Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “kernel”

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 1,243 records · Page 69
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Unprecedented cloud resolution in a GPU-enabled full-physics atmospheric climate simulation on OLCF’s summit supercomputer

Clouds represent a key uncertainty in future climate projection. While explicit cloud resolution remains beyond our computational grasp for global climate, we can incorporate important cloud effects through a computational middle ground called the Multi-scale Modeling Framework (MMF), also known as Super Parameterization. This algorithmic approach embeds high-resolution Cloud Resolving Models (CRMs) to represent moist convective processes within each grid column in a Global Climate Model (GCM). The MMF code requires no parallel data transfers and provides a self-contained target for acceleration. This study investigates the performance of the Energy Exascale Earth System Model-MMF (E3SM-MMF) code on the OLCF Summit supercomputer at an unprecedented scale of simulation. Hundreds of kernels in the roughly 10K lines of code in the E3SM-MMF CRM were ported to GPUs with OpenACC directives. A high-resolution benchmark using 4600 nodes on Summit demonstrates the computational capability of the GPU-enabled E3SM-MMF code in a full physics climate simulation.

58 GEOSCIENCES↗
DOE OSTI2021

EXAGRAPH: Graph and combinatorial methods for enabling exascale applications

Combinatorial algorithms in general and graph algorithms in particular play a critical enabling role in numerous scientific applications. However, the irregular memory access nature of these algorithms makes them one of the hardest algorithmic kernels to implement on parallel systems. With tens of billions of hardware threads and deep memory hierarchies, the exascale computing systems in particular pose extreme challenges in scaling graph algorithms. The codesign center on combinatorial algorithms, ExaGraph, was established to design and develop methods and techniques for efficient implementation of key combinatorial (graph) algorithms chosen from a diverse set of exascale applications. Algebraic and combinatorial methods have a complementary role in the advancement of computational science and engineering, including playing an enabling role on each other. In this paper, we survey the algorithmic and software development activities performed under the auspices of ExaGraph from both a combinatorial and an algebraic perspective. In particular, we detail our recent efforts in porting the algorithms to manycore accelerator (GPU) architectures. We also provide a brief survey of the applications that have benefited from the scalable implementations of different combinatorial algorithms to enable scientific discovery at scale. We believe that several applications will benefit from the algorithmic and software tools developed by the ExaGraph team.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

End-to-end GPU acceleration of low-order-refined preconditioning for high-order finite element discretizations

In this article, we present algorithms and implementations for the end-to-end GPU acceleration of matrix-free low-order-refined preconditioning of high-order finite element problems. The methods described here allow for the construction of effective preconditioners for high-order problems with optimal memory usage and computational complexity. The preconditioners are based on the construction of a spectrally equivalent low-order discretization on a refined mesh, which is then amenable to, for example, algebraic multigrid preconditioning. The constants of equivalence are independent of mesh size and polynomial degree. For vector finite element problems in H(curl) and H(div) (e.g., for electromagnetic or radiation diffusion problems), a specially constructed interpolation–histopolation basis is used to ensure fast convergence. Detailed performance studies are carried out to analyze the efficiency of the GPU algorithms. The kernel throughput of each of the main algorithmic components is measured, and the strong and weak parallel scalability of the methods is demonstrated. The different relative weighting and significance of the algorithmic components on GPUs and CPUs is discussed. Results on problems involving adaptively refined nonconforming meshes are shown, and the use of the preconditioners on a large-scale magnetic diffusion problem using all spaces of the finite element de Rham complex is illustrated.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Numerical eigen-spectrum slicing, accurate orthogonal eigen-basis, and mixed-precision eigenvalue refinement using OpenMP data-dependent tasks and accelerator offload

Performing a variety of numerical computations efficiently and, at the same time, in a portable fashion requires both an overarching design followed by a number of implementation strategies. All of these are exemplified below as we present transitioning the PLASMA numerical library from relying on dependence-driven large tasks to achieving utilization of fine grain tasking and offload to hardware accelerators while keeping its core dependence sets: OpenMP source code pragmas and runtime for most system-level functionality and basic low-level numerical kernels provided directly by hardware vendors or open source projects with vendor contributions. We also present new algorithmic methods and their efficient parallel implementations including fine grained tasking for eigen-spectrum slicing and offload for mixed-precision eigenvalue refinement. We provide performance, scaling, and numerical results showing sizable gains over the available solutions from either the open source and vendor-provided packages.

Luszczek, Piotr↗
DOE OSTI2024

A graphics processing unit accelerated sparse direct solver and preconditioner with block low rank compression

We present the GPU implementation efforts and challenges of the sparse solver package STRUMPACK. The code is made publicly available on github with a permissive BSD license. STRUMPACK implements an approximate multifrontal solver, a sparse LU factorization which makes use of compression methods to accelerate time to solution and reduce memory usage. Multiple compression schemes based on rank-structured and hierarchical matrix approximations are supported, including hierarchically semi-separable, hierarchically off-diagonal butterfly, and block low rank. Here, in this paper, we present the GPU implementation of the block low rank (BLR) compression method within a multifrontal solver. Our GPU implementation relies on highly optimized vendor libraries such as cuBLAS and cuSOLVER for NVIDIA GPUs, rocBLAS and rocSOLVER for AMD GPUs and the Intel oneAPI Math Kernel Library (oneMKL) for Intel GPUs. Additionally, we rely on external open source libraries such as SLATE (Software for Linear Algebra Targeting Exascale), MAGMA (Matrix Algebra on GPU and Multi-core Architectures), and KBLAS (KAUST BLAS). SLATE is used as a GPU-capable ScaLAPACK replacement. From MAGMA we use variable sized batched dense linear algebra operations such as GEMM, TRSM and LU with partial pivoting. KBLAS provides efficient (batched) low rank matrix compression for NVIDIA GPUs using an adaptive randomized sampling scheme. The resulting sparse solver and preconditioner runs on NVIDIA, AMD and Intel GPUs. Interfaces are available from PETSc, Trilinos and MFEM, or the solver can be used directly in user code. We report results for a range of benchmark applications, using the Perlmutter system from NERSC, Frontier from ORNL, and Aurora from ALCF. For a high frequency wave equation on a regular mesh, using 32 Perlmutter compute nodes, the factorization phase of the exact GPU solver is about 6.5× faster compared to the CPU-only solver. The BLR-enabled GPU solver is about 13.8× faster than the CPU exact solver. For a collection of SuiteSparse matrices, the STRUMPACK exact factorization on a single GPU is on average 1.9× faster than NVIDIA’s cuDSS solver.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

UMap: An application-oriented user level memory mapping library

Exploiting the prominent role of complex memories in exascale node architecture, the UMap page fault handler offers new capabilities to access large memory-mapped data sets directly. UMap provides flexible configuration options to customize page handling to each application, including analysis of massive observational and simulation data sets. The high-performance design features I/O decoupling, dynamic load balancing, and application-level controls. Page faults triggered by application threads and processes accessing data mapped to a UMapp’ed region are handled via the Linux userfaultfd protocol, an asynchronous message-oriented kernel-user communication mechanism that avoids the context switch penalty of traditional signal fault handlers. UMap is fully open source. In this paper, we give an overview of the UMap library architecture, its extensible plugin architecture, and the use/performance of UMap in emerging heterogeneous memory hierarchies such as near-node Non-volatile Memory (NVM) and network attached memories. We highlight new capabilities in two pagefault management plugins, the NetworkStore and SparseStore. We demonstrate the integration between UMap and multiple ECP products including Caliper, Metall, ZFP, Mochi, and Ripples.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A two-level GPU-accelerated incomplete LU preconditioner for general sparse linear systems

This paper presents a parallel preconditioning approach based on incomplete LU (ILU) factorizations in the framework of Domain Decomposition (DD) for general sparse linear systems. We focus on distributed memory parallel architectures, specifically, those that are equipped with graphic processing units (GPUs). In addition to block-Jacobi, we present general purpose two-level ILU Schur complement-based approaches, where different strategies are presented to solve the coarse-level reduced system. These strategies are combined with modified ILU methods in the construction of the coarse-level operator, in order to effectively remove smooth errors by targeting an algebraically smooth vector. We leverage available GPU-based sparse matrix kernels to accelerate the setup and the solve phases of the proposed ILU preconditioner. We evaluate the efficiency of the proposed methods as a smoother for algebraic multigrid (AMG) and as a preconditioner for Krylov subspace methods on challenging anisotropic diffusion problems and a collection of general sparse matrices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Numerical investigation of a fueled pre-chamber spark-ignition natural gas engine

Pre-chamber spark-ignition (PCSI) is a leading advanced ignition concept for internal combustion engines with the potential to enable diesel-like efficiency in medium-duty/heavy-duty (MD/HD) natural gas (NG) engines. By leveraging distributed ignition sources from multiple turbulent jets, the PCSI technology can deliver extremely short combustion duration in ultra-lean mixtures and significantly improve the engine thermal efficiency. However, in the automotive industry there is a lack of adequate science base and predictive simulation tools required for commercial development of PCSI engines. Here, in this study, Reynolds-Average Navier-Stokes simulations are carried out to describe the combustion process in lean-burn NG engines, focusing on the combustion modeling approach. Two combustion models, multi-zone well-stirred reactor (MZ-WSR) and G-equation, are used to simulate the combustion process in an MD NG engine equipped with a fueled-PCSI system for four operating conditions close to the lean operating limit. A skeletal chemical mechanism and a laminar flame speed tabulation are used to compute the combustion accurately. Simulation results are compared with experimental data regarding measured cylinder pressure, heat release rate, and combustion duration. By dividing the PCSI combustion process into four distinct phases, the difference between the two models’ results for each phase is analyzed in detail. The MZ-WSR model overestimates the combustion duration for early flame kernel growth in the pre-chamber due to the lack of a specific formulation to take turbulence-chemistry interaction into account. Despite the prolonged combustion duration and low pressure built-up inside the pre-chamber, the model matches the combustion rate in the main-chamber. In contrast, the G-equation model delivers good agreements for the pre-chamber combustion and turbulent jet-driven combustion processes. However, the model starts to underestimate the combustion rate in the main-chamber, especially under ultra-lean mixture conditions. Finally, improvements are needed for both models to simulate the later combustion stage that occurred in the near-wall regions.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2024

An FPGA-based hardware accelerator supporting sensitive sequence homology filtering with profile hidden Markov models

Abstract Background Sequence alignment lies at the heart of genome sequence annotation. While the BLAST suite of alignment tools has long held an important role in alignment-based sequence database search, greater sensitivity is achieved through the use of profile hidden Markov models (pHMMs). Here, we describe an FPGA hardware accelerator, called HAVAC, that targets a key bottleneck step (SSV) in the analysis pipeline of the popular pHMM alignment tool, HMMER. Results The HAVAC kernel calculates the SSV matrix at 1739 GCUPS on a $$\sim$$ ∼ $3000 Xilinx Alveo U50 FPGA accelerator card, $$\sim$$ ∼ 227× faster than the optimized SSV implementation in nhmmer . Accounting for PCI-e data transfer data processing, HAVAC is 65× faster than nhmmer’s SSV with one thread and 35× faster than nhmmer with four threads, and uses $$\sim$$ ∼ 31% the energy of a traditional high end Intel CPU. Conclusions HAVAC demonstrates the potential offered by FPGA hardware accelerators to produce dramatic speed gains in sequence annotation and related bioinformatics applications. Because these computations are performed on a co-processor, the host CPU remains free to simultaneously compute other aspects of the analysis pipeline.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Disparities in the air quality monitoring stations and PM₂.₅ in Chicago’s air quality landscape

Fine particulate matter (PM₂.₅) poses significant public and environmental health risks in urban areas. Chicago’s dense industry and traffic create variable air quality, yet monitoring is unevenly distributed, resulting in undersampling of air quality data in some city areas. This study applied a hybrid approach using GIS-based kernel density mapping, interpolation modeling (IDW, Spline, Kriging) of USEPA monitoring data, multi-scale temporal trend analyses (hourly to annual), and ESDA. Accordingly, the density surface showed that monitors are concentrated in the affluent north, northwest, and southwest sides of Chicago (up to ~ 0.07 stations per sq mile), while the south and southeast regions, with predominantly minority communities, have virtually no coverage. Overall, citywide coverage is minimal (~ 4–5 monitors total; ~0.02 per sq mile; ≈1 per 600,000 residents). Temporal analyses showed that the city’s mean annual PM₂.₅ (~ 10.8 µg/m³) exceeds USEPA/WHO standards (9 µg/m³), with summer means (~ 17.1 µg/m³) significantly higher than other seasons. Diurnally, a clear pattern was observed, with PM₂.₅ concentrations peaking overnight (00:00–03:00) and during the morning rush hours, and dipping during midday to late afternoon. Spatial distribution of PM₂.₅ identified hotspots near O’Hare Airport, the downtown Loop area, and south-side neighborhoods, contrasting with lower concentrations on the north side, revealing Chicago’s socioeconomic divides and resulting environmental inequities. The findings underscore the need for expanded monitoring and targeted interventions in under-monitored, high-pollution communities to advance equitable community health.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Modeling spatial patterns of longleaf pine needle dispersal using long-term data

Abstract Background Predicting patterns of fire behavior and effects in frequent fire forests relies on an understanding of fine-scale spatial patterns of available fuels. Leaf litter is a significant canopy-derived fine fuel in fire-maintained forests. Litter dispersal is dependent on foliage production, stand structure, and wind direction, but the relative importance of these factors is unknown. Results Using a 10-year litterfall dataset collected within eighteen 4-ha longleaf pine ( Pinus palustris Mill.) plots varying in canopy spatial pattern, we compared four spatially explicit models of annual needle litter dispersal: a model based only on basal area, an overstory abundance index (OAI) model, both isotropic and anisotropic litter kernel models, and a null model that assumed no spatial relationship. The best model was the anisotropic model (R 2 = 0.656) that incorporated tree size, location, and prevailing wind direction, followed by the isotropic model (R 2 = 0.612), basal area model (R 2 = 0.488), OAI model (R 2 = 0.416), and the null model (R 2 = 0.08). Conclusions As with previous studies, the predictive capability of the litter models was robust when internally validated with a subset of the original dataset (R 2 = 0.196–0.549); however, the models were less robust when challenged with an independent dataset (R 2 = 0.122–0.319) from novel forest stands. Our model validation underscores the need for rigorous tests with independent, external datasets to confirm the validity of litter dispersal models. These models can be used in the application of prescribed fire to estimate fuel distribution and loading, as well as aid in the fine tuning of fire behavior models to better understand fire outcomes across a range of forest canopy structures.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Three-dimensional nanoscale reduced-angle ptycho-tomographic imaging with deep learning (RAPID)

X-ray ptychographic tomography is a nondestructive method for three dimensional (3D) imaging with nanometer-sized resolvable features. The size of the volume that can be imaged is almost arbitrary, limited only by the penetration depth and the available scanning time. Here we present a method that rapidly accelerates the imaging operation over a given volume through acquiring a limited set of data via large angular reduction and compensating for the resulting ill-posedness through deeply learned priors. The proposed 3D reconstruction method “RAPID” relies initially on a subset of the object measured with the nominal number of required illumination angles and treats the reconstructions from the conventional two-step approach as ground truth. It is then trained to reproduce equal fidelity from much fewer angles. After training, it performs with similar fidelity on the hitherto unexamined portions of the object, previously not shown during training, with a limited set of acquisitions. In our experimental demonstration, the nominal number of angles was 349 and the reduced number of angles was 21, resulting in a x140 aggregate speedup over a volume of 4.48 x 93.18 x 3.92 μm 3 and with (14nm) 3 feature size, i.e. ~ 10 8 voxels. RAPID’s key distinguishing feature over earlier attempts is the incorporation of atrous spatial pyramid pooling modules into the deep neural network framework in an anisotropic way. We found that adjusting the atrous rate improves reconstruction fidelity because it expands the convolutional kernels’ range to match the physics of multi-slice ptychography without significantly increasing the number of parameters.

47 OTHER INSTRUMENTATION↗
DOE OSTI2022

Early arrival waveform inversion using data uncertainties and matching filters with application to near-surface seismic refraction data

We develop an early arrival waveform inversion (EAWI) technique for high-resolution near-surface velocity estimation by iteratively updating the P-wave velocity model to minimize the difference between the observed and calculated seismic refraction data. Traditional EAWI uses a least-squares penalty function and an acoustic forward-modeling engine. Conventional least-squares error is sensitive to data with low signal-to-noise ratio (S/N) and iterations of EAWI stop at a local-minimum data misfit or at the preassigned maximum number of iterations. These stopping criteria can result in overfitting the data. In addition, fitting the elastic field data with an acoustic modeling engine can introduce artifacts in velocity estimation, especially in land data with significant elastic effects. To overcome these challenges, we develop a robust EAWI (REAWI) method by (1) incorporating the data uncertainties into the penalty function and (2) mitigating the elastic effects using a matching filter workflow. The data uncertainties are estimated from waveform reciprocal errors. When full-waveform reciprocity is not available, trace interpolation is applied. The proposed method prevents closely fitting data with low S/N, avoids overall overfitting by stopping the iterations when a normalized chi-square ([Formula: see text]) waveform misfit of one is achieved, and is less affected by elastic effects. Numerical examples and application to near-surface refraction data at a groundwater contamination site suggest that the final REAWI models are more accurate than the corresponding EAWI models, at the same level of misfit. This is the first known application of a matching filter workflow to real land data. The final REAWI models satisfy an appropriate misfit between the real data and predicted elastic P-wave data, making this approach in this respect equivalent to elastic waveform inversion. We also develop a method to analyze model constraint by examining the energy of the wavefield Fréchet derivative thereby avoiding the influence of the data residuals in traditional Fréchet kernels.

Geochemistry & Geophysics↗
DOE OSTI2023

Co-Harvest and Anaerobic Co-Storage of Corn Grain and Stover as Biomass Feedstocks

This research investigated the utility of co-harvesting and anaerobic co-storage of corn grain and stover to positively influence their physical and chemical characteristics as a biomass feedstock. Corn grain and stover were harvested in 2019 and 2020 with a self-propelled forage harvester. Stover yield, moisture content, and composition were altered by the harvest date, stubble height, and header configuration. Harvest date had the utility of varying the stover moisture content (p < 0.001) from 42.3% to 53.5% (w.b.) and 43.1% to 53.9% (w.b.) for the 2019 and 2020 harvest years, respectively. Stubble height was also utilized to vary stover moisture content. A negative linear relationship was established between stubble height and stover moisture content for the early (R2 = 0.76) and late harvest (R2 = 0.91) dates for both years. Stover yield also showed a negative linear relationship (R2 = 0.76) with stubble height over both years. Regardless of the stubble height, the row-crop header collected more stover (p < 0.001) than the ear-snapper header. In 2020, harvested stover ranged from 5.0 to 10.5 Mg ha-1, with ha-1 representing 41% to 85% of the total available stover. In both years, stover ash content was less than 64 g kg-1. Material stored in pilot-scale silos (19 L) was well conserved during anaerobic storage, with average DM losses of 4.8% and 3.4% in 2019 and 2020, respectively. Grain moisture content averaged 23.6% (w.b.) at harvest, and 31.0% (w.b.) after storage as moisture migrated from the moist stover to the drier grain. Harvesting whole-plant corn with a forage harvester had the unwanted effect of reducing the particle size of the grain fraction, which would complicate downstream utilization. However, reducing the harvester cutterhead speed increased the fraction of intact kernels from 47% to 85% by mass. The studied system was a viable alternative to conventional corn grain and stover systems for producing feedstocks for biochemical conversion.

09 BIOMASS FUELS↗
DOE OSTI2022

Thermal performance sensitivity to radial power flattening considerations in Nuclear Thermal Propulsion Reactors

Nuclear thermal propulsion is currently being developed for manned Mars and Lunar missions by the National Aeronautics and Space Administration due to its increased specific impulse and significant thrust to reduce the total trip time and cost of deep space missions. To maximize the performance of the engine the exit hydrogen propellant temperature should, ideally, be uniformly distributed among all the fuel elements. Uniform exit temperatures imply that the power must be flattened as much as possible. This paper leverages decoupled neutronic and thermal-hydraulic models to assess the thermal performance and reactivity tradeoffs associated with flattening the radial power distribution by adjusting the fuel kernel loading for each fuel element within the core. The results suggest that a uniform radial power distribution for a single control drum position will not necessarily yield an optimal core design due to the higher variance in power and thus maximum fuel and exit propellant temperature as the control drums rotate out during operation. The results presented here only point to the design characteristics be considered when core-level multi-physics simulations are applied in the context of a specific mission. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2020

A Cold Moderator For Sub-Thermal Neutron Flux Enhancement at the RPI-LINAC

The importance of high-quality thermal neutron scattering kernels coupled with a lack of accurate sub-thermal total cross section data prompted researchers at Rensselaer Polytechnic Institute (RPI) to design a cold moderation system to enhance sub-thermal neutron flux (defined as flux below 5 meV), for nuclear data measurements at the RPI Gaerttner LINAC facility. Using a series of neutronic and heat transfer simulations obtained with MCNP and COMSOL, an optimized final design was produced. The final design exceeded current target sub-thermal neutron producing capabilities by a factor of 8 in the sub-thermal region and is easily configurable with existing neutron producing targets.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2025

Micromechanical Properties of the SiC and Pyrolytic Carbon Layers in Tristructural-Isotropic Coated Particles

Tristructural isotropic (TRISO) coated particle fuel was initially developed for high temperature gas cooled reactors (HTGR) and has been proposed for several other advanced reactor concepts. The design of TRISO particles focuses on preventing the release of fission products in normal and off-normal reactor conditions. The particle design features an actinide bearing fuel kernel that is surrounded by three pyrolytic carbon (PyC) layers and a silicon carbide layer (SiC). The mechanical stability of the particle and fission product retention for both metallic and gaseous fission products depend on the SiC layer. Post irradiation examination (PIE) of TRISO fuel from the first two US DOE Advanced Gas Reactor Fuel Development and Qualification Program irradiation campaigns, AGR-1 and AGR-2, had identified a low rate of particles exhibiting cracking in the SiC layer that did not propagate across the SiC layer. While cracking in the SiC is rare for test conditions and particles associated with the AGR program, understanding the stress state and mechanical properties of the SiC and PyC layers related to particle architecture can aid predicting thermomechanical response of TRISO fuel under the prescribed operation envelope and beyond as well as aiding in the development of similar fuel concepts for other advanced reactors. The presented investigation shows the relationship of the mechanical properties and mechanical response (e.g., understanding crack propagation) of the SiC and PyC layers relative to position within the particle. Testing was conducted on the inner and outer PyC layers of TRISO particles to quantify differences in mechanical behavior.

Montoya, Katherine [ORNL] (ORCID:0000000326955086)↗
DOE OSTI2023

Accelerated deep self-supervised ptycho-laminography for three-dimensional nanoscale imaging of integrated circuits

Three-dimensional inspection of nanostructures such as integrated circuits is important for security and reliability assurance. Two scanning operations are required: ptychographic to recover the complex transmissivity of the specimen, and rotation of the specimen to acquire multiple projections covering the 3D spatial frequency domain. Two types of rotational scanning are possible: tomographic and laminographic. For flat, extended samples, for which the full 180° coverage is not possible, the latter is preferable because it provides better coverage of the 3D spatial frequency domain compared to limited-angle tomography. It is also because the amount of attenuation through the sample is approximately the same for all projections. However, both techniques are time consuming because of extensive acquisition and computation time. Here, we demonstrate the acceleration of ptycho-laminographic reconstruction of integrated circuits with 16 times fewer angular samples and 4.67 times faster computation by using a physics-regularized deep self-supervised learning architecture. We check the fidelity of our reconstruction against a densely sampled reconstruction that uses full scanning and no learning. As already reported elsewhere [ Opt. Express 28 , 12872 ( 2020 ) OPEXFF 1094-4087 10.1364/OE.379200 ], we observe improvement of reconstruction quality even over the densely sampled reconstruction, due to the ability of the self-supervised learning kernel to fill the missing cone.

47 OTHER INSTRUMENTATION↗