Search NASA⌕ Search

SEARCH · Search NASA

Results for “Process optimization and mapping”

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 OSTI2023

Multi-level optimization with the koopman operator for data-driven, domain-aware, and dynamic system security

Cyber-Physical Systems (CPSs) like the power grid are critically important but also increasingly vulnerable; ensuring reliable system operation in the face of disruptions is becoming more and more challenging. Multi-Level Optimization (MLO) is a powerful way to model adversarial interactions, which naturally makes it applicable to studying CPS security. However, MLO typically does not address underlying system dynamics, and incorporating nonlinear dynamics is generally infeasible. In this paper, we show how to combine MLO with the Koopman Operator (KO) to remedy this. The KO maps nonlinear dynamics to a lifted space in which those dynamics are linear, thus making it ideal for use with MLO. Moreover, the structure of the KO also provides convenient ways to incorporate domain knowledge into the data-driven process of learning the KO representation of a given system. Here we then demonstrate the use of MLO-KO on a small example problem taken from the power grid domain, discuss the scalability and computational cost of MLO-KO, and identify future research directions for this work.

42 ENGINEERING↗
DOE OSTI2022

Invertible Neural Networks for Airfoil Design

We report the airfoil design problem, in which an engineer seeks a shape with desired performance characteristics, is fundamental to aerodynamics. Design workflows traditionally rely on iterative optimization methods using low-fidelity integral boundary-layer methods as higher-fidelity adjoint-based computational fluid dynamics methods are computationally expensive. Surrogate-based approaches can accelerate the design process but still rely on some iterative inverse design procedure. In this work, we leverage emerging invertible neural network (INN) tools to enable the rapid inverse design of airfoil shapes for wind turbines. INNs are specialized deep-learning models with well-defined inverse mappings. When trained appropriately, INN surrogate models are capable of forward prediction of aerodynamic and structural quantities for a given airfoil shape as well as inverse recovery of airfoil shapes with specified aerodynamic and structural characteristics. The INN approach offers a roughly 100 times speed-up compared to adjoint-based methods for inverse design. We demonstrate the INN tool for inverse design on three test cases of 100 airfoils each that satisfy the performance characteristics close to those of airfoils used in wind-turbine blades. All generated shapes satisfy the desired aerodynamic characteristics, demonstrating the success of the INN approach for inverse design of airfoils.

17 WIND ENERGY↗
DOE OSTI2026

Stakeholder-guided holistic, Adaptive Framework for enhancing community Energy Resilience (SAFER) (Final Technical Report)

The Stakeholder-guided holistic, Adaptive Framework for enhancing community Energy Resilience (SAFER) project advances resilience science and engineering by addressing challenges in rural Kansas communities where aging infrastructure, extreme weather, and socioeconomic disparities heighten vulnerability to energy disruptions. Traditional approaches often focus on technical performance while overlooking community concerns and priorities. SAFER responds by integrating community perspectives with advanced analytical frameworks to create a holistic model for measuring and improving resilience. Project objectives included developing novel resilience metrics, advancing modeling frameworks that capture interdependencies across infrastructures, and embedding community-centric indicators directly into planning processes for distributed energy resources. The key technical innovations included the creation of self-organizing map (SOM)-based indices for objective resilience quantification, hetero-functional graph theory (HFGT) models linking power, water, transportation, and community assets, and graph neural network (GNN) tools for identifying critical nodes in complex systems. Community-centric energy planning was demonstrated through optimal siting and sizing of (photovoltaic) PV and battery storage, ensuring resilience enhancements also addressed energy burden and energy insecurity. SAFER engaged community partners in Dodge City and Ford County through surveys, focus groups, and workshops, generating more than 600 responses that established baseline measures of energy burden, financial insecurity, and willingness-to-pay to avoid outages. This data, organized in terms of a community capitals framework, informed the development of weighted reliability indices that better reflect community costs than traditional utility metrics. SAFER’s GNN-based critical node identification framework identified expert-labelled critical nodes with over 99% accuracy, while also uncovering additional functionalities essential for proactive resilience planning. The project’s models demonstrated that optimal PV and storage deployment could improve resilience indices by over 11 percent, with dispatch strategies further enhancing outcomes, confirming both the technical effectiveness and economic feasibility of these approaches. Through its combined emphasis on rigorous modeling, community-focused planning, and community engagement, SAFER advances the state of resilience research while delivering direct benefits to rural communities. The project provides tools, guidelines, and resilience heatmaps that help utilities, local governments, and residents better anticipate disruptions, prioritize investments, and strengthen the capacity to withstand and recover from energy-related hazards. Furthermore, the developed HFG and GNN frameworks are designed for transferability, allowing them to be adapted for resilience planning in other communities with minimal retraining. This inductive learning capability provides a scalable pathway to extend the SAFER project’s impact. Thus, creating a foundation for a nationally applicable model of infrastructure resilience. Additionally, the HFG can also be extended to include other FEMA community lifelines.

14 SOLAR ENERGY↗
DOE OSTI2022

Optimization of stochastic feature properties in laser powder bed fusion

Process parameter selection in laser powder bed fusion (LPBF) controls the as-printed dimensional tolerances, pore formation, surface quality and microstructure of printed metallic structures. Measuring the stochastic mechanical performance for a wide range of process parameters is cumbersome both in time and cost. As such, in this study, we overcome these hurdles by using high-throughput tensile (HTT) testing of over 250 dogbone samples to examine process-driven performance of strut-like small features, ~1 mm 2 in austenitic stainless steel (316 L). The output mechanical properties, porosity, surface roughness and dimensional accuracy were mapped across the printable range of laser powers and scan speeds using a continuous wave laser LPBF machine. Tradeoffs between ductility and strength are shown across the process space and their implications are discussed. While volumetric energy density deposited onto a substrate to create a melt-pool can be a useful metric for determining bulk properties, it was not found to directly correlate with output small feature performance.

316 L stainless steel↗
DOE OSTI2024

Multi-objective Bayesian active learning for MeV-ultrafast electron diffraction

Ultrafast electron diffraction using MeV energy beams(MeV-UED) has enabled unprecedented scientific opportunities in the study of ultrafast structural dynamics in a variety of gas, liquid and solid state systems. Broad scientific applications usually pose different requirements for electron probe properties. Due to the complex, nonlinear and correlated nature of accelerator systems, electron beam property optimization is a time-taking process and often relies on extensive hand-tuning by experienced human operators. Algorithm based efficient online tuning strategies are highly desired. Here, we demonstrate multi-objective Bayesian active learning for speeding up online beam tuning at the SLAC MeV-UED facility. The multi-objective Bayesian optimization algorithm was used for efficiently searching the parameter space and mapping out the Pareto Fronts which give the trade-offs between key beam properties. Such scheme enables an unprecedented overview of the global behavior of the experimental system and takes a significantly smaller number of measurements compared with traditional methods such as a grid scan. This methodology can be applied in other experimental scenarios that require simultaneously optimizing multiple objectives by explorations in high dimensional, nonlinear and correlated systems.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

TEAMER: Pitching Foil Crossflow Turbine Efficiency Data

This dataset documents the efficiency testing of a pitching foil crossflow turbine, conducted at the University of New Hampshire's (UNH) Chase Ocean Engineering Laboratory tow tank facility. The tests explored various pitch phases and amplitudes, ranging from 0 to 18 degrees, across different flow speeds and turbine RPMs. Specifically, testing was performed at pitch amplitudes of 0,3,6,9,12,15 and 18 degrees. The results indicate that turbine efficiency improves with increased pitch amplitude, reaching optimal performance at 12 degrees. Modifications were made to the test frame by UNH to enhance the stiffness of the support structure. The testing specifically mapped the efficiency of a single quadrant of the Ocean Renewable Power Company (ORPC) Autonomous Turbine Generator Unit (ATGU), covering the full range of allowable pitching amplitudes and phases. The dataset provides both raw and processed data, including drag and torque measurements, and contains Python scripts used for data processing and visualization. This research was supported by funding from TEAMER RFTS 1 (Request for Technical Support).

16 TIDAL AND WAVE POWER↗
DOE OSTI2022

HFTS-2 (Final Report)

This report delves into the research program, identified tasks, various studies, observations, results, and conclusions under the Hydraulic Fracturing Test Site (HFTS-2) project. This test site is located in Loving County close to the Texas, New Mexico border. The program targeted the Wolfcamp formation in Delaware Basin. The aim of the project was to improve our understanding of hydraulic fracturing processes to not only optimize unconventional resource development in US Basins but to minimize environmental footprint as well. Various new diagnostic technologies were demonstrated at this site, including distributed acoustic microseismic, high resolution strain using Rayleigh frequency band (DSS-RFS), and in-fill proppant logging tool. Significant conclusions from this study include: 1) lateral and vertical fracture growth mapping as well as proppant transport behavior, 2) Use of DSS-RFS as a diagnostic tool to understand cluster level behavior, 3) Validation of proppant log, both at core scale as well as in-fill well implementation, and 4) New modeling tools to map far-field strain response using distributed strain sensing data. Specifically, far-field fracture distribution is influenced by linear fracture corridors defined by the stress-state as well as clustering behavior influenced by spatio-temporal effects. Depletion impacts on new stimulation was demonstrated and technically quantified. Significant vertical fracture height growth was observed. However, our understanding is limited by available gauge data availability. Numerical strain modeling techniques have been developed and demonstrated to help understand observed strain response behavior. Finally, proppant log has been further validated as a useful tool to map propped zones within the SRV. It has also shown very strong correlation with observed drawdown behavior at this test site. This report will detail many of these studies as well as technical results and conclusions from said studies. These include general subsurface characterization work, completion designs as implemented, studies looking at fracture geometry, high resolution microseismic study including source mechanisms, advanced DSS-RFS studies including drainage characterization, core characterization results, proppant analysis (core and in-fill child well), geochemistry, etc.

02 PETROLEUM↗
DOE OSTI2020

BeyondPlanck I. Global Bayesian analysis of the Planck Low Frequency Instrument data

We describe the BeyondPlanck project in terms of motivation, methodology and main products, and provide a guide to a set of companion papers that describe each result in fuller detail. Building directly on experience from ESA's Planck mission, we implement a complete end-to-end Bayesian analysis framework for the Planck Low Frequency Instrument (LFI) observations. The primary product is a joint posterior distribution P(omega|d), where omega represents the set of all free instrumental (gain, correlated noise, bandpass etc.), astrophysical (synchrotron, free-free, thermal dust emission etc.), and cosmological (CMB map, power spectrum etc.) parameters. Some notable advantages of this approach are seamless end-to-end propagation of uncertainties; accurate modeling of both astrophysical and instrumental effects in the most natural basis for each uncertain quantity; optimized computational costs with little or no need for intermediate human interaction between various analysis steps; and a complete overview of the entire analysis process within one single framework. As a practical demonstration of this framework, we focus in particular on low-l CMB polarization reconstruction, paying special attention to the LFI 44 GHz channel. We find evidence of significant residual systematic effects that are still not accounted for in the current processing, but must be addressed in future work. These include a break-down of the 1/f correlated noise model at 30 and 44 GHz, and scan-aligned stripes in the Southern Galactic hemisphere at 44 GHz. On the Northern hemisphere, however, we find that all results are consistent with the LCDM model, and we constrain the reionization optical depth to tau = 0.067 +/- 0.016, with a low-resolution chi-squared probability-to-exceed of 16%. The marginal CMB dipole amplitude is 3359.5 +/- 1.9 uK. (Abridged.)

Andersen, KJ↗
DOE OSTI2021

Sustainable Lactic Acid Production from Lignocellulosic Biomass

Lignocellulosic biomass is a promising feedstock for sustainable biofuels and bioproducts. Among emerging bioproducts, lactic acid has attracted significant interest because of its growing application in many industries (e.g., packaging, medical, and pharmaceutical). In this study, BioSTEAM—an open-source platform—was leveraged for the design, simulation, and evaluation (via techno-economic analysis, TEA, and life cycle assessment, LCA) of lignocellulosic lactic acid biorefineries. With a minimum product selling price (MPSP) between 1.38 and 1.91 kg –1 (5th–95th percentiles, baseline at 1.57 kg –1 ), the biorefinery was capable of producing market-competitive lactic acid (market price between 1.7 and 2.1 kg –1 ), and its performance could be further enhanced (e.g., MPSP down to 1.09 kg –1 , global warming potential of 2.79 kg CO 2 -eq·kg –1 , and fossil energy consumption of 31.7 MJ·kg –1 ) with advancements in key technological parameters (fermentation yield and separation process conversions) and optimization in process operation. Sensitivity analyses focused on the fermentation unit (across titer, yield, and productivity; neutral vs low-pH fermentation) and feedstock characteristics (carbohydrate content and price) were also included to quantify their impact on the sustainability of the biorefinery. Overall, this research highlights the ability of agile TEA/LCA to screen promising biorefinery designs, prioritize research needs, and establish a road map for the continued development of bioproducts and biofuels.

09 BIOMASS FUELS↗
DOE OSTI2025

Critical statistical assessment of data in metal additive manufacturing

Obtaining high quality data reflecting the relationships between the additive manufacturing (AM) process parameters, material microstructure and mechanical properties is crucial for the use of machine learning in AM. A database of over 4,000 data entries of metal AM was created thanks to a large number of literature studies on key process parameters and indicators of build quality. Meta-analysis reveals critical biases in the literature. Firstly, majority of studies report only high quality builds, these imbalances in reporting result in weak correlation between process parameters, properties and consolidation, limiting the ability of machine learning models to generalize beyond optimized conditions. Nevertheless, the trained models accurately predict yield strength ($R^2 = 0.85$), suggesting that certain process–property relationships are effectively captured within these models. Secondly, quantitative microstructural data are largely absent, limiting the learning of the microstructure-mechanical properties relationships. Finally, current process window identification is based largely on the consolidation, despite significant uncertainty in its measurement. It is important to identify the process map on the basis of not only the consolidation, but also mechanical behaviour under loading. Such a identification shows that 316 L and Inconel have much larger process map (i.e. highly printable) in comparison to the AlSi10Mg and Ti6Al4V.

Additive manufacturing↗
DOE OSTI2021

Latent map Gaussian processes for mixed variable metamodeling

Gaussian processes (GPs) are ubiquitously used in sciences and engineering as metamodels. Standard GPs, however, can only handle numerical or quantitative variables. Here we introduce latent map Gaussian processes (LMGPs) that inherit the attractive properties of GPs and are also applicable to mixed data which have both quantitative and qualitative inputs. The core idea behind LMGPs is to learn a continuous, low-dimensional latent space or manifold which encodes all qualitative inputs. To learn this manifold, we first assign a unique prior vector representation to each combination of qualitative inputs. We then use a low-rank linear map to project these priors on a manifold that characterizes the posterior representations. As the posteriors are quantitative, they can be directly used in any standard correlation function such as the Gaussian or Matern. Hence, the optimal map and the corresponding manifold, along with other hyperparameters of the correlation function, can be systematically learned via maximum likelihood estimation. Through a wide range of analytic and real-world examples, we demonstrate the advantages of LMGPs over state-of-the-art methods in terms of accuracy and versatility. In particular, we show that LMGPs can handle variable-length inputs, have an explainable neural network interpretation, and provide insights into how qualitative inputs affect the response or interact with each other. We also employ LMGPs in Bayesian optimization and illustrate that they can discover optimal compound compositions more efficiently than conventional methods that convert compositions to qualitative variables via manual featurization.

42 ENGINEERING↗
DOE OSTI2021

Case Study of Using Kokkos and SYCLs Performance-Portable Frameworks for Milc-Dslash Benchmark on NVIDIA, AMD and Intel GPUs

Six of the top ten supercomputers in the TOP500 list from June 2021 rely on NVIDIA GPUs to achieve their peak compute bandwidth. With the announcement of Aurora, Frontier, and El Capitan, Intel and AMD have also entered the domain of providing GPUs for scientific computing. A consequence of the increased diversity in the GPU landscape is the emergence of portable programming models such as Kokkos, SYCL, OpenCL, and OpenMP, which allow application developers to maintain a single-source code across a diverse range of hardware architectures. While the portable frameworks try to optimize the compute resource usage on a given architecture, it is the programmers responsibility to expose parallelism in an application that can take advantage of thousands of processing elements available on GPUs. In this paper, we introduce a GPU-friendly parallel implementation of Milc-Dslash that exposes multiple hierarchies of parallelism in the algorithm. Milc-Dslash was designed to serve as a benchmark with highly optimized matrix-vector multiplications to measure the resource utilization on the GPU systems. The parallel hierarchies in the Milc-Dslash algorithm are mapped onto a target hardware using Kokkos and SYCL programming models. We present the performance achieved by Kokkos and SYCL implementations of Milc-Dslash on NVIDIA A100 GPU, AMD MI100 GPU, and Intel Gen9 GPU. Additionally, we compare the Kokkos and SYCL performances with those obtained from the versions written in CUDA and HIP programming models on NVIDIA A100 GPU and AMD MI100 GPU, respectively.

Dufek, Amanda S↗
DOE OSTI2026

Extremized nonlinear and linearized responses in soft metamaterials enabled by gradient-based design and grayscale digital light processing

In this study, we develop a gradient-based design approach that exploits grayscale digital light processing (DLP) 3D printing for extremizing the nonlinear and linearized response of soft metamaterials — materials that harness engineered geometric instabilities to undergo large and programmable changes in configuration. Grayscale DLP approaches modulate local mechanical properties at the pixel scale by tuning the light intensity within a single grayscale image, unlocking an exceptionally large design space. To effectively navigate this space, we develop smooth mappings between local light intensity values and global quantities of interest that characterize the behavior of soft metamaterials. Enabling these smooth mappings are robust and differentiable nonlinear finite element simulations powered by a trust region solver. A PDE-constrained optimization problem is then solved to invert these mappings and produce light intensity distributions that endow the printed part with varying stiffness and flexibility in distinctive regions. It is shown that optimizing the distribution of soft and stiff phases throughout a metamaterial structure results in markedly different buckling and self-contact configurations to drive extremized nonlinear compression and linearized vibration responses. Optimized light intensity distributions are translated to grayscale images and directly used to print soft metamaterial samples, showing remarkable agreement between the buckling and self-contact response in simulated and measured deformed configurations.

Additive manufacturing↗
DOE OSTI2023

BeyondPlanck: I. Global Bayesian analysis of the Planck Low Frequency Instrument data

We describe the BEYONDPLANCK project in terms of our motivation, methodology, and main products, and provide a guide to a set of companion papers that describe each result in more detail. Building directly on experience from ESA’s Planck mission, we implemented a complete end-to-end Bayesian analysis framework for the Planck Low Frequency Instrument (LFI) observations. The primary product is a full joint posterior distribution P(ω | d), where ω represents the set of all free instrumental (gain, correlated noise, bandpass, etc.), astrophysical (synchrotron, free-free, thermal dust emission, etc.), and cosmological (cosmic microwave background – CMB – map, power spectrum, etc.) parameters. Some notable advantages of this approach compared to a traditional pipeline procedure are seamless end-to-end propagation of uncertainties; accurate modeling of both astrophysical and instrumental effects in the most natural basis for each uncertain quantity; optimized computational costs with little or no need for intermediate human interaction between various analysis steps; and a complete overview of the entire analysis process within one single framework. As a practical demonstration of this framework, we focus in particular on low-ℓ CMB polarization reconstruction with Planck LFI. In this process, we identify several important new effects that have not been accounted for in previous pipelines, including gain over-smoothing and time-variable and non-1/f correlated noise in the 30 and 44 GHz channels. Modeling and mitigating both previously known and newly discovered systematic effects, we find that all results are consistent with the ΛCDM model, and we constrained the reionization optical depth to τ = 0.066 ± 0.013, with a low-resolution CMB-based χ 2 probability to exceed of 32%. This uncertainty is about 30% larger than the official pipelines, arising from taking a more complete instrumental model into account. The marginal CMB solar dipole amplitude is 3362.7 ± 1.4 μK, where the error bar was derived directly from the posterior distribution without the need of any ad hoc instrumental corrections. We are currently not aware of any significant unmodeled systematic effects remaining in the Planck LFI data, and, for the first time, the 44 GHz channel is fully exploited in the current analysis. We argue that this framework can play a central role in the analysis of many current and future high-sensitivity CMB experiments, including LiteBIRD, and it will serve as the computational foundation of the emerging community-wide COSMOGLOBE effort, which aims to combine state-of-the-art radio, microwave, and submillimeter data sets into one global astrophysical model.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Optimizing long-term monitoring of radiation air-dose rates after the Fukushima Daiichi Nuclear Power Plant

Radiation air dose rates near the Fukushima Daiichi Nuclear Power Plant (FDNPP) have been steadily decreasing over the past eight years since the release of radioactive elements in March 2011. Currently, the radiation monitoring program is expected to transition to long-term monitoring after most of the remediation activities are completed. The main long-term monitoring objectives are to (1) confirm the continuing reduction of contaminant and hazard levels, (2) provide assurance for the public, (3) accumulate the basic datasets for scientific knowledge and future preparation, and (4) detect changes or anomalies in contaminant mobility (if they occur), or any unexpected processes or events. In this work, we have developed a methodology for optimizing the monitoring locations of radiation air dose-rate monitoring. Our approach consists of three steps in order to determine monitoring locations in a systematic manner: (1) prioritizing the critical locations, such as schools or regulatory requirement locations, (2) diversifying locations that cover the key environmental controls that are known to influence contaminant mobility and distributions, and (3) capturing the heterogeneity of radiation air-dose rates across the domain. Therefore, for the second step, we use a Gaussian mixture model to identify the representative locations among multiple environmental variables, such as elevation and land-cover types. For the third step, we use a Gaussian process model to capture and estimate the heterogeneity of air-dose rates across the domain. Employing an integrated dose-rate map derived from Bayesian geostatistical methods as a reference map, we distribute the monitoring locations in such a way as to capture the heterogeneity of the reference map. Our results have shown that this approach allows us to select monitoring locations in a systematic manner such that the heterogeneity of air dose rates is captured by the minimal number of monitoring locations.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2021

Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication.

There is a growing interest in custom spatial accelerators for machine learning applications. These accelerators employ a spatial array of processing elements (PEs) interacting via custom buffer hierarchies and networks-on-chip. The efficiency of these accelerators comes from employing optimized dataflow (i.e., spatial/temporal partitioning of data across the PEs and fine-grained scheduling) strategies to optimize data reuse. The focus of this work is to evaluate these accelerator architectures using a tiled general matrix-matrix multiplication (GEMM) kernel. To do so, we develop a framework that finds optimized mappings (dataflow and tile sizes) for a tiled GEMM for a given spatial accelerator and workload combination, leveraging an analytical cost model for runtime and energy. Finally, our evaluations over five spatial accelerators demonstrate that the tiled GEMM mappings systematically generated by our framework achieve high performance on various GEMM workloads and accelerators.

42 ENGINEERING↗
DOE OSTI2022

Considerations in upconversion: A practical guide to sum-frequency generation spectrometer design and implementation

Here, in this tutorial review, we discuss how the choice of upconversion pulse shape in broadband vibrational sum-frequency generation (SFG) spectrometer design impacts the chemical or physical insights one can obtain from a set of measurements. A time-domain picture of a vibrational coherence being mapped by a second optical field is described and the implications of how this mapping, or upconversion process, takes place are given in the context of several popular and emerging approaches found in the literature. Emphasis is placed on broadband frequency-domain measurements, where the choice of upconversion pulse enhances or limits the information contained in the SFG spectrum. We conclude with an outline for a flexible approach to SFG upconversion using pulse-shaping methods and a simple guide to design and optimize the associated instrumentation.

47 OTHER INSTRUMENTATION↗
DOE OSTI2021

Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication

There is a growing interest in custom spatial accelerators for machine learning applications. These accelerators employ a spatial array of processing elements (PEs) interacting via custom buffer hierarchies and networks-on-chip. The efficiency of these accelerators comes from employing optimized dataflow (i.e., spatial/temporal partitioning of data across the PEs and fine-grained scheduling) strategies to optimize data reuse. The focus of this work is to evaluate these accelerator architectures using a tiled general matrix-matrix multiplication (GEMM) kernel. To do so, we develop a framework that finds optimized mappings (dataflow and tile sizes) for a tiled GEMM for a given spatial accelerator and workload combination, leveraging an analytical cost model for runtime and energy. Our evaluations over five spatial accelerators demonstrate that the tiled GEMM mappings systematically generated by our framework achieve high performance on various GEMM workloads and accelerators.

43 PARTICLE ACCELERATORS↗