Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Energy usage”

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 1,657 records · Page 92

Resource-Efficient, Hierarchical Auto-Tuning of a Hybrid Lattice Boltzmann Computation on the Cray XT4

We apply auto-tuning to a hybrid MPI-pthreads lattice Boltzmann computation running on the Cray XT4 at National Energy Research Scientific Computing Center (NERSC). Previous work showed that multicore-specific auto-tuning can improve the performance of lattice Boltzmann magnetohydrodynamics (LBMHD) by a factor of 4x when running on dual- and quad-core Opteron dual-socket SMPs. We extend these studies to the distributed memory arena via a hybrid MPI/pthreads implementation. In addition to conventional auto-tuning at the local SMP node, we tune at the message-passing level to determine the optimal aspect ratio as well as the correct balance between MPI tasks and threads per MPI task. Our study presents a detailed performance analysis when moving along an isocurve of constant hardware usage: fixed total memory, total cores, and total nodes. Overall, our work points to approaches for improving intra- and inter-node efficiency on large-scale multicore systems for demanding scientific applications.

97 MATHEMATICS AND COMPUTING↗

RCSB Protein Data Bank: Celebrating 50 years of the PDB with new tools for understanding and visualizing biological macromolecules in 3D

We report the Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB), funded by the US National Science Foundation, National Institutes of Health, and Department of Energy, has served structural biologists and Protein Data Bank (PDB) data consumers worldwide since 1999. RCSB PDB, a founding member of the Worldwide Protein Data Bank (wwPDB) partnership, is the US data center for the global PDB archive housing biomolecular structure data. RCSB PDB is also responsible for the security of PDB data, as the wwPDB-designated Archive Keeper. Annually, RCSB PDB serves tens of thousands of three-dimensional (3D) macromolecular structure data depositors (using macromolecular crystallography, nuclear magnetic resonance spectroscopy, electron microscopy, and micro-electron diffraction) from all inhabited continents. RCSB PDB makes PDB data available from its research-focused RCSB.org web portal at no charge and without usage restrictions to millions of PDB data consumers working in every nation and territory worldwide. In addition, RCSB PDB operates an outreach and education PDB101.RCSB.org web portal that was used by more than 800,000 educators, students, and members of the public during calendar year 2020. This invited Tools Issue contribution describes (i) how the archive is growing and evolving as new experimental methods generate ever larger and more complex biomolecular structures; (ii) the importance of data standards and data remediation in effective management of the archive and facile integration with more than 50 external data resources; and (iii) new tools and features for 3D structure analysis and visualization made available during the past year via the RCSB.org web portal.

59 BASIC BIOLOGICAL SCIENCES↗

In situ strain relaxation comparison between GaAsBi and GaInAs grown by molecular-beam epitaxy

The strain relaxation of GaAsBi is studied in order to determine both the maximum thickness before dislocations form for various misfits and the potential of GaAsBi for usage in the compositionally graded buffer of lattice-mismatched devices. Low-misfit GaAsBi epilayers are grown and compared with GaInAs, a well-studied material currently used in compositional grades. Relaxation behavior and dislocation energetics are compared using in situ wafer curvature. Samples are grown using molecular-beam epitaxy under similar conditions with misfit up to -0.81%. GaAsBi begins to relax at a lower thickness than GaInAs of the same mismatch. This leads to the majority of GaAsBi strain relaxation occurring with less material growth than comparable GaInAs. However, GaAsBi has greater residual strain than GaInAs after 2 {mu}m of growth. These results indicate that GaAsBi requires less elastic energy to form dislocations and more elastic energy to either encourage glide or multiply dislocations than GaInAs. GaAsBi shows less surface roughness than GaInAs for all samples, ruling out roughness as a source of dislocations and hindered glide in these alloys.

36 MATERIALS SCIENCE↗

Ferroelectric materials toward next-generation electromechanical technologies

Ferroelectric materials have been widely used in various electromechanical devices, from ultrasonic transducers and actuators to mechanical energy harvesters. The key performance metrics of these devices, such as sensitivity, efficiency, and bandwidth of ultrasonic transducers, are largely determined by the piezoelectric properties. This Review highlights recent research progress in improving the piezoelectricity of ferroelectric materials and offers potential strategies for further enhancement to meet the ever-increasing demands for high-performance piezoelectric devices and systems. Here, it provides insights into the future development of ferroelectrics to address the increasing demands of emerging applications, including photoacoustic imaging and piezoelectric fans and motors in integrated circuit–enabled electronic devices. Additionally, it emphasizes the need to consider environmental impacts across the entire life cycle of ferroelectrics, from sourcing and manufacturing to usage and disposal.

36 MATERIALS SCIENCE↗

LSENS: A General Chemical Kinetics and Sensitivity Analysis Code for homogeneous gas-phase reactions. Part 1: Theory and numerical solution procedures

LSENS, the Lewis General Chemical Kinetics and Sensitivity Analysis Code, has been developed for solving complex, homogeneous, gas-phase chemical kinetics problems and contains sensitivity analysis for a variety of problems, including nonisothermal situations. This report is part 1 of a series of three reference publications that describe LENS, provide a detailed guide to its usage, and present many example problems. Part 1 derives the governing equations and describes the numerical solution procedures for the types of problems that can be solved. The accuracy and efficiency of LSENS are examined by means of various test problems, and comparisons with other methods and codes are presented. LSENS is a flexible, convenient, accurate, and efficient solver for chemical reaction problems such as static system; steady, one-dimensional, inviscid flow; reaction behind incident shock wave, including boundary layer correction; and perfectly stirred (highly backmixed) reactor. In addition, the chemical equilibrium state can be computed for the following assigned states: temperature and pressure, enthalpy and pressure, temperature and volume, and internal energy and volume. For static problems the code computes the sensitivity coefficients of the dependent variables and their temporal derivatives with respect to the initial values of the dependent variables and/or the three rate coefficient parameters of the chemical reactions.

Radhakrishnan, Krishnan↗

Calculations of High-Temperature Jet Flow Using Hybrid Reynolds-Average Navier-Stokes Formulations

Two multiscale-type turbulence models are implemented in the PAB3D solver. The models are based on modifying the Reynolds-averaged Navier Stokes equations. The first scheme is a hybrid Reynolds-averaged- Navier Stokes/large-eddy-simulation model using the two-equation k(epsilon) model with a Reynolds-averaged-Navier Stokes/large-eddy-simulation transition function dependent on grid spacing and the computed turbulence length scale. The second scheme is a modified version of the partially averaged Navier Stokes model in which the unresolved kinetic energy parameter f(sub k) is allowed to vary as a function of grid spacing and the turbulence length scale. This parameter is estimated based on a novel two-stage procedure to efficiently estimate the level of scale resolution possible for a given flow on a given grid for partially averaged Navier Stokes. It has been found that the prescribed scale resolution can play a major role in obtaining accurate flow solutions. The parameter f(sub k) varies between zero and one and is equal to one in the viscous sublayer and when the Reynolds-averaged Navier Stokes turbulent viscosity becomes smaller than the large-eddy-simulation viscosity. The formulation, usage methodology, and validation examples are presented to demonstrate the enhancement of PAB3D's time-accurate turbulence modeling capabilities. The accurate simulations of flow and turbulent quantities will provide a valuable tool for accurate jet noise predictions. Solutions from these models are compared with Reynolds-averaged Navier Stokes results and experimental data for high-temperature jet flows. The current results show promise for the capability of hybrid Reynolds-averaged Navier Stokes and large eddy simulation and partially averaged Navier Stokes in simulating such flow phenomena.

Abdol-Hamid, Khaled S.↗

Investigation of ion irradiation effects on mineral analogues of concrete aggregates

Irradiation can cause prominent damage to reactor concrete aggregates leading to amorphization, strength and modulus decrease, radiation induced volume expansion (RIVE) and micro-cracking, which limits their long-term performance. Here, to develop an improved understanding of irradiation effects in concrete, three mineral analogues of concrete aggregates (limestone, marble and quartzite) were irradiated by 5.5 MeV He ions and 13 MeV Ni ions to surface doses of 0.011 displacements per atom (dpa) and 0.23 dpa, respectively, at room temperature. The two different ion species allow irradiation spectrum effects (ionizing and displacive) to be examined. Irradiation induced cracks were observed in He irradiated limestone and marble, and Ni irradiated quartzite. Full amorphization was observed in Ni irradiated quartzite with 14.3 % RIVE, and ∼25 % hardness and modulus decrease, while almost no change was observed in He irradiated quartzite except 4.35 % RIVE, revealing a possible ionization enhanced diffusion effect for high energy light ions. Furthermore, partial amorphization was observed in Ni irradiated marble and limestone matrix with a 12 % hardness decrease in marble while no amorphization was observed for He irradiation with a 20 % hardness increase in limestone matrix. The role of knock-on damage and irradiation spectrum on amorphization, volumetric expansion and mechanical property changes are discussed. Moreover, the onset and critical doses for amorphization and RIVE in quartz are obtained for ion irradiations at room temperature. The dose dependence of RIVE exhibits a delay compared to the amorphization behavior. The superior irradiation resistance of calcite phase compared to quartz phase implies there could be advantages to using calcareous aggregates and lowering the usage of siliceous aggregates for concrete in nuclear power plants for extended operation beyond 60 years. However, other effects such as corrosion, aging and reactions during severe accidents should also be considered, and further investigations are needed.

Amorphous↗

Empirical Investigations of the Opportunity Limits of Automatic Residential Electric Load Shaping

Residential electric load shaping is often modeled as infrequent, utility-initiated, short-duration deferral of peak demand through direct load control. In contrast, modeled herein is the potential for frequent, transactive, intraday, consumer-configurable load shaping for storage-capable thermostatically controlled electric loads (TCLs), including refrigerators, freezers, and hot water heaters. Unique to this study are 28 months of 15-minute-interval observations of usage in 101 homes in the Pacific Northwest United States that specify exact start, duration, and usage patterns of approximately 25 submetered loads per home. The magnitudes of the load shift from voluntarily-participating TCL appliances are aggregated to form hourly upper and lower load-shaping limits for the coordination of electrical generation, transmission, distribution, storage, and demand. Empirical data are statistically analyzed to define metrics that help quantify load-shaping opportunities.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Cysteine Usage in Sulfolobus Spindle-Shaped Virus 1 And Extension to Hyperthermophilic Viruses in General

Fuselloviridae are ubiquitous crenarchaeal viruses found in high-temperature acidic hot springs worldwide. The type virus, Sulfolobus spindle-shaped virus 1 (SSV1), has a double-stranded DNA genome that contains 34 open reading frames (ORFs). Fuselloviral genomes show little similarity to other organisms, generally precluding functional predictions. However, tertiary protein structure can provide insight into protein function. We have thus undertaken a systematic investigation of the SSV1 proteome and report here on the F112 gene product. Biochemical, proteomic and structural studies reveal a monomeric intracellular protein that adopts a winged helix DNA binding fold. Notably, the structure contains an intrachain disulfide bond, prompting analysis of cysteine usage in this and other hyperthermophilic viral genomes. The analysis supports a general abundance of disulfide bonds in the intracellular proteins of hyperthermophilic viruses, and reveals decreased cysteine content in the membrane proteins of hyperthermophilic viruses infecting Sulfolobales. The evolutionary implications of the SSV1 distribution are discussed.

15 GEOTHERMAL ENERGY↗

Coal-based Bricks & Blocks (CBBs): Process Development to Prototype Fabrication Coupled with Techno-Economic Analysis and Market Survey

The decline of coal use for energy production provides an abundance of local feedstock for new innovative uses and value-added products. Expanding the U.S. coal-value chain to manufacture high-value carbon products can strengthen the nation’s energy and mineral security, enhance the U.S. national defense security, increase the United States’ economic prosperity, while achieving U.S. environmental objectives. The primary project goal is a relatively light-weight composite product with superior (or comparable) compressive strength. Both virgin and, where available, post-consumer recycled thermoplastic versions are tested for each thermoplastic species. A useful attribute of thermoplastics as binders is that they can be heated to their melting point, cooled, and reheated again without significant degradation. A key advantage of thermoset CBBs is that they require only mixing and molding. CBB advantages include low cost, availability, binding ability and processability. Coal-based bricks and blocks (CBBs) weigh about 50% less than clay bricks and can be manufactured with an interlocking design to promote ease of use for the novice builder. CBB formulation is evaluated according to a design-of-experiments (DoE) approach. DoE variables are a) relative weight fractions of binder, b) relative proportions of large versus small (milled) anthracite size fractions, and c) additive percentage. Fabrication methods include hot-press molding and extrusion, the later being the most commercially viable. CBBs are tested for compressive strength, modulus of rupture (by flexure test) and water absorption per ASTM C67, with density determined by the Archimedes drainage method. Fractured interfaces are examined by SEM (Scanning Electron Microscopy) to resolve fracture dynamics and interior microstructure uniformity. Differential scanning calorimetry (DSC) is used to compare plastic transition temperatures i.e., glass and melting temperatures to contrast virgin with post-consumer recycled thermoplastics and optimize their usage. These results are used in the DoE analysis to identify the binder and relative weight percentages for optimum strength, density, and porosity. Overall, CBBs possess strength comparable to clay-based bricks but are non-permeable and hydrophobic, and hence resistant to degradation by freeze-fracturing, corrosion, and efflorescence. The strongest composites have been made with the following thermoplastic binders (in order of strength): thermoset, high-density polyethylene crosslink resin, high-density polyethylene, nylon 6/6, and polypropylene. Results from a techno-economic analysis TEA show economy of scale for CBBs by modularization and reveal the binder as the cost driver for material costs. Ideally, the incorporation of post-consumer recycled thermoplastic will decrease material acquisition costs and increase product sustainability. Notably, CBBs do not require the high temperature calcination needed to produce cement, nor do they require firing in the 1600-2400 °F range for three days using natural gas, as do clay brick equivalents. Instead, CBBs are heated to a modest <600 °F according to the melt flow index of the thermoplastic binder. Existing anthracite mines can be expanded to produce CBBs to reduce aggregate transportation costs and emissions that exist for clay bricks. TEA reflects this reduced energy cost while a comparative CO2 emission analysis quantifies the reduced environmental footprint. The market survey identifies several commercialization opportunities, dependent upon the brick classification.

01 COAL, LIGNITE, AND PEAT↗

Calculations vs. Measurements for Remnant Dose Rates from SNS Spent Structures

The Spallation Neutron Source (SNS) in Oak Ridge, Tennessee, is an accelerator driven neutron scattering facility for materials research. Presently SNS is capable to operate at 1.4 MW proton beam power incident on a mercury target with a proton beam energy of 1 GeV and 60 Hz repetition rate. SNS target system components are periodically replaced because they reach their end-of-life due to radiation induced material damage. Target vessel, which houses mercury target, is exchanged about two-three times per year and the proton beam window (PBW) is exchanged every two – three years.Each spent structure that leaves the SNS site requires supporting documentation with radionuclide inventory and dose rate prediction for the time of the transportation. Neutronics analyses are performed, assuming realistic irradiation history and decay case to ensure that the container/package, housing the structure, is compliant with the waste management regulations. Analyses are complex due to geometry, multi-code usage and following data treatment.To validate analyses, measurements of dose rates from the spent target vessel # 13 and PBW module #5 were performed. Neutronics analyses were performed to calculate residual dose rates from both structures for the time of measurements.

Popova, Irina I.↗

A Segregated Approach for Modeling the Electrochemistry in the 3-D Microstructure of Li-Ion Batteries and Its Acceleration Using Block Preconditioners

Abstract Battery performance is strongly correlated with electrode microstructure. Electrode materials for lithium-ion batteries have complex microstructure geometries that require millions of degrees of freedom to solve the electrochemical system at the microstructure scale. A fast-iterative solver with an appropriate preconditioner is then required to simulate large representative volume in a reasonable time. In this work, a finite element electrochemical model is developed to resolve the concentration and potential within the electrode active materials and the electrolyte domains at the microstructure scale, with an emphasis on numerical stability and scaling performances. The block Gauss-Seidel (BGS) numerical method is implemented because the system of equations within the electrodes is coupled only through the nonlinear Butler–Volmer equation, which governs the electrochemical reaction at the interface between the domains. The best solution strategy found in this work consists of splitting the system into two blocks—one for the concentration and one for the potential field—and then performing block generalized minimal residual preconditioned with algebraic multigrid, using the FEniCS and the Portable, Extensible Toolkit for Scientific Computation libraries. Significant improvements in terms of time to solution (six times faster) and memory usage (halving) are achieved compared with the MUltifrontal Massively Parallel sparse direct Solver. Additionally, BGS experiences decent strong parallel scaling within the electrode domains. Last, the system of equations is modified to specifically address numerical instability induced by electrolyte depletion, which is particularly valuable for simulating fast-charge scenarios relevant for automotive application.

25 ENERGY STORAGE↗

An evolutionarily conserved RNase-based mechanism for repression of transcriptional positive autoregulation

It is known that environmental context influences the degree of regulation at the transcriptional and post-transcriptional levels. However, the principles governing the differential usage and interplay of regulation at these two levels are not clear. Here, we show that the integration of transcriptional and post-transcriptional regulatory mechanisms in a characteristic network motif drives efficient environment-dependent state transitions. Through phenotypic screening, systems analysis, and rigorous experimental validation, we discovered an RNase (VNG2099C) in Halobacterium salinarum that is transcriptionally co-regulated with genes of the aerobic physiologic state but acts on transcripts of the anaerobic state. Through modelling and experimentation we show that this arrangement generates an efficient state-transition switch, within which RNase-repression of a transcriptional positive autoregulation (RPAR) loop is critical for shutting down ATP-consuming active potassium uptake to conserve energy required for salinity adaptation under aerobic, high potassium, or dark conditions. Subsequently, we discovered that many Escherichia coli operons with energy-associated functions are also putatively controlled by RPAR indicating that this network motif may have evolved independently in phylogenetically distant organisms. Thus, our data suggest that interplay of transcriptional and post-transcriptional regulation in the RPAR motif is a generalized principle for efficient environment-dependent state transitions across prokaryotes.

59 BASIC BIOLOGICAL SCIENCES↗

Haptic seat for fuel economy feedback

A process of providing driver fuel economy feedback is disclosed in which vehicle sensors provide for haptic feedback on fuel usage. Such sensors may include one or more of a speed sensors, global position satellite units, vehicle pitch/roll angle sensors, suspension displacement sensors, longitudinal accelerometer sensors, throttle position in sensors, steering angle sensors, break pressure sensors, and lateral accelerometer sensors. Sensors used singlely or collectively can provide enhanced feedback as to various environmental conditions and operating conditions such that a more accurate assessment of fuel economy information can be provided to the driver.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Advanced Transmission Technologies –GETs and HPCs Session 3: HPCs and Building Actions Plans to Digital Assurance Risks

The third session of the Idaho National Laboratory’s (INL) Technical Assistance for Digital Assurance (TADA) program, held on November 11, 2025, centered on High Performance Conductors (HPCs) and the formulation of action plans to address digital assurance risks associated with Grid-Enhancing Technologies (GETs). This session convened experts from utilities, vendors, and government agencies to examine the technical, operational, and cybersecurity aspects of HPC deployment. Discussions highlighted the benefits of HPCs, such as their ability to rapidly increase transmission capacity using existing corridors, improve grid resilience, reduce system losses, and align with FERC Orders 2023 and 1920. Participants evaluated supply chain and digital assurance risks, including reliance on imported materials, limited domestic manufacturing capacity, workforce shortages, and traceability issues. The session also emphasized the importance of digital trust, integration-layer cybersecurity, and unified risk frameworks, introducing tools like intrusion detection systems, encryption, zero trust networking, and firmware integrity. Recaps of earlier workshops on Dynamic Line Ratings (DLRs), Advanced Power Flow Control (APFC), and Transmission Topology Optimization (TTO) underscored institutional barriers and integration challenges. Action plans were proposed to mitigate issues such as inconsistent cybersecurity practices, SBOM usage, supply chain visibility, operator trust, and misaligned incentives. Additionally, INL presented its supply chain risk management tools and Cyber-Informed Engineering (CIE) principles to support secure procurement and system design. The session concluded with a commitment to share key takeaways, incorporate cohort feedback into future policy development, and continue collaborative engagement through upcoming pilot activities. Session 3 of 3.

24 - POWER TRANSMISSION AND DISTRIBUTION↗

The Economic Value of PV and Net Metering to Residential Customers in California

In this paper, we analyze the bill savings from PV for residential customers of the California's two largest electric utilities, under existing net metering tariffs as well as under several alternative compensation mechanisms. We find that economic value of PV to the customer is dependent on the structure of the underlying retail electricity rate and can vary quite significantly from one customer to another. In addition, we find that the value of the bill savings from PV generally declines with PV penetration level, as increased PV generation tends to offset lower-priced usage. Customers in our sample from both utilities are significantly better off with net metering than with a feed-in tariff where all PV generation is compensated at long-run avoided generation supply costs. Other compensation schemeswhich allow customers to displace their consumption with PV generation within each hour or each month, and are also based on the avoided costs, yield similar value to the customer as net metering.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Demonstration Assessment of Light Emitting Diode (LED) Commercial Garage Lights In the Providence Portland Medical Center, Portland, Oregon

This U.S. Department of Energy GATEWAY Demonstration project studied the applicability of light-emitting diode (LED) luminaires for commercial parking garage applications. High-pressure sodium (HPS) area luminaires were replaced with new LED area luminaires. The project was supported under the U.S. Department of Energy (DOE) Solid State Lighting Program. Other participants in the demonstration project included Providence Portland Medical Center in Portland, Oregon, the Energy Trust of Oregon, and Lighting Sciences Group (LSG) Inc. Pacific Northwest National Laboratory (PNNL) conducted the measurements and analysis of the results. PNNL manages GATEWAY demonstrations for DOE and represents their perspective in the conduct of the work. Quantitative and qualitative measurements of light and electrical power were taken at the site for both HPS and LED light sources. Economic costs were estimated and garage users’ responses to the new light sources were gauged with a survey. Six LED luminaires were installed in the below-ground parking level A, replacing six existing 150W HPS lamps spread out over two rows of parking spaces. Illuminance measurements were taken at floor level approximately every 4 ft on a 60-ft x 40-ft grid to measure light output of these LED luminaires which were termed the “Version 1” luminaires. PNNL conducted power measurements of the circuit in the garage to which the 6 luminaires were connected and determined that they drew an average of 82 W per lamp. An improved LED luminaire, Version 2, was installed in Level B of the parking garage. Illuminance measurements were not made of this second luminaire on site due to higher traffic conditions, but photometric measurements of this lamp and Version 1 were made in an independent testing laboratory and power usage for Version 2 was also measured. Version 1 was found to produce 3600 lumens and Version 2 was found to produce 4700 lumens of light and to consume 78 Watts. Maximum and minimum light levels were measured for the HPS and LED Version 1 luminaires and projected for the Version 2 luminaires. Maximum light levels were 23.51 foot candles, 20.54 fc, and 26.7 fc respectively and minimum light levels were 1.49 fc, 1.45 fc, and 1.88 fc. These results indicate very similar or even slightly higher light levels produced by the LED lamps, despite the higher lumen output of the HPS lamp. The LED lamps provide higher luminaire efficacy because all of the light is directed down and out. None of it is “lost” in the fixture. Also the HPS luminaire had poorly designed optics and a plastic covering that tended to get dirty and cracked, further decreasing the realized light output.[is this an accurate way to say this?] Consumer perceptions of the Version 2 LED were collected via a written survey form given to maintenance and security personnel. More than half felt the LED luminaires provided more light than the HPS lamps and a majority expressed a preference for the new lamps when viewing the relamped area through a security camera. Respondents commented that the LED luminaires were less glary, created less shadows, had a positive impact on visibility, and improved the overall appearance of the area. PNNL conducted an economic analysis and found that the Version 1 lamp produced annual energy savings of 955 kWh and energy cost savings of $76.39 per lamp at electricity rates of 6.5 cents per kWh and $105.03 at 11 cents per kWh. PNNL found that the Version 2 lamp produced annual energy savings of 991 kWh and energy cost savings of $79.26 per lamp at electricity rates of 6.5 cents per kWh and $108.98 at 11 cents per kWh. PNNL also calculated simple payback and found that Version 1 showed paybacks of 5.4 yrs at 6.5c/kWh and 4.1 yrs at 11c/kWh while Version 2 showed paybacks of 5.2 yrs at 6.5c/kWh and 3.9 yrs at 11c/kWh.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Optimization with the OpenACC-to-FPGA framework on the Arria 10 and Stratix 10 FPGAs

The reconfigurable computing paradigm with field programmable gate arrays (FPGAs) has received renewed interest in the high-performance computing field due to FPGAs’ unique combination of performance and energy efficiency. However, difficulties in programming and optimizing FPGAs have prevented them from being widely accepted as general-purpose computing devices. In accelerator-based heterogeneous computing, portability across diverse heterogeneous devices is also an important issue, but the unique architectural features in FPGAs make this difficult to achieve. To address these issues, a directive-based, high-level FPGA programming and optimization framework was previously developed. In this work, developed optimizations were combined holistically using the directive-based approach to show that each individual benchmark requires a unique set of optimizations to maximize performance. We perform this exploration on Intel Arria 10 and Stratix 10 FPGAs. We also explored the relationships between performance, resource usages, and compilation times, and investigated implications for performance portability. Finally, we present an initial evaluation of a real-world proxy application, LULESH.

97 MATHEMATICS AND COMPUTING↗