a project as soon as it is downloaded. As part of another optimization, this innovation programmatically tells Eclipse to stop building while checkouts are happening, which dramatically reduces lock contention and enables plug-ins to continue downloading until all of them finish. Furthermore, the software re-enables automatic building, and forces Eclipse to do a clean build once it finishes checking out all of the plug-ins. This software is fully generic and does not contain any NASA-specific code. It can be applied to any Eclipse-based repository with a similar structure. It also can apply build parameters and preferences automatically at the end of the checkout.

Crockett, Thomas M.↗
NASA NTRS2015

Multiscale and Multiphysics Modeling of Additive Manufacturing of Advanced Materials

The objective of this proposed project is to research and develop a prediction tool for advanced additive manufacturing (AAM) processes for advanced materials and develop experimental methods to provide fundamental properties and establish validation data. Aircraft structures and engines demand materials that are stronger, useable at much higher temperatures, provide less acoustic transmission, and enable more aeroelastic tailoring than those currently used. Significant improvements in properties can only be achieved by processing the materials under nonequilibrium conditions, such as AAM processes. AAM processes encompass a class of processes that use a focused heat source to create a melt pool on a substrate. Examples include Electron Beam Freeform Fabrication and Direct Metal Deposition. These types of additive processes enable fabrication of parts directly from CAD drawings. To achieve the desired material properties and geometries of the final structure, assessing the impact of process parameters and predicting optimized conditions with numerical modeling as an effective prediction tool is necessary. The targets for the processing are multiple and at different spatial scales, and the physical phenomena associated occur in multiphysics and multiscale. In this project, the research work has been developed to model AAM processes in a multiscale and multiphysics approach. A macroscale model was developed to investigate the residual stresses and distortion in AAM processes. A sequentially coupled, thermomechanical, finite element model was developed and validated experimentally. The results showed the temperature distribution, residual stress, and deformation within the formed deposits and substrates. A mesoscale model was developed to include heat transfer, phase change with mushy zone, incompressible free surface flow, solute redistribution, and surface tension. Because of excessive computing time needed, a parallel computing approach was also tested. In addition, after investigating various methods, a Smoothed Particle Hydrodynamics Model (SPH Model) was developed to model wire feeding process. Its computational efficiency and simple architecture makes it more robust and flexible than other models. More research on material properties may be needed to realistically model the AAM processes. A microscale model was developed to investigate heterogeneous nucleation, dendritic grain growth, epitaxial growth of columnar grains, columnar-to-equiaxed transition, grain transport in melt, and other properties. The orientations of the columnar grains were almost perpendicular to the laser motion's direction. Compared to the similar studies in the literature, the multiple grain morphology modeling result is in the same order of magnitude as optical morphologies in the experiment. Experimental work was conducted to validate different models. An infrared camera was incorporated as a process monitoring and validating tool to identify the solidus and mushy zones during deposition. The images were successfully processed to identify these regions. This research project has investigated multiscale and multiphysics of the complex AAM processes thus leading to advanced understanding of these processes. The project has also developed several modeling tools and experimental validation tools that will be very critical in the future of AAM process qualification and certification.

Liou, Frank↗
NASA NTRS2002

Design of a Dual Waveguide Normal Incidence Tube (DWNIT) Utilizing Energy and Modal Methods

This report investigates the partition design of the proposed Dual Waveguide Normal Incidence Tube (DWNIT). Some advantages provided by the DWNIT are (1) Assessment of coupling relationships between resonators in close proximity, (2) Evaluation of "smart liners", (3) Experimental validation for parallel element models, and (4) Investigation of effects of simulated angles of incidence of acoustic waves. Energy models of the two chambers were developed to determine the Sound Pressure Level (SPL) drop across the two chambers, through the use of an intensity transmission function for the chamber's partition. The models allowed the chamber's lengthwise end samples to vary. The initial partition design (2" high, 16" long, 0.25" thick) was predicted to provide at least 160 dB SPL drop across the partition with a compressive model, and at least 240 dB SPL drop with a bending model using a damping loss factor of 0.01. The end chamber sample transmissions coefficients were set to 0.1. Since these results predicted more SPL drop than required, a plate thickness optimization algorithm was developed. The results of the algorithm routine indicated that a plate with the same height and length, but with a thickness of 0.1" and 0.05 structural damping loss, would provide an adequate SPL isolation between the chambers.

Betts, Juan F.↗
NASA NTRS1992

Physiological asymmetry in etiolated pea epicotyls: relation to patterns of auxin distribution and phototropic behavior

Etiolated pea seedlings require transformation of Pr phytochrome to Pfr before they display optimal phototropic response to unilateral blue light. This study investigates the possible role of auxin transport in explaining these phenomena. Labeled [2-14C]IAA applied to the intact terminal buds of dark-grown and red light-treated pea seedlings was measured 210 min later on the shaded and illuminated sides of the epicotyl as a function of direction and duration of irradiation with blue light. Totally darkened epicotyls show an asymmetry in distribution of radioactivity in the upper growth zone of the epicotyl, in favor of the side under the concave part of the apical hook. Red light, which greatly potentiates curvature toward subsequent unilateral blue light, lowers this asymmetry. Blue light directed to the epicotyl of red-pretreated plants in a plane parallel to the hook and from the side bearing the convex portion of the hook induces positive phototropic curvature as well as a surplus of radioactivity on the illuminated side of the upper epicotyl and on the shaded side of the lower growth zone of the epicotyl. Light directed to the side bearing the concave part of the hook also causes an accumulation of counts in the upper part of the lighted side but produces neither curvature of the epicotyl nor accumulation of counts in the lower shaded side. Because of this built-in physiological asymmetry in the growth zone just below the apical hook, it is difficult to explain the effects of red and blue light on curvature in terms of patterns of auxin distribution alone.

NASA Program Space Biology↗
NASA NTRS2021

USA Crop Yield Estimation with MODIS NDVI: Are Remotely Sensed Models Better Than Simple Trend Analyses?

Crop yield forecasting is performed monthly during the growing season by the United States Department of Agriculture’s National Agricultural Statistics Service. The underpinnings are long-established probability surveys reliant on farmers’ feedback in parallel with biophysical measurements. Over the last decade though, satellite imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS) has been used to corroborate the survey information. This is facilitated through the Global Inventory Modeling and Mapping Studies/Global Agricultural Monitoring system, which provides open access to pertinent real-time normalized difference vegetation index (NDVI) data. Hence, two relatively straightforward MODIS-based modeling methods are employed operationally. The first model constitutes mid-season timing based on the maximum peak NDVI value, while the second is reflective of late-season timing by integrating accumulated NDVI over a threshold value. Corn model results nationally show the peak NDVI method provides a R^(2) of 0.88 and a coefficient of variation (CV) of 3.5%. The accumulated method, using an optimally derived 0.58 NDVI threshold, improves the performance to 0.93 and 2.7%, respectively. Both these models outperform simple trend analysis, which is 0.48 and 7.4%, correspondingly. For soybeans the R^(2) results of the peak NDVI model are 0.62, and 0.73 for the accumulated using a 0.56 threshold. CVs are 6.8% and 5.7%, respectively. Spring wheat’s R2performance with the accumulated NDVI model is 0.60 but just 0.40 with peak NDVI. The soybean and spring wheat models perform similarly to trend analysis. Winter wheat and upland cotton show poor model performance, regardless of method. Ultimately, corn yield forecasting derived from MODIS imagery is robust, and there are circumstances when forecasts for soybeans and spring wheat have merit too.

crop yield↗
DOE OSTI2020

TeraChem: A graphical processing unit-accelerated electronic structure package for large-scale ab initio molecular dynamics

TeraChem was born in 2008 with the goal of providing fast on-the-fly electronic structure calculations to facilitate ab initio molecular dynamics studies of large biochemical systems such as photoswitchable proteins and multichromophoric antenna complexes. Originally developed for videogaming applications, graphics processing units (GPUs) offered a low-cost parallel computer architecture that became more accessible for general-purpose GPU computing with the release of CUDA in 2007. The evaluation of the electron repulsion integrals (ERIs) is a major bottleneck in electronic structure codes and provides an attractive target for acceleration on GPUs. Thus, highly efficient routines for evaluation of and contractions between the ERIs and density matrices were implemented in TeraChem. Here, electronic structure methods were developed and implemented to leverage these integral contraction routines, resulting in the first quantum chemistry package designed from the ground up for GPUs. This GPU acceleration makes TeraChem capable of performing large-scale ground and excited state calculations in the gas and condensed phase. Today, TeraChem's speed forms the basis for a suite of quantum chemistry applications, including optimization and dynamics of proteins, automated and interactive chemical discovery tools, and large-scale nonadiabatic dynamics simulations.

74 ATOMIC AND MOLECULAR PHYSICS↗