Search NASA⌕ Search

SEARCH · Search NASA

Results for “structured data”

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 613 records · Page 34

Evidence for a Single Holocene Paleoseismic Event on the Pajarito Fault, Northern New Mexico

Low-slip rate fault systems tend to be less studied than their high-slip rate counterparts, and paleoseismic techniques used to study them may pose challenges in interpretation that differ from high-slip rate systems. A good example of this is the Pajarito fault system (PFS), a normal fault complex within the Rio Grande rift. Despite numerous previous paleoseismic trenching studies conducted on the PFS between 1990 and 2003, considerable uncertainty remains regarding its Holocene paleoseismic history, particularly for the primary Pajarito fault (PF). To further clarify the PF paleoseismic history, we present data from paleoseismic investigations of 6 trenches at 3 distinct locations along the PF. Though the totality of the age and structural data obtained in this study is complex and not entirely consistent with any one interpretation, a single Holocene paleoearthquake occurring younger than ∼1,600 to 2,300 kcal yr BP is the simplest interpretation. It is possible that the PF records two Holocene events, with a penultimate event 6.9–2.4 kcal yr BP event and the aforementioned most recent event (MRE) between 2.3 and 1.6 kcal yr BP. However, only a single wall of one trench, out of a total of 12 walls in our 6 trenches, provides evidence supporting that interpretation. This study finds evidence of a single late Holocene paleoseismic event on the PF and sparse evidence for 2 Holocene paleoseismic events on the PF and highlights the benefits of logging multiple trench walls to better understand the complexity that results from this low-slip rate, low-deposition-rate fault system.

58 GEOSCIENCES↗

SUSPECT: MINLP special structure detector for Pyomo

We present SUSPECT, an open source toolkit that symbolically analyzes mixed-integer nonlinear optimization problems formulated using the Python algebraic modeling library Pyomo. We present the data structures and algorithms used to implement SUSPECT. SUSPECT works on a directed acyclic graph representation of the optimization problem to perform: bounds tightening, bound propagation, monotonicity detection, and convexity detection. We show how the tree-walking rules in SUSPECT balance the need for lightweight computation with effective special structure detection. SUSPECT can be used as a standalone tool or as a Python library to be integrated in other tools or solvers. Here, we highlight the easy extensibility of SUSPECT with several recent convexity detection tricks from the literature. We also report experimental results on the MINLPLib 2 dataset.

97 MATHEMATICS AND COMPUTING↗

Refinement of macromolecular structures against neutron data with SHELXL2013

Some of the improvements in SHELX2013 make SHELXL convenient to use for refinement of macromolecular structures against neutron data without the support of X-ray data. The new NEUT instruction adjusts the behaviour of the SFAC instruction as well as the default bond lengths of the AFIX instructions. This work presents a protocol on how to use SHELXL for refinement of protein structures against neutron data. It includes restraints extending the Engh & Huber [Acta Cryst. (1991), A47, 392-400] restraints to H atoms and discusses several of the features of SHELXL that make the program particularly useful for the investigation of H atoms with neutron diffraction. SHELXL2013 is already adequate for the refinement of small molecules against neutron data, but there is still room for improvement, like the introduction of chain IDs for the refinement of macromolecular structures.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Insights into Substrate Specificity and Metal Activation of Mammalian Tetrahedral Aspartyl Aminopeptidase

Aminopeptidases are key enzymes involved in the regulation of signaling peptide activity. Here, we present a detailed biochemical and structural analysis of an evolutionary highly conserved aspartyl aminopeptidase called DNPEP. We show that this peptidase can cleave multiple physiologically relevant substrates, including angiotensins, and thus may play a key role in regulating neuron function. Using a combination of x-ray crystallography, x-ray absorption spectroscopy, and single particle electron microscopy analysis, we provide the first detailed structural analysis of DNPEP. We show that this enzyme possesses a binuclear zinc-active site in which one of the zinc ions is readily exchangeable with other divalent cations such as manganese, which strongly stimulates the enzymatic activity of the protein. The plasticity of this metal-binding site suggests a mechanism for regulation of DNPEP activity. We also demonstrate that DNPEP assembles into a functionally relevant tetrahedral complex that restricts access of peptide substrates to the active site. These structural data allow rationalization of the enzyme's preference for short peptide substrates with N-terminal acidic residues. This study provides a structural basis for understanding the physiology and bioinorganic chemistry of DNPEP and other M18 family aminopeptidases.

59 BASIC BIOLOGICAL SCIENCES↗

X-ray crystallography–based structural elucidation of enzyme-bound intermediates along the 1-deoxy-d-xylulose 5-phosphate synthase reaction coordinate

1-Deoxy-D-xylulose 5-phosphate synthase (DXPS) uses thiamine diphosphate (ThDP) to convert pyruvate and D-glyceraldehyde 3-phosphate ( D -GAP) into 1-deoxy-D-xylulose 5-phosphate (DXP), an essential bacterial metabolite. DXP is not utilized by humans; hence, DXPS has been an attractive antibacterial target. In this work, we investigate DXPS from Deinococcus radioduransN (DrDXPS), showing that it has similar kinetic parameters K m D-GAP AND K m pyruvate (54 ± 3 and 11 ± 1 μM, respectively) and comparable catalytic activity (kcat = 45 ± 2 min –1 ) with previously studied bacterial DXPS enzymes and employing it to obtain missing structural data on this enzyme family. In particular, we have determined crystallographic snapshots of DrDXPS in two states along the reaction coordinate: a structure of DrDXPS bound to C2α-phosphonolactylThDP (PLThDP), mimicking the native pre-decarboxylation intermediate C2α-lactylThDP (LThDP), and a native post-decarboxylation state with a bound enamine intermediate. The 1.94-Å-resolution structure of PLThDP-bound DrDXPS delineates how two active-site histidine residues stabilize the LThDP intermediate. Meanwhile, the 2.40-Å-resolution structure of an enamine intermediate-bound DrDXPS reveals how a previously unknown 17-Å conformational change removes one of the two histidine residues from the active site, likely triggering LThDP decarboxylation to form the enamine intermediate. These results provide insight into how the bi-substrate enzyme DXPS limits side reactions by arresting the reaction on the less reactive LThDP intermediate when its cosubstrate is absent. They also offer a molecular basis for previous low-resolution experimental observations that correlate decarboxylation of LThDP with protein conformational changes.

1-deoxy-D-xylulose 5-phosphate synthase (DXPS)↗

TEQUILA: a platform for rapid development of quantum algorithms

Variational quantum algorithms are currently the most promising class of algorithms for deployment on near-term quantum computers. In contrast to classical algorithms, there are almost no standardized methods in quantum algorithmic development yet, and the field continues to evolve rapidly. As in classical computing, heuristics play a crucial role in the development of new quantum algorithms, resulting in a high demand for flexible and reliable ways to implement, test, and share new ideas. In this paper, inspired by this demand, we introduce TEQUILA, a development package for quantum algorithms in PYTHON, designed for fast and flexible implementation, prototyping and deployment of novel quantum algorithms in electronic structure and other fields. TEQUILA operates with abstract expectation values which can be combined, transformed, differentiated, and optimized. On evaluation, the abstract data structures are compiled to run on state of the art quantum simulators or interfaces.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Towards improved and more routine Earth system model evaluation in CMIP

The Coupled Model Intercomparison Project (CMIP) has successfully provided the climate community with a rich collection of simulation output from Earth system models (ESMs) that can be used to understand past climate changes and make projections and uncertainty estimates of the future. Confidence in ESMs can be gained because the models are based on physical principles and reproduce many important aspects of observed climate. More research is required to identify the processes that are most responsible for systematic biases and the magnitude and uncertainty of future projections so that more relevant performance tests can be developed. At the same time, there are many aspects of ESM evaluation that are well established and considered an essential part of systematic evaluation but have been implemented ad hoc with little community coordination. Given the diversity and complexity of ESM analysis, we argue that the CMIP community has reached a critical juncture at which many baseline aspects of model evaluation need to be performed much more efficiently and consistently. We provide a perspective and viewpoint on how a more systematic, open, and rapid performance assessment of the large and diverse number of models that will participate in current and future phases of CMIP can be achieved, and announce our intention to implement such a system for CMIP6. Accomplishing this could also free up valuable resources as many scientists are frequently "re-inventing the wheel" by re-writing analysis routines for well-established analysis methods. A more systematic approach for the community would be to develop and apply evaluation tools that are based on the latest scientific knowledge and observational reference, are well suited for routine use, and provide a wide range of diagnostics and performance metrics that comprehensively characterize model behaviour as soon as the output is published to the Earth System Grid Federation (ESGF). The CMIP infrastructure enforces data standards and conventions for model output and documentation accessible via the ESGF, additionally publishing observations (obs4MIPs) and reanalyses (ana4MIPs) for model intercomparison projects using the same data structure and organization as the ESM output. This largely facilitates routine evaluation of the ESMs, but to be able to process the data automatically alongside the ESGF, the infrastructure needs to be extended with processing capabilities at the ESGF data nodes where the evaluation tools can be executed on a routine basis. Efforts are already underway to develop community-based evaluation tools, and we encourage experts to provide additional diagnostic codes that would enhance this capability for CMIP. And, at the same time, we encourage the community to contribute observations and reanalyses for model evaluation to the obs4MIPs and ana4MIPs archives. The intention is to produce through the ESGF a widely accepted quasi-operational evaluation framework for CMIP6 that would routinely execute a series of standardized evaluation tasks. Over time, as this capability matures, we expect to produce an increasingly systematic characterization of models which, compared with early phases of CMIP, will more quickly and openly identify the strengths and weaknesses of the simulations. This will also reveal whether long-standing model errors remain evident in newer models and will assist modelling groups in improving their models. Finally, this framework will be designed to readily incorporate updates, including new observations and additional diagnostics and metrics as they become available from the research community.

54 ENVIRONMENTAL SCIENCES↗

A Comparison of Shared Memory Parallel Programming Models

The dominant parallel programming models for shared memory computers, Pthreads and OpenMP, are both thread-centric in that they are based on explicit management of tasks and enforce data dependencies and output ordering through task management. By comparison, the Cray XMT programming model is data-centric where the primary concern of the programmer is managing data dependencies, allowing threads to progress in a data flow fashion. The XMT implements this programming model by associating tag bits with each word of memory, affording efficient fine grained data synchronization independent of the number of processors or how tasks are scheduled. When task management is implicit and synchronization is abundant, efficient, and easy to use, programmers have viable alternatives to traditional thread-centric algorithms. In this paper we compare the amount of available parallelism relative to the amount of work in a variety of different algorithms and data structures when synchronization does not need to be rationed, as well as identify opportunities for platform and performance portability of the data-centric programming model on multi-core processors.

99 GENERAL AND MISCELLANEOUS↗

Generative Thermodynamic Computing

Here, we introduce a generative modeling framework for thermodynamic computing, in which structured data are synthesized from noise by the natural time evolution of a physical system governed by Langevin dynamics. While conventional diffusion models use neural networks to perform denoising, here the information needed to generate structure from noise is encoded by the dynamics of a thermodynamic system. Training proceeds by maximizing the probability with which the computer generates the reverse of a noising trajectory, which ensures that the computer generates data with minimal heat emission. We demonstrate this framework within a digital simulation of a thermodynamic computer. If realized in analog hardware, such a system would function as a generative model that produces structured samples without the need for artificially injected noise or active control of denoising.

Whitelam, Stephen [Lawrence Berkeley National Labo↗

Clustering of nonvalently bonded NO{sub 2}…O{sub 2}N fragments at C(sp{sup 3}) atoms

Based on the 'invariom' approach and the 'Atoms in Molecules' theory, a simplified method is proposed to reveal and analyze bonding interatomic interactions using large arrays of structural data. The technique was tested taking nonvalent interactions between nitro groups bonded to C(sp{sup 3}) atoms as examples. Geometric clustering of structural fragments was performed and the interaction strength and covalent contribution distributions were analyzed. Relationships between the fragment clustering results and nonvalent bonding pattern features are discussed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Influence of chemical substitution on the photoluminescence of Sr{sub (1−x)}Pb{sub x}WO{sub 4} solid solution

The solid solution Sr{sub 1−x}Pb{sub x}WO{sub 4} based on luminescent tungstates SrWO{sub 4} and PbWO{sub 4} has been synthesized by solid-state reaction for all compositions 0≤x≤1. Using Rietveld method, the structural data of all polycrystalline samples have been refined and crystal cell parameters exhibited a linear behavior as a function of x. All substituted structures are of scheelite type. Scanning electron microscopy showed that a high level of crystallization characterized the samples, with modifications in sizes and shapes depending on composition x. Infrared and Raman spectroscopy have been performed to characterize the evolution of vibrational modes with substitution rate. Finally, a systematic study of luminescence under X-ray excitation has been performed: in the composition range x=0.2 to 0.4, intensities of emission exhibited increased values. The luminescence profiles have been interpreted in terms of four Gaussian components, two of them depending on substitution rate. - Graphical abstract: Photoluminescence under X-ray excitation of Sr{sub 1−x}Pb{sub x}WO{sub 4} solid solution: (left) decomposition of emission large band into four components for composition x=0.3; (right) variation of total emission intensity with composition x. - Highlights: • Structural study of solid solution Sr{sub 1−x}Pb{sub x} WO{sub 4} using Rietveld analyses. • Variation of Debye–Waller factor with composition x, disorder parameter. • Existence of four components of luminescence under X-ray excitation. • Effect of substitution of Sr by Pb on two components, due to Pb–O–W interactions. • Enhancement of luminescence intensity in a specific composition range.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Structure-Based Alteration of Substrate Specificity and Catalytic Activity of Sulfite Oxidase from Sulfite Oxidation to Nitrate Reduction

Eukaryotic sulfite oxidase is a dimeric protein that contains the molybdenum cofactor and catalyzes the metabolically essential conversion of sulfite to sulfate as the terminal step in the metabolism of cysteine and methionine. Nitrate reductase is an evolutionarily related molybdoprotein in lower organisms that is essential for growth on nitrate. In this study, we describe human and chicken sulfite oxidase variants in which the active site has been modified to alter substrate specificity and activity from sulfite oxidation to nitrate reduction. On the basis of sequence alignments and the known crystal structure of chicken sulfite oxidase, two residues are conserved in nitrate reductases that align with residues in the active site of sulfite oxidase. On the basis of the crystal structure of yeast nitrate reductase, both positions were mutated in human sulfite oxidase and chicken sulfite oxidase. The resulting double-mutant variants demonstrated a marked decrease in sulfite oxidase activity but gained nitrate reductase activity. An additional methionine residue in the active site was proposed to be important in nitrate catalysis, and therefore, the triple variant was also produced. The nitrate reducing ability of the human sulfite oxidase triple mutant was nearly 3-fold greater than that of the double mutant. To obtain detailed structural data for the active site of these variants, we introduced the analogous mutations into chicken sulfite oxidase to perform crystallographic analysis. The crystal structures of the Mo domains of the double and triple mutants were determined to 2.4 and 2.1 {angstrom} resolution, respectively.

59 BASIC BIOLOGICAL SCIENCES↗

Structural study of hcp and liquid iron under shock compression up to 275 GPa

Here we combine nanosecond laser shock compression with in situ picosecond x-ray diffraction to provide structural data on iron up to 275 GPa. We constrain the extent of hcp-liquid coexistence, the onset of total melt, and the structure within the liquid phase. Our results indicate that iron, under shock compression, melts completely by 258(8) GPa. A coordination number analysis indicates that iron is a simple liquid at these pressure-temperature conditions. We also perform texture analysis between the ambient body-centered-cubic (bcc) α, and the hexagonal-closed-packed (hcp) high-pressure ε-phase. We rule out the Rong-Dunlop orientation relationship (OR) between the α and ε - phase s. However, we cannot distinguish between three other closely related ORs: Burger's, Mao-Bassett-Takahashi, and Potter's OR. The solid-liquid coexistence region is constrained from a melt onset pressure of 225(3) GPa from previously published sound speed measurements and full melt [246.5(1.8)–258(8) GPa] from x-ray diffraction measurements, with an associated maximum latent heat of melting of 623 J/g. This value is lower than recently reported theoretical estimates and suggests that the contribution to the earth's geodynamo energy budget from heat release due to freezing of the inner core is smaller than previously thought. Melt pressures for these nanosecond shock experiments are consistent with gas gun shock experiments that last for microseconds, indicating that the melt transition occurs rapidly.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Dark progression reveals slow timescales for radiation damage between T = 180 and 240 K

Between T = 180 and 240 K, radiation damage progresses on minute timescales when the X-rays are off, suggesting that a fraction of damage at higher temperatures may be outrun using currently available sources and detectors. Can radiation damage to protein crystals be ‘outrun’ by collecting a structural data set before damage is manifested? Recent experiments using ultra-intense pulses from a free-electron laser show that the answer is yes. Here, evidence is presented that significant reductions in global damage at temperatures above 200 K may be possible using conventional X-ray sources and current or soon-to-be available detectors. Specifically, ‘dark progression’ (an increase in damage with time after the X-rays have been turned off) was observed at temperatures between 180 and 240 K and on timescales from 200 to 1200 s. This allowed estimation of the temperature-dependent timescale for damage. The rate of dark progression is consistent with an Arrhenius law with an activation energy of 14 kJ mol{sup −1}. This is comparable to the activation energy for the solvent-coupled diffusive damage processes responsible for the rapid increase in radiation sensitivity as crystals are warmed above the glass transition near 200 K. Analysis suggests that at T = 300 K data-collection times of the order of 1 s (and longer at lower temperatures) may allow significant reductions in global radiation damage, facilitating structure solution on crystals with liquid solvent. No dark progression was observed below T = 180 K, indicating that no important damage process is slowed through this timescale window in this temperature range.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Custom Accessors: Enabling Scalable Data Ingestion, (Re-)Organization, and Analysis on Distributed Systems

The emerging class of high velocity and high volume data analytic workflows comprise interwoven data ingestion, organization, and processing stages, with ingestion and organization steps often contributing comparable or even higher computational costs than actual processing steps. Since complex workflows consist of a variety of phases that view and use data differently, being able to construct efficient, scalable, distributed data structures (arrays, vectors, sets, maps, and multi-maps) is essential and requires custom methods to extend and shrink containers, analyze and position data, and, maintain globallyconsistent meta-data. In this paper, we propose a novel datastructure access paradigm based on the concept of Accessors. At a high level, accessors are customizable callable objects that can modify the behavior of insert, read, update, and delete operations for distributed containers while preserving atomicity guarantees. Accessors provide a very clean and natural way to implement a variety of programming patterns, e.g., conditional insertion/deletion and cascading computations, which would be otherwise hard (or even impossible) to express in parallel and distributed settings without using locks. We demonstrate the practicality and usefulness of our approach with two representative use cases and study the performance of these applications on a distributed High-Performance Computing system. Our analysis highlights that our proposed abstraction allows for an effective overlapping and concurrent execution of different workflow steps (e.g., data ingestion and analysis), which in a conventional analytics pipeline would execute sequentially, contributing cumulatively to the overall latency.

Castellana, Vito G. [BATTELLE (PACIFIC NW LAB)] (O↗

GROWdb US River Systems - Samples

GROW Overview We developed the Genome Resolved Open Watersheds database (GROWdb), which aims to increase genomic sampling and understanding of global river microbiomes. An emphasis of GROWdb is to create a publicly available and ever-expanding microbial genome database that is focused on rivers while being interoperable with databases from other ecosystems. GROWdb is based on a network-of-networks approach to move beyond a small collection of well-studied rivers, towards a spatially distributed, global network of systematic observations. GROWdb represents the first microbial, river-focused resource parsed at various scales from genes to MAGs to community level including expression and potential based measurements that will be of interest to microbiologists, ecologists, geochemists, hydrologists, and modelers. Dataset Acknowledgement GROWdb contains data from various research campaigns, please acknowledge the following data generators, as appropriate: WHONDRS derived genomes or samples - include this statement in your acknowledgements: “This study used data from the Worldwide Hydrobiogeochemistry Observation Network for Dynamic River Systems (WHONDRS) under the River Corridor Science Focus Area (SFA) at the Pacific Northwest National Laboratory (PNNL) that was generated at the U.S. Department of Energy (DOE) Joint Genome Institute User Facility. PNNL is operated by Battelle Memorial Institute for the U.S. DOE under Contract No. DE-AC05-76RL01830. The SFA is supported by the U.S. DOE, Office of Biological and Environmental Research (BER), Environmental System Science (ESS) Program.” Total Samples loaded onto this Narrative: 178 Note: Not all GROW samples may be loaded into KBase Data Availability The data underlying GROWdb are accessible across various platforms to ensure all levels of data structure are widely available. First, all reads and MAGs are publicly hosted on National Center for Biotechnology (NCBI) under Bioproject PRJNA946291. Second, all data related data presented here including MAG annotations, extended data tables, phylogenetic tree files, antibiotic resistance gene database files, and MAG abundance tables are available in Zenodo (link). Beyond the flat database files listed above, our aim for GROWdb was to maximize data use by making the data available in searchable and interactive platforms including the National Microbiome Data Collaborative (NMDC) data portal, the Department of Energy’s Systems Biology Knowledgebase (KBase), and a GROW specific user interface released here, GROWdb Explorer. Each platform provides different ways to interact with GROWdb: NMDC GROWdb formed a pilot project for the NMDC. Specifically, individual GROWdb datasets (metagenomes, metatranscriptomes, etc) are easily accessible and searchable through the NMDC data portal, where they are systematically connected to each other and to a rich suite of sample information and standard analysis results, following Findable, Accessible, Interoperable, and Reusable (FAIR) data practices. KBase GROWdb is publicly available within KBase, including samples (this Narrative), MAGs, and corresponding genome scale metabolic models. Access within KBase allows for immediate access and reuse of data, including comparison to private data using KBase’s 500+ analysis tools. Other linked narratives in KBase: GROW Metagenome Assembled Genomes (MAGs) GROW Metabolic Models GROWdb Explorer GROWdb data is also explorable through a graphical user interface built through the Colorado State University Geospatial Centroid (https://geocentroid.shinyapps.io/GROWdatabase/), allowing users to search and graph microbial and spatial data simultaneously. In summary, this microbial genome resource represents the first publicly available genome collection from rivers and offers data that can be leveraged across microbiome studies. GROWdb is an expanding repository to incorporate and unify global river multi-omic data for the future.

59 BASIC BIOLOGICAL SCIENCES↗

Exploring Integer Sum Reduction using Atomics on Intel CPU

Atomic functions are useful in updating a shared variable by multiple threads, barrier synchronizations, constructing complex data structures, and building high-level frameworks. In this paper, we focus on the evaluation and analysis of integer sum reduction, a common data parallel primitive. We convert the sequential reduction into parallel OpenCL implementations on the CPU. We also develop three micro kernels, which allow us to understand the relationships between the kernel performance and the operations involved in reduction. The results of the micro kernels show that increasing the work-group size linearly can linearly improve the kernel performance. There is a sweet spot in the relationship between the work-group size and barrier synchronization overhead. The performance of the atomics over local memory are not sensitive to the work-group size. The sum reduction kernel with vectorized memory accesses can improve the performance of the baseline kernel for a wide range of work-group sizes. However, the vectorization efficiency shrinks with the growing work-group size. We also find that the vendor’s default OpenCL kernel optimization does not improve the kernel performance. On average, disabling the optimization can reduce the execution time of the kernel with vectorized memory accesses by 15%. We attribute the performance drop to the fact that the default kernel optimizations instantiate a large number of atomics over global memory when implicitly vectorizing the kernel computation.

atomics↗

PSelInv—A Distributed Memory Parallel Algorithm for Selected Inversion: The Symmetric Case

We describe an efficient parallel implementation of the selected inversion algorithm for distributed memory computer systems, which we call PSelInv. The PSelInv method computes selected elements of a general sparse matrix Athat can be decomposed as A= LU, where L is lower triangular and U is upper triangular. The implementation described in this article focuses on the case of sparse symmetric matrices. It contains an interface that is compatible with the distributed memory parallel sparse direct factorization SuperLU-DIST. Yet, the underlying data structure and design of PSelInv allows it to be easily combined with other factorization routines, such as PARDISO. We discuss general parallelization strategies such as data and task distribution schemes. Specifically, we describe how to exploit the concurrency exposed by the elimination tree associated with the LU factorization of A. We demonstrate the efficiency and accuracy of PSelInv by presenting several numerical experiments. In particular, we show that PSelInv can run efficiently on more than 4,000 cores for a modestly sized matrix. We also show how PSelInv can be used to accelerate large-scale electronic structure calculations.

97 MATHEMATICS AND COMPUTING↗