Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance bound”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Understanding Strong Scaling on GPUs Using Empirical Performance Saturation Size

The roofline model provides a concise overview of the maximum performance capabilities of a given computer system through a combination of peak memory bandwidth and compute performance rates. The increasing complexity of scheduling and cache in recent GPUs, however, has introduced complicated performance variability that is not captured by arithmetic intensity alone. This work examines the effect of problem size and GPU launch configurations on roofline performance for V100, A100, MI100, and MI250X graphics processing units. We introduce an extended roofline model that takes problem size into account, and find that strong scaling on GPUs can be characterized by saturation problem sizes as additional key metrics. Saturation problem sizes break up a plot of GPU performance vs. problem size into three distinct performance regimes– size-limited, cache-bound, and DRAM-bound. With our extended roofline model, we are able to provide a robust view of these performance regimes across recent GPU architectures.

Eberius, David↗
DOE OSTI2020

Pyrazine-Fused Porous Graphitic Framework-Based Mixed Matrix Membranes for Enhanced Gas Separations

Membrane-based separations can mitigate the capital- and energy-intensive challenges associated with traditional thermally driven processes. To further push the boundary of gas separations, mixed matrix membranes (MMMs) have been extensively exploited; however, identifying an optimal nanofiller to boost the separation performance of MMMs beyond Robeson permeability-selectivity upper bounds remains an ongoing challenge. In this study, a new class of MMMs based on pyrazine-fused crystalline porous graphitic frameworks (PGFs) is reported. At a loading of 6 wt % PGFs, the MMMs surpass the current H 2 /CH 4 Robeson upper bound, ideally suited for applications such as H 2 regeneration. In addition, the fabricated MMMs exhibit appealing CO 2 separation performance, closely approaching the current Robeson upper bounds for CO 2 separation. Compared with the pristine polymeric membranes, the PGF-based MMMs display a record-high enhancement of gas permeability over 120% while maintaining intrinsic gas selectivities. Highlighting the crucial role of the crystallinity of nanofillers, this study demonstrates a facile and effective approach in formulating high-performance MMMs, complementing state-of-the-art membrane formation processes. The design principles open the door to energy-efficient separations of gas mixtures with enhanced productivity compatible with the current membrane manufacturing.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Tunable Catalytic Performance of Palladium Nanoparticles for H 2 O 2 Direct Synthesis via Surface-Bound Ligands

There is a critical need for sustainable routes to produce hydrogen peroxide, H 2 O 2 . A promising approach involves direct synthesis from molecular hydrogen and oxygen at (sub)ambient temperatures using unmodified supported Pd catalysts, which are limited by low selectivities. Controlling the environment of Pd active sites via surface ligands is shown to enhance selectivity. Trends among a myriad of surface ligands (i.e., phosphines, thiols, weakly bound molecules) suggest that those containing H-bonding groups lead to the highest H 2 O 2 production, potentially by affecting reaction energetics via H-bonding with key intermediates. These insights lay the groundwork for ligand design to achieve the optimal catalyst performance for H 2 O 2 synthesis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Benchmark CO 2 separation achieved by highly fluorinated nanoporous molecular sieve membranes from nonporous precursor via in situ cross-linking

Molecular sieve membranes with rigid micropores and CO 2 -philic functionalities within the architectures are promising candidates in CO 2 separation. However, the development of ctieve approaches for their fabrication still remains a significant challenge. In this study, an in situ cross-linking strategy is developed for the preparation of nanoporous fluorinated moleffecular sieve membranes using commercially available dense and non-porous polystyrene (M PS ) as a precursor template. Based on the dehydrative Friedel-Crafts reactions with highly fluorinated benzylic alcohols, M PS membranes are cross-linked in situ upon exposure to Brønsted acid (CF 3 SO 3 H), affording fluorinated microporous polymeric membranes with surface areas up to 523 m 2 g –1 and the presence of micropores centered at 1.1–1.3 nm as well as ultra-micropores (~0.6 nm). The obtained modified membranes exhibit good ideal CO 2 permeability of 797 barrer and CO 2 /N 2 selectivity of 28.5. In addition, high fluorine content (up to 28.5 wt%) and good thermal stability made the cross-linked membranes promising candidates to produce fluorinated carbon molecular sieve membranes with improved textural properties, exhibiting surface areas up to 1020 m 2 g –1 and ultra-micropores of ~0.4 nm. These membranes achieve superior CO 2 /N 2 separation performances exceeding the Robeson upper bound limit (2008).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Concordant Mode Approach (CMA): Vibrational Analysis of New and Upgraded Intermolecular Benchmarks for Noncovalent Bonding

The Concordant Mode Approach (CMA) is a novel method that offers tremendous potential for increasing the system size and the level of theory attainable in quantum chemical computations of molecular vibrational frequencies. To investigate the extension of CMA to intermolecular vibrations, computations with coupled cluster singles and doubles with perturbative triples theory [CCSD(T)] using two augmented correlation-consistent polarized-valence triple-ζ basis sets (aug-cc-pVTZ or h-aug-ccpVTZ) were performed on 17 prototypical loosely bound complexes of hydrogen-bonded, dispersion, and mixed character. These Level A results provide new and upgraded benchmarks for noncovalent bonding and a severe test for CMA vibrational analyses. The Level A target frequencies were recovered remarkably well using second-order Møller−Plesset perturbation theory (MP2) with h-aug-cc-pVTZ for generating the underlying (Level B) normal modes of the CMA scheme. Employing this Level B within the lowest-rung CMA-0A method reproduces the 435 benchmark frequencies with a mean absolute error (MAE) of 0.23 cm −1 and a corresponding standard deviation (σ) of 0.84 cm −1 ; strikingly, the corresponding subset of 106 interfragment frequencies exhibits MAE = 0.34 cm −1 and σ = 0.90 cm −1 . Subsequent application of the higher-rung CMA-2A scheme eliminates all outliers and reduces the overall MAE to a minuscule 0.08 cm −1 with the inclusion of only 3.0% of the off-diagonal couplings not accounted for by CMA-0A. Accordingly, the highly efficient CMA methodology proves to be robust even for vibrations on flat potential energy surfaces.

Aromatic compounds↗
DOE OSTI2024

Regioisomeric Spirobifluorene CANAL Ladder Polymers and Their Gas Separation Performance

For this work, we synthesized and characterized two isomeric microporous hydrocarbon ladder polymers from catalytic arene norbornene annulation (CANAL) of regioisomeric bis-norbornene fused spirobifluorenes, where the ladder chains are connected either through the same fluorene unit or across two different fluorene units in spirobifluorene. This pair of isomeric polymers was used to investigate the effect of ladder macromolecular structures on the microporosity and transport properties. Both polymers form mechanically intact films with thermal stability up to 480 °C and relatively high BET surface areas. The polymer formed from 2,7-dibromospirobifluorene showed higher BET surface area and higher gas permeability than the polymer from 2,2'-dibromospirobifluorene despite similar intersegmental spacing as indicated by X-ray scattering. The aging behavior for both polymers followed the same trend as the previously reported CANAL-fluorene polymers, with dramatically increased permselectivities over time, resulting in gas separation performance above the 2008 upper bounds for H 2 /CH 4 and O 2 /N 2 .

36 MATERIALS SCIENCE↗
DOE OSTI2020

Destructive Error Interference in Product-Formula Lattice Simulation

Quantum computers can efficiently simulate the dynamics of quantum systems. Here, we study the cost of digitally simulating the dynamics of several physically relevant systems using the first-order product-formula algorithm. We show that the errors from different Trotterization steps in the algorithm can interfere destructively, yielding a much smaller error than previously estimated. In particular, we prove that the total error in simulating a nearest-neighbor interacting system of n sites for time t using the first-order product formula with r time slices is O ( n t / r + n t 3 / r 2 ) when n t 2 / r is less than a small constant. Given an error tolerance ϵ , the error bound yields an estimate of max { O ( n 2 t / ϵ ) , O ( n 2 t 3 / 2 / ϵ 1 / 2 ) } for the total gate count of the simulation. The estimate is tighter than previous bounds and matches the empirical performance observed in Childs et al. [ Proc. Natl. Acad. Sci. U.S.A. 115 , 9456 (2018) ]. We also provide numerical evidence for potential improvements and conjecture an even tighter estimate for the gate count.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Revisiting Theoretical Limits for One Degree-of-Freedom Wave Energy Converters

This work revisits the theoretical limits of one degree-of-freedom wave energy converters (WECs). This work considers the floating sphere used in the OES Task 10 WEC modeling and verification effort for analysis. Analytical equations are derived to determine bounds on displacement amplitude, time-averaged power (TAP), and power-take-off (PTO) force. A unique result found shows that the TAP absorbed by a WEC can be defined solely by the inertial properties and radiation hydrodynamic coefficients. In addition, a unique expression for the PTO force was derived that provides lower and upper bounds when resistive control is used to maximize power generation. For complex conjugate control, this same expression only provides a lower bound, as there is theoretically no upper bound. These bounds assist in comparing the performance of the floating sphere if it were to extract energy using surge or heave motion. The analysis shows because of differences in hydrodynamic coefficients for each oscillating mode, there are different frequency ranges that provide better power capture efficiency. The influence of a motion constraint on TAP while utilizing a nonideal power take-off is examined and found to reduce the losses associated with bidirectional energy flow. The expression to calculate TAP with a nonideal PTO is modified by the electrical conversion efficiency and the ratio of the PTO spring and damping coefficients. Furthermore, the PTO spring and damping coefficients were separated in the expression, allowing for limits to be set on the PTO coefficients to ensure net power generation.

16 TIDAL AND WAVE POWER↗
DOE OSTI2023

Parametric Study of 239 Pu and 235 U Mass Equivalency in Spherical Geometries

In this technical document, parametric studies were performed to determine if a bounding mass equivalency can be used to simplify multi-material models in MCNP6 for criticality evaluations. The studies explored the simplification of MCNP6 models by substituting materials, particularly 239 Pu for 235 U. The initial hypothesis for the material simplification is based on the subcritical mass limits from the ANSI/ANS- 8.1 standard which states that for water reflected metal spheres the limits are 20.1 kg 235 U and 5.0 kg 239 Pu. Using those masses, it was hypothesized that a 20.1 kg sphere of 235 U and a 5.0 kg sphere of 239 Pu could be considered to have the same multiplication. Given that equivalency, 4kg 235 U could be modeled by 1kg 239 Pu (i.e., a mass ratio of 4:1). The parametric studies compare the materials 239 Pu and 235 U independently and in multi-material spherical models. Critical mass, critical thickness, critical mass ratio and critical thickness ratio values are presented as functions of inner radius. keff values are presented as a function of inner radius at different mass ratios and total masses. Results indicate that the conservative bounding mass equivalency is one. That is, if substituting 239 Pu for 235 U it should be done on a gram for gram basis.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

Mojo: MLIR-based Performance-Portable HPC Science Kernels on GPUs for the Python Ecosystem

We explore the performance and portability of the novel Mojo language for scientific computing workloads on GPUs. As the first language based on the LLVM’s Multi-Level Intermediate Representation (MLIR) compiler infrastructure, Mojo aims to close performance and productivity gaps by combining Python’s interoperability and CUDA-like syntax for compile-time portable GPU programming. We target four scientific workloads: a seven-point stencil (memory-bound), BabelStream (memory-bound), miniBUDE (compute-bound), and Hartree–Fock (compute-bound with atomic operations); and compare their performance against vendor baselines on NVIDIA H100 and AMD MI300A GPUs. We show that Mojo’s performance is competitive with CUDA and HIP for memory-bound kernels, whereas gaps exist on AMD GPUs for atomic operations and for fast-math compute-bound kernels on both AMD and NVIDIA GPUs. Although the learning curve and programming requirements are still fairly low-level, Mojo can close significant gaps in the fragmented Python ecosystem in the convergence of scientific computing and AI.

Godoy, William [ORNL] (ORCID:0000000225905178)↗
DOE OSTI2021

Tuning the Adsorption Properties of Metal–Organic Frameworks through Coadsorbed Ammonia

In this study, we report a novel strategy to increase the gas adsorption selectivity of metal organic framework materials by coadsorbing another molecular species. Specifically, we find that addition of tightly bound NH 3 molecules in the well-known metal–organic framework MOF-74 dramatically alters its adsorption behavior of C 2 H 2 and C 2 H 4 . Combining in situ infrared spectroscopy and ab initio calculations, we find that—as a result of coadsorbed NH 3 molecules attaching to the open metal sites—C 2 H 2 binds more strongly and diffuses much faster than C 2 H 4 , occupying the available space adjacent to metal-bound NH 3 molecules. Most remarkably, C 2 H 4 is now almost completely excluded from entering the MOF once C 2 H 2 has been loaded. This finding dispels the widespread belief that strongly coadsorbed species in nanoporous materials always undermine their performance in adsorbing or separating weakly bound target molecules. Furthermore, it suggests a new route to tune the adsorption behavior of MOF materials through harnessing the interactions among coadsorbed guests.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

A Framework for Error-Bounded Approximate Computing, with an Application to Dot Products

Approximate computing techniques, which trade off the computation accuracy of an algorithm for better performance and energy efficiency, have been successful in reducing computation and power costs in several domains. However, error sensitive applications in high-performance computing are unable to benefit from existing approximate computing strategies that are not developed with guaranteed error bounds. While approximate computing techniques can be developed for individual high-performance computing applications by domain specialists, this often requires additional theoretical analysis and potentially extensive software modification. Hence, the development of low-level error-bounded approximate computing strategies that can be introduced into any high-performance computing application without requiring additional analysis or significant software alterations is desirable. In this paper, we provide a contribution in this direction by proposing a general framework for designing error-bounded approximate computing strategies and apply it to the dot product kernel to develop \bf qdot---an error-bounded approximate dot product kernel. Following the introduction of qdot, here we perform a theoretical analysis that yields a deterministic bound on the relative approximation error introduced by qdot. Empirical tests are performed to illustrate the tightness of the derived error bound and to demonstrate the effectiveness of qdot on a synthetic dataset, as well as two scientific benchmarks---the conjugate gradient (CG) and power methods. In some instances, using qdot for the dot products in CG can result in many components being quantized to half precision without increasing the iteration count required for convergence to the same solution as CG using a double precision dot product.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Correlated conformational dynamics of the human GluN1-GluN2A type N-methyl-D-aspartate (NMDA) receptor

N-Methyl-D-aspartate receptors (NMDARs) are glutamate-gated ion channels found in the nerve cell membranes. As a result of overexcitation of NMDARs, neuronal death occurs and may lead to diseases such as epilepsy, stroke, Alzheimer’s disease, and Parkinson’s disease. In this study, human GluN1- GluN2A type NMDAR structure is modeled based on the X-ray structure of the Xenopus laevis template and missing loops are added by ab-initio loop modeling. The final structure is chosen according to two different model assessment scores. To be able to observe the structural changes upon ligand binding, glycine and glutamate molecules are docked into the corresponding binding sites of the receptor. Subsequently, molecular dynamics simulations of 1.3 μs are performed for both apo and ligand-bound structures. Structural parameters, which have been considered to show functionally important changes in previous NMDAR studies, are monitored as conformational rulers to understand the dynamics of the conformational changes. Moreover, principal component analysis (PCA) is performed for the equilibrated part of the simulations. Additionally, from these analyses, the differences in between apo and ligand-bound simulations can be summarized as the following: The girdle right at the beginning of the pore loop, which connects M2 and M3 helices of the ion channel, partially opens. Ligands act like an adhesive for the ligand-binding domain (LBD) by keeping the bi-lobed structure together and consequently this is reflected to the overall dynamics of the protein as an increased correlation of the LBD with especially the amino-terminal domain (ATD) of the protein.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Anderson Localization of Composite Particles

In this work, we investigate the effect of coupling between translational and internal degrees of freedom of composite quantum particles on their localization in a random potential. We show that entanglement between the two degrees of freedom weakens localization due to the upper bound imposed on the inverse participation ratio by purity of a quantum state. We perform numerical calculations for a two-particle system bound by a harmonic force in a 1D disordered lattice and a rigid rotor in a 2D disordered lattice. We illustrate that the coupling has a dramatic effect on localization properties, even with a small number of internal states participating in quantum dynamics.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Probing Postmeasurement Entanglement without Postselection

We study the problem of observing quantum collective phenomena emerging from large numbers of measurements. These phenomena are difficult to observe in conventional experiments because, in order to distinguish the effects of measurement from dephasing, it is necessary to postselect on sets of measurement outcomes with Born probabilities that are exponentially small in the number of measurements performed. An unconventional approach, which avoids this exponential “postselection problem”, is to construct cross-correlations between experimental data and the results of simulations on classical computers. However, these cross-correlations generally have no definite relation to physical quantities. We first show how to incorporate classical shadows into this framework, thereby allowing for the construction of quantum information-theoretic cross-correlations. We then identify cross-correlations that both upper and lower bound the measurement-averaged von Neumann entanglement entropy, as well as cross-correlations that lower bound the measurement-averaged purity and entanglement negativity. These bounds show that experiments can be performed to constrain postmeasurement entanglement without the need for postselection. To illustrate our technique, we consider how it could be used to observe the measurement-induced entanglement transition in Haar-random quantum circuits. We use exact numerical calculations as proxies for quantum simulations and, to highlight the fundamental limitations of classical memory, we construct cross-correlations with tensor-network calculations at finite bond dimension. Our results reveal a signature of measurement-induced criticality that can be observed using a quantum simulator in polynomial time and with polynomial classical memory. Published by the American Physical Society 2024

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Optimizing Error-Bounded Lossy Compression for Scientific Data on GPUs

Error-bounded lossy compression is a critical technique for significantly reducing scientific data volumes. With ever-emerging heterogeneous high-performance computing (HPC) architecture, GPU-accelerated error-bounded compressors (such as CUSZ and cuZFP) have been developed. However, they suffer from either low performance or low compression ratios. To this end, we propose CUSZ+ to target both high compression ratios and throughputs. We identify that data sparsity and data smoothness are key factors for high compression throughputs. Our key contributions in this work are fourfold: (1) We propose an efficient compression workflow to adaptively perform run-length encoding and/or variable-length encoding. (2) We derive Lorenzo reconstruction in decompression as multidimensional partial-sum computation and propose a fine-grained Lorenzo reconstruction algorithm for GPU architectures. (3) We carefully optimize each of CUSZ kernels by leveraging state-of-the-art CUDA parallel primitives. (4) We evaluate CUSZ+ using seven real-world HPC application datasets on V100 and A100 GPUs. Experiments show CUSZ+ improves the compression throughputs and ratios by up to 18.4x and 5.3x, respectively, over CUSZ on the tested datasets.

Tian, Jiannan↗
DOE OSTI2023

An Assessment of the Laminar Hypersonic Double-Cone Experiments in the LENS-XX Tunnel

This is an investigation on two experimental datasets of laminar hypersonic flows, over a double-cone geometry, acquired in Calspan—University at Buffalo Research Center’s Large Energy National Shock (LENS)-XX expansion tunnel. These datasets have yet to be modeled accurately. A previous paper suggested that this could partly be due to mis-specified inlet conditions. The authors of this paper solved a Bayesian inverse problem to infer the inlet conditions of the LENS-XX test section and found that in one case they lay outside the uncertainty bounds specified in the experimental dataset. However, the inference was performed using approximate surrogate models. Here in this paper, the experimental datasets are revisited and inversions for the tunnel test-section inlet conditions are performed with a Navier–Stokes simulator. The inversion is deterministic and can provide uncertainty bounds on the inlet conditions under a Gaussian assumption. It was found that deterministic inversion yields inlet conditions that do not agree with what was stated in the experiments. An a posteriori method is also presented to check the validity of the Gaussian assumption for the posterior distribution. This paper contributes to ongoing work on the assessment of datasets from challenging experiments conducted in extreme environments, where the experimental apparatus is pushed to the margins of its design and performance envelopes.

42 ENGINEERING↗
DOE OSTI2024

Shifting Between Compute and Memory Bounds: A Compression-Enabled Roofline Model

In the evolving landscape of high-performance computing, especially to fight the end of Moore’s Law and Dennard’s Scaling, the ability to shift between compute-bound and memory-bound states is critical for enhancing adaptability and flexibility to diverse system and domain-specific architectures. Such capability is vital for optimizing performance across distinguished hardware configurations, such as accelerators, memory hierarchies, and cache systems. Despite that ad hoc optimization techniques, such as compressed/approximate computation, have been enabled for compute-/data-intensive computing for improved performance in distinct hardware settings, there lacks an understanding of 1) the rational behind performance improvement; 2) capability of different optimizations; 3) what optimization to respond to specific computational and memory demands. This work proposes a compression-enabled roofline model to facilitate this adaptability with data compression techniques to balance and transform between computational and memory demands. This model enables applications to adjust in response to the specific strengths and limitations of the underlying hardware and system to optimize resource utilization. The effectiveness of this approach is demonstrated with matrix multiplication kernels on different input sizes, with turning on/off various compression techniques, including 1) low-precision floating point; 2) sparse matrix formulation; and 3) compressed arrays with ZFP. By reducing memory transfer volumes and cache misses and increasing data locality and computational intensity through compression, the specific roofline model can transform between compute and memory bounds to align more efficiently with system capabilities. This advancement not only improves overall performance but also maximizes adaptability in diverse computing environments.

Naraparaju, Ramasoumya [University of Washington]↗