Search NASA⌕ Search

SEARCH · Search NASA

Results for “sum reduction”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Evaluating the Performance of Integer Sum Reduction on an Intel GPU

Sum reduction is a primitive operation in parallel computing while SYCL is a promising heterogeneous programming language. In this paper, we describe the SYCL implementations of integer sum reduction using atomic functions, shared local memory, vectorized memory accesses, and parameterized workload sizes. Evaluating the reduction kernels shows that we can achieve 1.4X speedup over the open-source implementations of sum reduction for a sufficiently large number of integers on an Intel integrated GPU.

Jin, Zheming↗
DOE OSTI2022

Integer Sum Reduction with OpenMP on an AMD MI100 GPU

Sum reduction is a primitive operation in parallel computing. Device offload support allows a user to use OpenMP directives to take advantage of a highly capable GPU. In this paper, we present the integer sum reduction annotated with the OpenMP directives and evaluate the performance impacts of tunable parameters with the AOMP and GCC compilers on an AMD MI100 GPU. In addition, we explain the implementations of the OpenMP reduction by the compilers. Sweeping over the pruned parameter space, we find that the speedup is approximately 20 with AOMP, and the reduction performance using AOMP is approximately 11% higher than that using GCC. However, the OpenMP offload performance is approximately 30% lower compared to the performance of the reductions written with rocThrust or hipCUB.

Jin, Zheming↗
DOE OSTI2019

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↗
DOE OSTI2021

Evaluating the Performance of Integer Sum Reduction in SYCL on GPUs

SYCL is a promising programming model for heterogeneous computing—allowing a single-source code to target devices from multiple vendors. One significant task performed on these accelerators is a primitive operation for integer sum reduction. This paper presents several SYCL implementations of integer sum reduction—using atomic functions, shared local memory, vectorized memory accesses and parameterized workload sizes—to compare the performance and maturity of SYCL against open-source vendor-specific implementations of the same reduction. For a sufficiently large number of integers, tuning the parameters of our SYCL implementations achieves 1.4X speedup over the open-source implementations on an Intel UHD630 integrated GPU. The SYCL reduction is 3% faster than the templated reduction in Thrust, and 0.3% faster than the device reduction in CUB on an Nvidia P100 GPU. The SYCL reduction is 1.9% faster than the templated reduction in Thrust, and 0.4% faster than the device reduction in CUB on an Nvidia V100 GPU.

Jin, Zheming↗
DOE OSTI2024

Sum Reduction with OpenMP Offload on NVIDIA Grace-Hopper System

We evaluate the performance of the baseline and optimized reductions in OpenMP on an NVIDIA Grace-Hopper system. We explore the impacts of the number of teams, the number of elements to sum per loop iteration, and simultaneous execution on the central-processing unit (CPU) and the GPU in the unified memory (UM) mode upon the reduction performance. The experimental results show that the optimized reductions are 6.120X to 20.906X faster than the baselines on the GPU, and their efficiency ranges from 89% to 95% of the theoretical GPU memory bandwidth. Depending on where an input array is allocated in the program when co-running the reduction on the CPU and GPU in the UM mode, the average speedup over the GPU-only execution is approximately 2.484 or 1.067, and the speedup of the optimized reductions over the baseline reductions ranges from 0.996 to 10.654 or from 0.998 to 6.729.

Jin, Zheming↗
DOE OSTI2025

A Study of Performance Portability of Low-bit Fused Matrix-Vector Multiplication Kernels in SYCL

Understanding the causes of performance gaps between a portable programming model and a vendor-specific programming model is important for improving performance portability. This paper studies performance portability of low-bit fused general matrix-vector multiplication kernels in SYCL on vendors’ graphics processing units (GPUs). This work introduces the use case, explains the kernel implementations in detail, evaluates the performance of the CUDA, HIP, and SYCL kernels on datacenter, desktop, and laptop GPUs, and investigates the causes of performance gaps. The results show that loop unrolling, kernel dispatch overhead, and sum reduction contribute to the gaps.

Jin, Zheming [ORNL] (ORCID:000000027197780X)↗
DOE OSTI2024

Exploring the influence of interfacial solvation on electrochemical CO 2 reduction using plasmon‐enhanced vibrational sum frequency generation spectroscopy

Although interfacial solvation plays an important role in determining carbon dioxide reduction (CO 2 R) kinetics, present understanding of the potential dependent properties of the electrochemical double layer under conditions relevant for CO 2 R remains limited. This article summarizes the development and recent applications of plasmon-enhanced vibrational sum frequency generation (VSFG) spectroscopy to study the effects of cation hydration and interfacial solvation on CO 2 R using CO as a vibrational Stark reporter. Results show that electrolyte cations retain their entire solvation shell upon adsorption to inactive sites, while active sites retain only a single water layer between the gold surface and the cation. Measurements also show that the total interfacial electric field can be separated into two contributions: one from the electrochemical double layer (Stern field) and another from the polar solvation environment (Onsager field). Surprisingly, correlating VSFG spectra with reaction kinetics reveals that it is the solvation-mediated Onsager field that governs the chemical reactivity at the electrode/electrolyte interface. Measuring the interfacial water spectra during electrocatalysis also provides evidence for the proton source during H2 evolution, which competes with CO 2 R in aqueous electrolyte. These findings highlight the importance of directly probing cation hydration and interfacial solvation, which mediates reaction kinetics at electrochemical interfaces.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS1989

Sulfate reduction and oxic respiration in marine sediments: implications for organic carbon preservation in euxinic environments

Compilations have been made of sulfate reduction rates and oxic respiration rates over the entire range of marine sedimentation rates, and sedimentary environments, including several euxinic sites. These data show, consistent with the findings of Jorgensen (1982, Nature, 296, 643-645), that sulfate reduction and oxic respiration oxidize equal amounts of organic carbon in nearshore sediments. As sedimentation rates decrease, oxic respiration, becomes progressively more important, and in deep-sea sediments 100-1000 times more organic carbon is oxidized by oxic respiration than by sulfate reduction. By contrast, nearly as much organic carbon is oxidized by sulfate reduction in euxinic sediments as is oxidized by the sum of sulfate reduction and oxic respiration in normal marine sediments of similar deposition rate. This observation appears at odds with the enhanced preservation of organic carbon observed in euxinic sediments. However, only small reductions in (depth-integrated) organic carbon decomposition rates (compared to normal marine) are required to give both high organic carbon concentrations and enhanced carbon preservation in euxinic sediments. Lower rates of organic carbon decomposition (if only by subtle amounts) are explained by the diminished ability of anaerobic bacteria to oxidize the full suite of sedimentary organic compounds.

NASA Center ARC↗
NASA NTRS1984

Investigation of the external flow analysis for density measurements at high altitude

The results of analysis performed on the external flow around the shuttle orbiter nose regions at the Shuttle Upper Atmosphere Mass Spectrometer (SUMS) inlet orifice are presented. The purpose of the analysis is to quantitatively characterize the flow conditions to facilitate SUMS flight data reduction and subsequent determination of orbiter aerodynamic force coefficients in the hypersonic rarefied flow regime. Experimental determination of aerodynamic force coefficients requires accurate simultaneous measurement of forces (or acceleration) and dynamic pressure along with independent knowledge of density and velocity. The SUMS provides independent measurement of dynamic pressure; however, it does so indirectly and requires knowledge of the relationship between measured orifice conditions and the dynamic pressure which can only be determined on the basis of molecule or theory for a winged configuration. Monte Carlo direct simulation computer codes were developed for both the flow field solution at the orifice and for the internal orifice flow. These codes were used to study issues associated with geometric modeling of the orbiter nose geometry and the modeling of intermolecular collisions including rotational energy exchange and a preliminary analysis of vibrational excitation and dissociation effects. Data obtained from preliminary simulation runs are presented.

Bienkowski, G. K.↗
DOE OSTI2019

Simulation of Local Error Correction of the Surface Shape by a Low-Dimensional Ion Beam

We propose an algorithm for solving the problem of local error correction of the surface shape by a low-dimensional ion beam. The algorithm presumes successive sampling running over protrusions relative to the average height aimed at searching for the optimal etching point satisfying the criterion for the reduction of the sum of derivative moduli on the etching spot. It is shown that the new approach makes it possible to considerably extend the range of spatial frequencies accessible to the action for a given dimension of an ion beam.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
NASA NTRS2014

Experiments on Exhaust Noise of Tightly Integrated Propulsion Systems

A wide-ranging series of tests have been completed that seek to map the effects of installation, including jet by jet interaction effects, on exhaust noise from various nozzles in forward flight. The primary data was far-field acoustic spectral directivity. The goals of the test series were (i) to generate enough data for empirical models of the different effects, and (ii) to provide data for advanced computational noise predictions methods applied to simplified yet realistic configurations. Data is presented that demonstrate several checks on data quality and that provide an overview of trends observed to date. Among the findings presented here: (i) Data was repeatable between jet rigs for single nozzles with and without surfaces to within +/- 0.5 dB. (ii) The presence of a second jet caused a strong reduction of the summed noise in the plane of the two plumes and an increase over the expected source doubling in most other azimuthal planes. (iii) The impact of the second jet was reduced when the jets were unheated. (iv) The impact of adding a second isolated rectangular jet was relatively independent of the nozzle aspect ratio up to aspect ratio 8:1. (v) Forward flight had similar impact on a high aspect ratio (8:1) jet as on an axisymmetric jet, except at the peak noise angle where the impact was less. (vi) The effect of adding a second round jet to a tightly integrated nozzle where the nozzle lip was less than a diameter from the surface was very dependent upon the length of the surface downstream of the nozzle. (vii) When the nozzles were rectangular and tightly integrated with the airframe surface the impact of a second jet was very dependent upon how close together the two jets were. This paper serves as an overview of the test; other papers presented in the same conference will give more detailed analysis of the results.

Noise↗
DOE OSTI2020

GMRES with embedded ensemble propagation for the efficient solution of parametric linear systems in uncertainty quantification of computational models

In a previous work, embedded ensemble propagation was proposed to improve the efficiency of sampling-based uncertainty quantification methods of computational models on emerging computational architectures. It consists of simultaneously evaluating the model for a subset of samples together, instead of evaluating them individually. A first approach introduced to solve parametric linear systems with ensemble propagation is ensemble reduction. In Krylov methods for example, this reduction consists in coupling the samples together using an inner product that sums the sample contributions. Ensemble reduction has the advantages of being able to use optimized implementations of BLAS functions and having a stopping criterion which involves only one scalar. However, the reduction potentially decreases the rate of convergence due to the gathering of the spectra of the samples. In this paper, we investigate a second approach: ensemble propagation without ensemble reduction in the case of GMRES. This second approach solves each sample simultaneously but independently to improve the convergence compared to ensemble reduction. This raises two new issues which are solved in this paper: the fact that optimized implementations of BLAS functions cannot be used anymore and that ensemble divergence, whereby individual samples within an ensemble must follow different code execution paths, can occur. We tackle those issues by implementing a high-performing ensemble GEMV and by using masks. The proposed ensemble GEMV leads to a similar cost per GMRES iteration for both approaches, i.e. with and without reduction. For illustration, we study the performances of the new linear solver in the context of a mesh tying problem. Furthermore, this example demonstrates improved ensemble propagation speed-up without reduction.

BLAS↗
NASA NTRS1988

Shuttle to Shuttle 2: Subsystem weight reduction potential (estimated 1992 technology readiness date)

The objective of this study was to make estimates of the weight savings that might be realized on all the subsystems on an advanced rocket-powered shuttle (designated Shuttle 2) by using advanced technologies having a projected maturity date of 1992. The current Shuttle with external tank was used as a baseline from which to make the estimates of weight savings on each subsystem. The subsystems with the greatest potential for weight reduction are the body shell and the thermal protection system. For the body shell, a reduction of 5.2 percent in the weight of the vehicle at main engine cutoff is projected through the application of new technologies, and an additional configuration-based reduction of 5 percent is projected through simplification of body shape. A reduction of 5 percent is projected for the thermal protection system through experience with the current Space Shuttle and the potential for reducing thermal protection system thicknesses in selected areas. Main propellant tanks are expected to increase slightly in weight. The main propulsion system is also projected to increase in weight because of the requirement to operate engines at derated power levels in order to accommodate one-engine-out capability. The projections for weight reductions through improvements in the remaining subsystems are relatively small. By summing all the technology factors, a projected reduction of 16 percent in the vehicle weight at main engine cutoff is obtained. By summarizing the configurational factors, a potential reduction of 12 percent in vehicle weight is obtained.

Macconochie, Ian O.↗
DOE OSTI2023

Truncated non-linear interferometer-based sensor system

A truncated non-linear interferometer-based sensor system includes an input that receives an optical beam and a non-linear amplifier that generates a probe beam and a conjugate beam from the optical beam. The system's local oscillators are related to the probe beam and the conjugate beam. The system includes a sensor that transduces an input with the probe beam and the conjugate beam. The transduction detects changes in the phase of each of the probe beam and the conjugate beam. The system's phase sensitive detectors detect phase modulations between the respective local oscillators, the probe beam, and the conjugate beam and outputs phase signals based on detected phase modulations. The system measures phase signals indicative of the sensor's input resulting from a sum or difference of the phase signals. The measurement exhibits a quantum noise reduction in an intensity difference, a phase sum, or an amplitude difference quadrature.

47 OTHER INSTRUMENTATION↗
NASA NTRS1994

An analysis of spectral envelope-reduction via quadratic assignment problems

A new spectral algorithm for reordering a sparse symmetric matrix to reduce its envelope size was described. The ordering is computed by associating a Laplacian matrix with the given matrix and then sorting the components of a specified eigenvector of the Laplacian. In this paper, we provide an analysis of the spectral envelope reduction algorithm. We described related 1- and 2-sum problems; the former is related to the envelope size, while the latter is related to an upper bound on the work involved in an envelope Cholesky factorization scheme. We formulate the latter two problems as quadratic assignment problems, and then study the 2-sum problem in more detail. We obtain lower bounds on the 2-sum by considering a projected quadratic assignment problem, and then show that finding a permutation matrix closest to an orthogonal matrix attaining one of the lower bounds justifies the spectral envelope reduction algorithm. The lower bound on the 2-sum is seen to be tight for reasonably 'uniform' finite element meshes. We also obtain asymptotically tight lower bounds for the envelope size for certain classes of meshes.

George, Alan↗
DOE OSTI2024

Truncated nonlinear interferometer-based sensor system

A truncated non-linear interferometer-based sensor system includes an input port that receives an optical beam and a non-linear amplifier that amplifies the optical beam with a pump beam and renders a probe beam and a conjugate beam. The system's local oscillators have a relationship with the respective beams. The system includes a sensor that transduces an input with the probe beam and the conjugate beam or their respective local oscillators. It includes one or more phase-sensitive detectors that detect a phase modulation between the respective local oscillators and the probe beam and the conjugate beam. Output from the phase-sensitive-detectors is based on the detected phase modulation. The phase-sensor-detectors include measurement circuitry that measure the phase signals. The measurement is the sum or difference of the phase signals in which the measured combination exhibit a quantum noise reduction in an intensity difference or a phase sum or an amplitude difference quadrature.

Pooser, Raphael C.↗
DOE OSTI2020

Corn stover harvest reduces soil CO 2 fluxes but increases overall C losses

Abstract Harvesting corn ( Zea mays L.) stover for use as bioenergy feedstock may provide short‐term economic benefits and perhaps improve grain yield in continuous corn systems, but excessive stover removal may lead to long‐term depletion of soil C stocks. To better quantify the impacts of stover harvest on the soil C balance, we investigated CO 2 fluxes under three harvest treatments (none, moderate, and high) in three continuous corn systems: (a) no tillage and no biochar applications, (b) chisel plowing with biochar amendments, and (c) chisel plowing without biochar amendments were quantified. We sampled static chambers 14, 13, and 15 times in 2010, 2011, and 2012, respectively, and the measurements were used to calibrate the Agricultural Production Systems sIMulator model. Although CO 2 ‐C emissions did not differ among the three systems, both moderate (approximately 30%) and high (approximately 60%) stover removal rates reduced simulated CO 2 ‐C emissions by nearly 10% and more than 22%, respectively. Despite these reductions in CO 2 flux, the sum of CO 2 ‐C and stover C exceeded CO 2 ‐C losses in those plots without stover removal. This finding suggested soil C depletion was occurring for the soil and crop management practices used, even though depletion was not evident after 3 years. Application of biochar increased soil C levels, suggesting it may be able to offset some C losses. We conclude that for sustainable, highly productive agroecosystems that include stover harvest, all agronomic practices must be optimized to minimize C losses and maintain soil quality.

54 ENVIRONMENTAL SCIENCES↗