Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Performance optimization”

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

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

At least 1,639 records · Page 91
All sourcesNASA NTRSDOE OSTI
NASA NTRS2018

The Utilization Profiles of the CCSDS Unified Space Link Protocol (USLP)

The purpose of this paper is to identify the utilization profiles for interfacing the Data Protocol Sublayer using the Unified Space Link Protocols (USLP) (reference 1) with the space link coding procedures as specified in the CCSDS Coding & Synchronization Blue Books (references 2 through 5), used in both telecommand and telemetry applications. This paper describes how the USLP Protocol utilizes the coding and synchronization sublayer to support: a. Direct to Earth (DTE) telemetry links for engineering and science data b. Direct to Earth (DTE) telemetry links for very high rate science data c. Direct from Earth (DFE) command, sequencing and flight software loads d. Space to Space Links (Proximity) utilized by orbiters for data exchange to/from surface bound assets. The CCSDS has divided the functions of the Data Link Layer into two sublayers: the Data Link Protocol Sublayer (DLP-SL) and the Coding and Synchronization Sublayer (CS-SL). The Data Link Protocol Sublayer (DLP-SL) interfaces to the users, accepting the data that is to be transported, on the sending side of the link, and delivering that data on the receiving end. The Transfer Frame is the data unit that is transferred across the Data Link Protocol Sublayer and the Coding and Synchronization Sublayer boundary. The Coding and Synchronization Sublayer (CS-SL) provides the encoding, randomization, and frame synchronization functions that prepares the USLP Transfer Frame for transport across the space link. The CS-SL is divided into 2 processes: 1) The Frame Interface Processes (FIP) performs the interface functions required to prepare the data for delivery to the Coding/Decoding Process (CDP). This process includes prepending a Frame Start Marker to the provided frame, when management has designated that the frame is not to be aligned to the codeblock or when there is no block code used. 2) The Coding/Decoding Process (CDP) performs the forward error correction processes that are used to optimize the performance of the link and minimize the error rate. The CDP creates the symbol stream that is delivered to the Physical Layer. The transfer of the USLP transfer frames across different types of space links is the focus of this paper. The Protocol Data Unit (PDU) that is passed in both directions between the Data Link Protocol Sublayer (DLP-SL) and Coding and Synchronization Sublayer (CS-SL) is the transfer frame. The USLP frame structure provides flexibility that can be constrained by the functions utilized within the CS-SL that prepare the transfer frame for transit. For example, the USLP transfer frame contains a length field that enables the frame to be of variable length but CS-SL under certain conditions may constrain the frame to be fixed in length. This paper describes 5 operational modes available for use by the Data Link Layer to provide data exchange across the USLP space link. These modes are different because different operational requirements apply to vastly different types of space links and thus the communications implementation requirements differ. The environmental issues include the power or energy available, the distance between the end points of the link, the complexity of the equipment available at those end points, the atmospheric conditions and radiometric frequency selection. The CS-SL utilizes different forward error correcting codes supported by specific operational modes to configure the data for transit. This paper describes all of the operational modes in a series of data models which decompose the functionality between the Data Link Protocol Sublayer and the Coding and Synchronization sublayer. The operational modes described are: 1. Uncoded Mode: has been used for short links that contain significant available power to provide an acceptable frame error rate. The frames in this mode can be variable in length and typically use an error detection algorithm (i.e., CRC) to determine if there are errors in the received frame. 2. Convolutional Only Mode: is currently the prime forward error correction coding used for the proximity links. The frames in this mode can be variable in length and typically use an error detection algorithm (i.e., CRC) to determine if there are errors in the received frame. 3. Variable Length Frame Aligned to Variable Length Codeblock (TC): is used for Direct from Earth links were power levels are high and the simple, least complex code i.e., the BCH code is used. This mode has been in use since the early 1970s. The BCH code is a short code and the decoder is easy to implement. 4. Fixed Length Frame Aligned to Fixed Length Codeblock (AOS/TM): was introduced when the concatenated Convolutional and Reed-Solomon Code was formulated to provide significant reduction in link data error rate and the ability to determine if there was an error in the decoded codeblock. The frame is aligned to the codeblock so that there is a one to one relationship of frame errors to codeblock errors without additional error detection coding being added. This mode requires the protocol frames to be the exact size of the message portion of the codeblock. 5. Frames Unaligned to Fixed Length Codeblocks (Currently used for very high rates and space to space links): This mode is currently used for missions that have a very high data rate that can be controlled adaptively as the environment changes and as the next generation operating mode for the proximity link. This mode from a coded data stream point of view is exactly like that described in 4. above, except that the frame need not be aligned to the codeblock. There is no requirement on frame length when using this mode. Thus when using USLP it can be used to support links that require short or long frames. There is also no mandatory requirement that frames cannot be separated by idle data reducing the tight data rate connection requirements between the data link protocol sublayer and the coding & synchronization sublayer. In conclusion, how these operational modes can be put to use in mission operational scenarios is described for Direct from Earth links (DFE), Direct to Earth links (DTE), and Proximity links.

Greenberg, E.↗
DOE OSTI2019

Analysis of Cygnus Electrical Signals

The United States initiated the Subcritical Experiment (SCE) program to support a stockpile stewardship mission shortly after the 1992 moratorium on underground nuclear testing was established. Many SCE’s are conducted at the Nevada National Security Site (NNSS) in Nevada. Cygnus is a high-energy radiation generating device (RGD) located and operated at the NNSS and is a primary diagnostic for the SCE program. The Cygnus Dual Beam Radiographic Facility consists of two identical radiographic sources, Cygnus 1 and Cygnus 2. From creation of the high power V-I drive to energy transport and X-ray conversion at the rod-pinch diode, the Cygnus machines utilize the following components: oil-filled Marx generator, water-filled pulse-forming line (PFL), water-filled coaxial transmission line (CTL), three-cell vacuum induction voltage adder (IVA), and rod-pinch diode. The diode pulse has the following electrical specifications: 2.25 MV, 60 kA, 60 ns. Each source has the following X-ray specifications: 1 mm diameter, 4 rad at 1 m, and 50 ns radiation pulse. SCEs are both single-event and high-value; therefore, a high level of performance in reliability and reproducibility are key issues of Cygnus. Prior to executing such a SCE, there are a formidable number of shots (e.g., each of the two Cygnus RGDs charging and discharging properly into the rod-pinch diode load) that must be executed to determine reliability and reproducibility of the Cygnus RGDs. For every shot on Cygnus, voltages and currents along the machine are recorded and analyzed. In this paper we summarize attributes of the voltage and current waveforms at different locations using distribution plots. These distribution plots are used to quantify the reliability and reproducibility for Cygnus. This analysis methodology will provide a standard to trouble shoot results of individual Cygnus shots, and also to monitor long term Cygnus performance. Accordingly, this type of extensive analysis will contribute to optimized Cygnus performance on SCEs.

42 ENGINEERING↗
DOE OSTI2017

Study on Neutronics Calculation Method of SCWR with Annular Fuel Assembly

Supercritical Water-Cooled Reactor (SCWR) is one of Gen-IV reactor systems for its advantages over LWRs, such as high thermal efficiency, simplified reactor system and economics. Annular fuel was proposed by MIT to supplant conventional solid fuel and improve power density in LWRs. Annular fuel is cooled from both inside and outside of the fuel. Thus a lower fuel temperature and cladding surface temperature can be achieved with higher power density. SCWRs will benefit from this advantage because high cladding temperature (usually over 650 degree C) is one challenge of SCWRs. Using annular fuel in SCWRs is very promising. This is also because the inner hole of the annual fuel can provide extra neutron moderation. Supercritical water density is too low to provide enough moderation. Different analysis codes have been developed all over the world to carry out SCWR pre-conceptual core design. Yamaji et al. developed SRAC/ASMBURN/COREBN to study Japan version of SCWR. Monti et al. used ERANOS to analyze HPLWR (High Performance Light Water Reactor). WIMS-D4/CITATION was also used to calculate HPLWR by Tashakor et al.. These codes employ a two-step method used in PWRs calculation. Lattice/Assembly calculation is carried out as the first step. In the second step, few group constant of different fuel assembly types is then used in core calculation. MCNP had also been used to study SCWR by Watta et al.. Due to large computation time cost, it's used mostly to analyze fuel assembly or core performance rather than pre-conceptual design, in which much optimization is needed. In this research, a neutronics calculation code for SCWR is developed using two-step method. This code is verified by comparing with Monte-Carlo code. Fuel assembly homogenization method is studied to improve the code accuracy.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2024

Active Learning for Metamaterial Optimization on HPC and QC Integrated Systems

Active learning algorithms, integrating machine learning, quantum computing and optics simulation in an iterative loop, offer a promising approach to optimizing metamaterials. However, these algorithms can face difficulties in optimizing highly complex structures due to computational limitations. High-performance computing (HPC) and quantum computing (QC) integrated systems can address these issues by enabling parallel computing. In this study, we develop an active learning algorithm working on HPC-QC integrated systems. We evaluate the performance of optimization processes within active learning (i.e., training a machine learning model, problem-solving with quantum computing, and evaluating optical properties through wave-optics simulation) for highly complex metamaterial cases. Our results showcase that utilizing multiple cores on the integrated system can significantly reduce computational time, thereby enhancing the efficiency of optimization processes. Therefore, we expect that leveraging HPC-QC integrated systems helps effectively tackle large-scale optimization challenges in general.

Kim, Seongmin↗
DOE OSTI2022

ML4Geo: Machine Learning based Well Design to Enhance Unconventional Energy Production

This project aimed at using artificial intelligence and machine learning methods to improve oil and gas operations. On the machine learning front, this focused on the development of differentiable programming methods that are relevant in this application area, such as incorporating physical simulations into machine learning workflows. On the oil and gas front, the focus was on optimizing well performance.

02 PETROLEUM↗
DOE OSTI2022

Optimization of key energy and performance metrics for drug product manufacturing

During the development of pharmaceutical manufacturing processes, detailed systems-based analysis and optimization are required to control and regulate critical quality attributes within specific ranges, to maintain product performance. As discussions on carbon footprint, sustainability, and energy efficiency are gaining prominence, the development and utilization of these concepts in pharmaceutical manufacturing are seldom reported, which limits the potential of pharmaceutical industry in maximizing key energy and performance metrics. Based on an integrated modeling and techno-economic analysis framework previously developed by the authors, this study presents the development of a combined sensitivity analysis and optimization approach to minimize energy consumption while maintaining product quality and meeting operational constraints in a pharmaceutical process. The optimal input process conditions identified were validated against experiments and good agreement resulted between simulated and experimental data. Here, the results also allowed for a comparison of the capital and operational costs for batch and continuous manufacturing schemes under nominal and optimized conditions. Using the nominal batch operations as a basis, the optimized batch operation results in a 71.7% reduction of energy consumption, whereas the optimized continuous case results in an energy saving of 83.3%.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2018

Optimization, Inference and Learning for Advanced Networks

This piece of software aims at providing an implementation of methods to perform efficient optimization, probabilistic inference and model selection and learning for various type of networked systems. The software contains a suite of methods and functions that: 1) perform probabilistic inference and random sampling on networked models 2) select or learn networked models from data 3) solve optimization tasks with a networked structure Network models that are supported by these features include, but are not limited to, factor graphs and Markov random fields such as Ising models, Potts models, p-spins models, Gaussian random fields and restricted Boltzmann machine.

Hijazi, Hassan↗
DOE OSTI2019

PABO: Pseudo Agent-Based Multi-Objective Bayesian Hyperparameter Optimization for Efficient Neural Accelerator Design

The ever increasing computational cost of Deep Neural Networks (DNN) and the demand for energy efficient hardware for DNN acceleration has made accuracy and hardware cost co-optimization for DNNs tremendously important, especially for edge devices. Owing to the large parameter space and cost of evaluating each parameter in the search space, manually tuning of DNN hyperparameters is impractical. Automatic joint DNN and hardware hyperparameter optimization is indispensable for such problems. Bayesian optimization-based approaches have shown promising results for hyperparameter optimization of DNNs. However, most of these techniques have been developed without considering the underlying hardware, thereby leading to inefficient designs. Further, the few works that perform joint optimization are not generalizable and mainly focus on CMOS-based architectures. In this work, we present a novel pseudo agent-based multiobjective hyperparameter optimization (PABO) for maximizing the DNN performance while obtaining low hardware cost. Compared to the existing methods, our work poses a theoretically different approach for joint optimization of accuracy and hardware cost and focuses on memristive crossbar based accelerators. PABO uses a supervisor agent to establish connections between the posterior Gaussian distribution models of network accuracy and hardware cost requirements. The agent reduces the mathematical complexity of the co-optimization problem by removing unnecessary computations and updates of acquisition functions, thereby achieving significant speed-ups for the optimization procedure. PABO outputs a Pareto frontier that underscores the trade-offs between designing high-accuracy and hardware efficiency. Our results demonstrate a superior performance compared to the state-of-the-art methods both in terms of accuracy and computational speed (~100x speed up).

Parsa, Maryam↗
DOE OSTI2016

Global optimization algorithms to compute thermodynamic equilibria in large complex systems with performance considerations

Several global optimization methods are reviewed that attempt to ensure that the integral Gibbs energy of a closed isothermal isobaric system is a global minimum to satisfy the necessary and sufficient conditions for thermodynamic equilibrium. In particular, the integral Gibbs energy function of a multicomponent system containing non-ideal phases may be highly non-linear and non-convex, which makes finding a global minimum a challenge. Consequently, a poor numerical approach may lead one to the false belief of equilibrium. Furthermore, confirming that one reaches a global minimum and that this is achieved with satisfactory computational performance becomes increasingly more challenging in systems containing many chemical elements and a correspondingly large number of species and phases. Several numerical methods that have been used for this specific purpose are reviewed with a benchmark study of three of the more promising methods using five case studies of varying complexity. A modification of the conventional Branch and Bound method is presented that is well suited to a wide array of thermodynamic applications, including complex phases with many constituents and sublattices, and ionic phases that must adhere to charge neutrality constraints. Also, a novel method is presented that efficiently solves the system of linear equations that exploits the unique structure of the Hessian matrix, which reduces the calculation from a O(N 3 ) operation to a O(N) operation. As a result, this combined approach demonstrates efficiency, reliability and capabilities that are favorable for integration of thermodynamic computations into multi-physics codes with inherent performance considerations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Ceramic Composite Inert Matrix Fuel Forms in High-Temperature Gas-Cooled Microreactors

Here, this work optimizes micro-prismatic high-temperature gas reactor (HTGR) designs to reduce the energy-normalized mass of spent nuclear fuel (SNF) and high-level waste (HLW) produced. The optimization was performed for the current graphite moderator and an inert matrix fuel (IMF) concept employing different composite moderators in a prismatic design architecture. The fuel matrix is magnesium oxide (MgO) with entrained tristructural-isotropic (TRISO) fuel. The moderator materials, including beryllium oxide (MgO-BeO) and beryllium (MgO-Be) at 40 vol % loading and yttrium hydride (MgO-YH x=1.9 ) and zirconium hydride (MgO-ZrH x=1.9 ) at 15 vol % loading, were entrained within the MgO host matrix. A generic graphite micro-prismatic HTGR is used as the baseline point design where the external dimensions are held constant. The composite moderator designs use 19.9% enriched uranium nitride TRISO fuel and hexagonal assemblies. For each IMF concept, an optimization study was performed to maximize the discharge burnup of the fuel by varying the TRISO packing fraction and the lattice pitch of the assemblies. The mass of SNF and HLW, other waste metrics, fuel cost, environmental impact metrics, and the activity of the SNF and HLW at 100 years and 100 000 years were calculated for the optimized IMF and graphite reference designs. The IMF results were subsequently compared to those of the graphite reference and the values for a light water reactor (LWR) and a small modular LWR. For the SNF and HLW, all the IMF concepts and the graphite reference produced less waste compared to the traditional LWR designs. However, the IMF concepts outperformed the graphite reference regarding the mass of SNF and HLW. For the other waste metrics, the IMF concepts showed reductions in fuel cost with improved environmental metrics relative to the graphite reference. Overall, the IMF concepts significantly reduced the SNF and HLW produced per unit of energy generated compared to traditional LWR designs.

TRISO↗
NASA NTRS1980

Application of an advanced trajectory optimization method to ramjet propelled missiles

The mission performance characteristics of ramjet-propelled missiles are highly dependent upon the trajectory flown. Integration of the trajectory profile with the ramjet propulsion system performance characteristics to achieve optimal missile performance is very complex. Past trajectory optimization methods have been extremely problem dependent and require a high degree of familiarity to achieve success. A general computer code (CTOP) has been applied to ramjet-powered missiles to compute open-loop optimal trajectories. CTOP employs Chebyshev polynomial representations of the states and controls. This allows a transformation of the continuous optimal control problem to one of parameter optimization. With this method, the trajectory boundary conditions are always satisfied. State dynamics and path constraints are enforced via penalty functions. The presented results include solutions to minimum fuel-to-climb, minimum time-to-climb, and minimum time-to-target intercept problems.

Paris, S. W.↗
DOE OSTI2021

Apparatus and method for optimizing quantifiable behavior in configurable devices and systems

An apparatus and method are provided to perform constrained optimization of a constrained property of an apparatus, which is complex due to having several components, and these components are configurable in real-time. The optimization is achieved by detecting values of the constrained property and a plurality of other properties of the apparatus when the apparatus is configured in a first subset of the plurality of configurations. A model is learned using the detected values of the constrained property. The model represents the constrained property and can also represent other properties as a function of the configurations. The model can also include estimated uncertainties of the constrained property in the model. Then, using the d model and the estimated uncertainties, the optimal configuration can be selected to minimize an error value (e.g., the difference between a desired value and an observed value of the at least one constrained property).

Hoffmann, Henry↗
DOE OSTI2026

Crystal generation using the fully differentiable pipeline and latent space optimization

We present a materials generation framework that couples a symmetry-conditioned variational autoencoder with a differentiable SO(3) power spectrum objective to steer candidates toward a specified local environment under the crystallographic constraints. In particular, we implement a fully differentiable pipeline that performs batch-wise optimization on both direct and latent crystallographic representations. Using the GPU acceleration, the implementation achieves about fivefold speed compared to our previous CPU workflow, while yielding comparable outcomes. In addition, we introduce the optimization strategy that alternatively performs optimization on the direct and latent crystal representations. This dual-level relaxation approach can effectively escape local minima defined by different objective gradients, thus increasing the success rate of generating complex structures satisfying the target local environments. This framework can be extended to systems consisting of multi-components and multi-environments, providing a scalable route to generate material structures with the target local environment.

conditional VAE↗
DOE OSTI2022

Research and Test Reactor Fuels

PRO-RR is the research reactor focused program element of the broader Proliferation Resistance Optimization program (PRO-X) under the National Nuclear Safety Administration (NNSA) in the U.S. Department of Energy (DOE). PRO-X provides a framework for integrating proliferation resistance in nuclear system designs to minimize weapons usable nuclear materials (WUNM) production and diversion pathways while optimizing systems performance for peaceful use missions. PRO-RR applies the PRO-X mission objectives to research reactor system design. This document serves as one of the foundational documents for the PRO-RR-Fuel System Design technical team by documenting current research reactor fuels usage. The PRO-RR-Fuel System Design technical team consists of subject matter experts from Argonne National Laboratory (Argonne) and Savannah River National Laboratory (SRNL). In order to determine the preferred fuel of use in upcoming research and test reactors to optimize proliferation resistance, performance, and safety, it is useful to assess the fuels that have been used in the past, or are currently in use. This report reviews the historical and current fuels used in research and test reactors to inform future fuel selection. Chapter 2 discusses the low-enriched uranium (LEU) fuels currently in use in terms of thermal power level and utilization of the reactor. Chapter 3 summarizes the fabrication processes for common fuel types. Chapter 4 discusses in detail the fuel types in use in research and test reactors. A review of the cladding types in use is presented in Chapter 5, and a historical review of research and test reactor fuel fabricators is presented in Chapter 6. The data collection strategy used the International Atomic Energy Agency (IAEA) research reactor database [1] as a starting point. Information on the fuel used was gathered on research reactors (other than critical assemblies) that were listed as operational, planned, or in temporary shutdown in the IAEA database. Data on the fuel type, geometry, enrichment, uranium loading, cladding type, and fabricator were collected for each of the reactors available in the public domain. Sources of data included conference papers, journal articles, and facility and fabricator websites. Data on research reactors operating on LEU fuels are presented in Appendix A, while Appendix B presents data collected on all reactors at the time of publication of this report. Appendix C presents data collected on reactors that were part of the M3 research and test reactor conversion program.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
NASA NTRS1989

Mercury ion thruster technology

The Mercury Ion Thruster Technology program was an investigation for improving the understanding of state-of-the-art mercury ion thrusters. Emphasis was placed on optimizing the performance and simplifying the design of the 30 cm diameter ring-cusp discharge chamber. Thruster performance was improved considerably; the baseline beam-ion production cost of the optimized configuration was reduced to Epsilon (sub i) perspective to 130 eV/ion. At a discharge propellant-utilization efficiency of 95 percent, the beam-ion production cost was reduced to about 155 eV/ion, representing a reduction of about 40 eV/ion over the corresponding value for the 30 cm diameter J-series thruster. Comprehensive Langmuir-probe surveys were obtained and compared with similar measurements for a J-series thruster. A successful volume-averaging scheme was developed to correlate thruster performance with the dominant plasma processes that prevail in the two thruster designs. The average Maxwellian electron temperature in the optimized ring-cusp design is as much as 1 eV higher than it is in the J-series thruster. Advances in ion-extraction electrode fabrication technology were made by improving materials selection criteria, hydroforming and stress-relieving tooling, and fabrications procedures. An ion-extraction performance study was conducted to assess the effect of screen aperture size on ion-optics performance and to verify the effectiveness of a beam-vectoring model for three-grid ion optics. An assessment of the technology readiness of the J-series thruster was completed, and operation of an 8 cm IAPS thruster using a simplified power processor was demonstrated.

Beattie, J. R.↗
DOE OSTI2016

Comparative Performance Analysis of Coarse Solvers for Algebraic Multigrid on Multicore and Manycore Architectures

In this paper, we study the performance of a two-level algebraic-multigrid algorithm, with a focus on the impact of the coarse-grid solver on performance. We consider two algorithms for solving the coarse-space systems: the preconditioned conjugate gradient method and a new robust HSS-embedded low-rank sparse-factorization algorithm. Our test data comes from the SPE Comparative Solution Project for oil-reservoir simulations. We contrast the performance of our code on one 12-core socket of a Cray XC30 machine with performance on a 60-core Intel Xeon Phi coprocessor. To obtain top performance, we optimized the code to take full advantage of fine-grained parallelism and made it thread-friendly for high thread count. We also developed a bounds-and-bottlenecks performance model of the solver which we used to guide us through the optimization effort, and also carried out performance tuning in the solver’s large parameter space. Finally, as a result, significant speedups were obtained on both machines.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Optimizing Asynchronous Multi-Level Checkpoint/Restart Configurations with Machine Learning

With the emergence of versatile storage systems, multi-level checkpointing (MLC) has become a common approach to gain efficiency. However, multi-level checkpoint/restart can cause enormous I/O traffic on HPC systems. To use multilevel checkpointing efficiently, it is important to optimize check-point/restart configurations. Current approaches, namely modeling and simulation, are either inaccurate or slow in determining the optimal configuration for a large scale system. In this paper, we show that machine learning models can be used in combination with accurate simulation to determine the optimal checkpoint configurations. We also demonstrate that more advanced techniques such as neural networks can further improve the performance in optimizing checkpoint configurations.

checkpoint-restart↗
DOE OSTI2024

ytopt: Autotuning Scientific Applications for Energy Efficiency at Large Scales

As we enter the exascale computing era, efficiently utilizing power and optimizing the performance of scientific applications under power and energy constraints has become critical and challenging. We propose a low-overhead autotuning framework to autotune performance and energy for various hybrid MPI/OpenMP scientific applications at large scales and to explore the tradeoffs between application runtime and power/energy for energy efficient application execution, then use this framework to autotune four ECP proxy applications—XSBench, AMG, SWFFT, and SW4lite. Our approach uses Bayesian optimization with a Random Forest surrogate model to effectively search parameter spaces with up to 6 million different configurations on two large-scale HPC production systems, Theta at Argonne National Laboratory and Summit at Oak Ridge National Laboratory. The experimental results show that our autotuning framework at large scales has low overhead and achieves good scalability. Using the proposed autotuning framework to identify the best configurations, we achieve up to 91.59% performance improvement, up to 21.2% energy savings, and up to 37.84% EDP (energy delay product) improvement on up to 4096 nodes.

Autotuning↗