Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “parallel codes”

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 757 records · Page 42
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Electric field and turbulence in global Braginskii simulations across the ASDEX Upgrade edge and scrape-off layer

Turbulence simulations in diverted geometry across the edge and scrape-off layer (SOL) of ASDEX Upgrade are performed with the GRILLIX code (Stegmeir et al 2019 Phys. Plasmas 26 052517). The underlying global (full-f) drift-reduced Braginskii model allows to concurrently study the self-consistent dynamics of the turbulence and the background as well as the evolution of toroidal and zonal flows. Different contributions to the radial electric field are identified. The dominant contribution on closed flux surfaces comes from the ion pressure gradient, due to the diamagnetic drift in the curved magnetic field. Large deviations can be induced, in particular, by the polarization particle flux, leading to zonal flows. The latter are driven by small-scale eddies, but do not exhibit much impact on the overall transport which is driven by ballooning modes at larger scales. Ion viscosity is found to be important in damping poloidal rotation through adjusting of the parallel velocity profile, but not via direct vorticity damping. The zonal flow drive peaks at the separatrix, where a strong shear layer forms due to the sheath-induced counter-propagating SOL flow, allowing for the formation of a transport barrier. The temperature profile across the separatrix is determined by the competition between cross-field transport and outflow in the SOL, the latter being largely controlled by the parallel heat conductivity.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Emu v1.1

Emu is a particle-in-cell code for solving the neutrino quantum kinetic equations in 1, 2, or 3 spatial dimensions with arbitrary angular resolution in order to simulate neutrino flavor transformation in neutron star merger and core collapse supernova environments. Emu represents the neutrino distribution function as a set of particles, each of which represent a collection of neutrinos and antineutrinos with unique position and momentum. Each particle carries two density matrices to define the flavor state of the neutrinos and antineutrinos it represents. Emu includes the vacuum, matter, and neutrino self-interaction potentials. Emu calculates the self-interaction potential using PIC deposition and interpolation algorithms that efficiently compute the local number density and flux at particle locations. Emu is implemented in C++ and is based on the AMReX library for high-performance, block-structured adaptive mesh refinement. Emu is parallelized with MPI + OpenMP for CPUs and MPI + CUDA for GPUs.

Willcox, Donald↗
DOE OSTI2024

Using SPHINX Accelerator to Measure Heat Transfer in Metal Foils [Slides]

Experiments on the SPHINX accelerator at Sandia National Laboratories (SNL) looked at electron beam heating of metal foils used in Bremsstrahlung flash x-ray production. Titanium and tantalum foils were exposed to electron beam energies of up to 2MeV to determine surface heating. As part of the experiment, the x-ray dose profiles were characterized using a scintillating panel (EJ-230) placed both perpendicular and parallel to the photon beam. This was followed by direct electron beam imaging using a fast-gated ICCD camera and a fused silica Cherenkov witness plate. Once characterized, the foil was directly measured using a mid-wave (3-5 micron) infrared camera (MWIR) to determine the peak surface temperatures and spatially-resolved thermal profiles. Additionally, an IR pyrometer was used in a single wavelength mode (2.3 µm) to determine localized surface temperatures as well. Results are compared with calculations of surface temperatures obtained from e-beam energy depositions combined with the specific heats of the foils. This is the first attempt to directly measure the surface temperature profile on a short-pulse (10 nanoseconds) flash x-ray source using MWIR imaging. These data are directly relevant for Monte Carlo/Integrated Tiger Series electron/photon transport code results being used in the design of future Bremsstrahlung sources such as CREST.

43 PARTICLE ACCELERATORS↗
DOE OSTI2020

Model Exploration of an Information-Based Healthcare Intervention Using Parallelization and Active Learning

This paper describes the application of a large-scale active learning method to characterize the parameter space of a computational agent-based model developed to investigate the impact of CommunityRx, a clinical information-based health intervention that provides patients with personalized information about local community resources to meet basic and self-care needs. Additionally, the diffusion of information about community resources and their use is modeled via networked interactions and their subsequent effect on agents' use of community resources across an urban population. A random forest model is iteratively fitted to model evaluations to characterize the model parameter space with respect to observed empirical data. We demonstrate the feasibility of using high-performance computing and active learning model exploration techniques to characterize large parameter spaces; by partitioning the parameter space into potentially viable and non-viable regions, we rule out regions of space where simulation output is implausible to observed empirical data. We argue that such methods are necessary to enable model exploration in complex computational models that incorporate increasingly available micro-level behavior data. We provide public access to the model and high-performance computing experimentation code.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Formation of edge pressure pedestal and reversed magnetic shear due to toroidal rotation in a tokamak equilibrium

Toroidal rotation is well known to play significant roles in the edge transport and L-H transition dynamics of tokamaks. Our recent calculation finds that a sufficiently strong localized toroidal rotation can directly bring out the formation of edge pressure pedestal with reversed magnetic shear that is reminiscent of an H-mode plasma, purely through the effects of toroidal rotation on the tokamak MHD equilibrium itself. In particular, the enhanced edge toroidal rotation enables a substantial peaking of the parallel current profile near edge in higher β regimes, which leads to the flattening or reversal of the local q (safety factor) profile. Here the formation of pressure pedestal along with the reversed magnetic shear region is shown to be the natural outcome of the MHD tokamak equilibrium in a self-consistent response to the presence of a localized toroidal rotation typically observed in H-mode or QH-mode.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

A whole-core steady-state thermal-hydraulic model for annular fuel type fluoride-salt-cooled reactors

A whole-core, steady-state thermal-hydraulic model is developed for the fluoride-salt-cooled small modular advanced high-temperature reactor (SmAHTR) that employs an annular fuel configuration. This pre-conceptual reactor design by Oak Ridge National Laboratory (ORNL) has the annular fuel and moderator pins arranged in a hexagonal layout. The FLiBe coolant flows from the bottom to the top of the core, parallel to the hexagonal bundle. The fuel and moderator pins in the core are discretized into finite volumes and the 3-D heat conduction equation is solved to obtain the temperature profile. Inter-fuel assembly conduction is also addressed. For this fuel assembly configuration, the coolant flows through two distinct regions – the hexagonal pin bundle and the annulus between the fuel pin and the tie rod. The fluid flow through the hexagonal bundles is modeled using the subchannel approach, in which the coolant region is discretized into corner, edge and interior subchannels and the resulting conservation equations are systematically solved. The 1-D mass, momentum and energy equations are solved for the annulus channels between the fuel pin and the tie rod. Pertinent closure models from the literature are employed to close the system of equations. We also performed a preliminary code-to-code comparison between the present model and a CFD model.. The resulting thermal-hydraulic model can provide temperature, flow rate and pressure drop profiles for the different solid and fluid regions throughout the entire core. Whole-core thermal-hydraulic results for a representative power profile are presented and discussed.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Excitation of fishbone-like mode in tokamaks due to bounce resonances of trapped thermal ions

When the drift kinetic effect of thermal ions is taken into account, a high-frequency fishbone-like mode (FLM) is found to be driven unstable by trapped thermal ions (TTIs) in tokamak plasmas, according to self-consistent magneto-hydrodynamic (MHD)-kinetic hybrid modeling utilizing the MARS-K code [Liu et al., Phys. Plasmas 15, 112503 (2008)] as well as an analytic theory. It is found that, similar to energetic particles, TTIs can also stabilize the internal kink mode, whereas the FLM is excited when the effective beta of TTIs exceeds a threshold value. The real frequency of the FLM is comparable to the bounce frequency of TTIs. The mode structure of the FLM can be significantly different from the conventional step-like function for the associated plasma radial displacement. This drift kinetic induced modification of the mode structure near the q = 1 surface is captured by non-perturbative MHD-kinetic hybrid computations with MARS-K. Furthermore, the FLM can only be triggered by TTIs at sufficiently high thermal temperatures. Both the FLM and the internal kink can be stabilized by sufficiently fast plasma toroidal rotation and parallel sound wave damping. These two conditions of high thermal temperature and (fast) flow stabilization, though making it challenging to observe the TTI-driven FLM in present day experiments, are favorable for the mode excitation in future reactor scale devices.

36 MATERIALS SCIENCE↗
DOE OSTI2024

FENIX: Towards a Fully Integrated Multiphysics Framework for Plasma Facing Component Modeling

Computational tools have a crucial role to play in accelerating the deployment of fusion as a clean, reliable, abundant, and sustainable energy source. Multiphysics, high-fidelity simulation capabilities can help model, study, and predict intricate interactions between materials performance, plasma exposure, neutron irradiation, and engineering processes. As such, they can assist in the resolution of scientific and engineering challenges underpinning design, construction, and commission of fusion power plants. To address these needs, ongoing efforts are leveraging the Multiphysics Object-Oriented Simulation Environment (MOOSE) framework and delivering new computational tools for the fusion community. These tools inherit crucial attributes from MOOSE. They are open-source, modular, integrated with nuclear industry-standard software quality assurance processes, and enable multiphysics, multi-fidelity, fully integrated, zero- to three-dimensional, and massively parallel simulations. After a short overview of these capabilities, we will present the development of Fusion ENergy Integrated multiphys-X (FENIX), a MOOSE-based application designed to enable plasma facing component design and performance evaluation. Throughout their lifetime, plasma facing components are exposed to extreme thermal loads, repeated thermal shocks, and irradiation by plasma ions, neutral particles, and high-energy neutrons. Consequently, designing a plasma facing component with acceptable lifetime degradation is extremely challenging. FENIX aims to model the multiphysics environment in which plasma facing components evolve to accelerate their design studies. To that end, FENIX couples existing MOOSE capabilities such as heat transfer, thermomechanics, and thermal hydraulics, with tritium transport via the MOOSE-based Tritium Migration Analysis Program, Version 8 (TMAP8), with neutronics via the MOOSE-based high-fidelity neutron-photon transport and fluid dynamics code Cardinal, and finally with Particle-in-Cell plasma simulation capabilities being developed in this project. In this study, we present the current FENIX capabilities and preliminary results of its application to model the Tritium Plasma Experiment set up at Idaho National Laboratory.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Particle-based modelling of axisymmetric tandem mirror devices

In this work, we describe the use of a 1D-2V quasi-neutral hybrid electrostatic PIC with Monte-Carlo Coulomb collisions and non-uniform magnetic field to model the parallel transport and confinement in an axisymmetric tandem mirror device. End-plugs, based on simple-mirrors, are positioned at each end of the device and fueled with neutral beams (25 and 100 keV) to produce a sloshing ion population and increase the density of the end-plugs relative to the central cell. Results show the formation of a potential difference barrier between the central cell and the end-plugs. This potential confines a large fraction of the low energy thermal ions in the central cell which would otherwise be lost in a simple mirror, demonstrating the advantage of the beam-driven tandem mirror configuration relative to simple mirrors. In addition, we explore the effect of end-plug electron temperature on the confinement time of the device and compare it with theoretical estimates. Finally, we discuss the limitations of the code in its present form and describe the next logical steps to improve its predictive capability such as a fully nonlinear Fokker–Planck collision operator, multiply nested flux surface solutions and modeling the exhaust region up to the wall.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

gRNA-SeqRET: a universal tool for targeted and genome-scale gRNA design and sequence extraction for prokaryotes and eukaryotes

High-throughput genetic screening is frequently employed to rapidly associate gene with phenotype and establish sequence-function relationships. With the advent of CRISPR technology, and the ability to functionally interrogate previously genetically recalcitrant organisms, non-model organisms can be investigated using pooled guide RNA (gRNA) libraries and sequencing-based assays to quantitatively assess fitness of every targeted locus in parallel. To aid the construction of pooled gRNA assemblies, we have developed an in silico design workflow for gRNA selection using the gRNA Sequence Region Extraction Tool (gRNA-SeqRET). Built upon the previously developed CCTop, gRNA-SeqRET enables automated, scalable design of gRNA libraries that target user-specified regions or whole genomes of any prokaryote or eukaryote. Additionally, gRNA-SeqRET automates the bulk extraction of any regions of sequence relative to genes or other features, aiding in the design of homology arms for insertion or deletion constructs. We also assess in silico the application of a designed gRNA library to other closely related genomes and demonstrate that for very closely related organisms Average Nucleotide Identity (ANI) > 95% a large fraction of the library may be of relevance. The gRNA-SeqRET web application pipeline can be accessed at https://grna.jgi.doe.gov. The source code is comprised of freely available software tools and customized Python scripts, and is available at https://bitbucket.org/berkeleylab/grnadesigner/src/master/ under a modified BSD open-source license (https://bitbucket.org/berkeleylab/grnadesigner).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Multiphysics Demonstration of Temperature-Driven Assembly Bowing in SFRs using MOOSE-Based Codes

Core bowing is an important passive safety mechanism in liquid metal cooled fast reactors. When the core restraint system is properly designed, temperature and flux gradients influence assemblies in the core to bow into less reactive configurations during accident scenarios, resulting in negative reactivity feedback. Prediction of core bowing involves complex interplay of radiation transport, impacts of fluid flow and heat transfer on duct temperature, and mechanical responses to the induced temperature and flux gradients. Under the U.S. Department of Energy Office of Nuclear Energy’s Advanced Modeling and Simulation (NEAMS) Program [1], an integrated multiphysics approach is being developed to model the core bowing phenomena in liquid metal-cooled fast reactors with the Multiphysics Object Oriented Simulation Environment (MOOSE) [2]. In this methodology, the MOOSE-based reactor physics code Griffin [3] will solve the neutron transport equation and determine the power distribution. With the detailed power distribution from Griffin, the subchannel analysis codes MOOSE-Subchannel [4] and Pronghorn [5] are utilized to calculate the assembly temperature distribution. MOOSE’s Solid Mechanics [6] and Contact [7] Modules are leveraged to calculate the thermal expansion and duct bowing displacement with the duct wall temperature from thermal hydraulics calculation. In this work, an initial one-way coupling demonstration of the integrated multiphysics approach has been performed on a seven-assembly problem based on the sodium-cooled fast reactor ABR-1000 design [8]. The neutronics calculation with Griffin is not yet involved in the current simulation. MOOSE-Subchannel and Pronghorn evaluate fluid and solid temperature based on a fixed power distribution. In addition, one-way coupling is utilized in this coupled calculation, via Pronghorn passing the duct temperature data to the MOOSE Solid Mechanics calculation. An assessment of the Solid Mechanics module was performed in parallel to verify duct bowing behavior with duct-to-duct contact phenomenon [9]. The displacement from MOOSE Solid Mechanics is not yet transferred back and utilized in the Pronghorn and MOOSE-Subchannel calculation. This model will be available on the National Reactor Innovation Center (NRIC) Virtual Test Bed (VTB) repository [10]. Future stages of this work will involve solving problems of increasing complexity as well as adding more physics (e.g. reactor physics) to the integrated workflow to reach the end goal of modeling the core bowing phenomenon with an integrated multiphysics workflow.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Performance Results on CPU/GPU Exascale Architectures for OMEGA: The Ocean Model for E3SM Global Applications

The US Department of Energy (DOE) conducts climate simulations on some of the world’s largest supercomputers. These exascale machines use heterogeneous architectures with both CPUs and GPUs, and scientific codes must adapt to make full use of this computing power. Los Alamos National Lab is developing Omega: The Ocean Model for E3SM Global Applications, which is specifically designed for modern exascale computers. It uses external libraries that have been optimized for a variety of architectures to run on different supercomputers. Omega is an unstructured-mesh ocean model based on TRiSK numerical methods. It will be the new ocean component of the DOE’s Energy Exascale Earth System Model (E3SM). The algorithms in Omega follow those of the current ocean component, MPAS-Ocean, but it will be written in C++ rather than Fortran to take advantage of the Kokkos performance portability library. Omega spatial operators are written as Kokkos kernels to run efficiently on both CPUs and GPUs. Work on Omega began in 2023 with a new C++ framework for unstructured mesh partitioning, halo exchanges, parallel IO, and Kokkos interfaces. The current version, Omega-0, is being developed to solve the shallow water equations and at present includes all of the tendency terms but not time stepping. Here we share the results of Omega-0 verification and performance testing. Verification includes unit tests implemented with CTest as well as convergence tests in Polaris, an in-house python package with a large suite of test problems. Performance tests compare simulations conducted on CPUs versus GPUs and across different architectures: tests are run on Frontier, which has AMD “Optimized 3rd Gen EPYC” CPUs and AMD MI250X GPUs, as well as Perlmutter, which is composed of AMD EPYC 7763 CPUs and NVIDIA A100 GPUs.

58 GEOSCIENCES↗
DOE OSTI2022

Large-Eddy Simulation of a wind turbine using a Filtered Actuator Line Model

When dealing with multirotor devices such as quadcopters or wind farms, the cost of blade-resolved large-eddy simulation (LES) becomes prohibitive. Combining LES with a family of lower-fidelity models, called actuator line models (ALMs), has grown in popularity in the past decade. ALM replaces full blade resolution with an array of actuator points or lines parameterized by aerodynamic lift/drag polar plots along the blades. Body forces computed based on these actuator points are then projected onto the LES flow mesh, mimicking the effect of rotating blades on the flow. However, the optimal projection radius and the associated LES grid size is often too restrictive for multirotor simulations. Recently, a new tip-correction-based filtered ALM (F-ALM) was proposed by Martinez-Tossas and Meneveau (2019), which allows coarser-than-optimal grids by avoiding the associated overprediction of thrust. In this work, F-ALM is implemented into a high-order, in-house LES code to simulate National Renewable Energy Laboratory Phase VI wind turbine. It is then followed by a comparison between the baseline ALM and the newly implemented F-ALM in terms of instantaneous and time-averaged flow fields and blade loads, revealing the advantages of F-ALM in preventing the overprediction of power on coarse grids. Finally, this encourages accurate and affordable simulations of multirotor devices in the future.

17 WIND ENERGY↗
DOE OSTI2025

Designing a Framework for Solving Multiobjective Simulation Optimization Problems

Multiobjective simulation optimization (MOSO) problems are optimization problems with multiple conflicting objectives, where evaluation of at least one of the objectives depends on a black-box numerical code or real-world experiment, which we refer to as a simulation. Whereas an extensive body of research is dedicated to developing new algorithms and methods for solving these and related problems, it is challenging and time-consuming to integrate these techniques into real-world production-ready solvers. This is partly because of the diversity and complexity of modern state-of-the-art MOSO algorithms and methods and partly because of the complexity and specificity of many real-world problems and their corresponding computing environments. The complexity of this problem is only compounded when introducing potentially complex and/or domain-specific surrogate-modeling techniques, problem formulations, design spaces, and data acquisition functions. Here, this paper carefully surveys the current state of the art in MOSO algorithms, techniques, and solvers, as well as problem types and computational environments where MOSO is commonly applied. We then present several key challenges in the design of a parallel multiobjective simulation optimization framework (ParMOO) and how they have been addressed. Finally, we provide two case studies demonstrating how customized ParMOO solvers can be quickly built and deployed to solve real-world MOSO problems.

engineering design optimization↗
DOE OSTI2024

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count. Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 2048 3 on a single node to 32768 3 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 2048 3 where communication is entirely intra-node, and for 32768 3 , where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 32768 3 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

3D fast Fourier transform↗
DOE OSTI2020

High-Resolution Comonomer Sequencing of Blocky Brominated Syndiotactic Polystyrene Copolymers Using 13C NMR Spectroscopy and Computer Simulations

This work demonstrates the first high-resolution comonomer sequencing of Blocky brominated syndiotactic polystyrene (sPS-co-sPS-Br) copolymers based on pentad assign- ments of the quaternary carbon region of the nuclear magnetic resonance spectrum. Copolymers containing p-bromostyrene (Br-Sty) units were prepared in matched sets using postpolymerization bromination methods carried out in the heterogeneous gel state (Blocky) and homogeneous solution state (Random). Quantitative information from the quaternary carbon spectra, heteronuclear multiple bond correlation spectroscopy, electronic structure calculations, and simulated statistically random copolymers was correlated to confirm the carbon resonance assignments for all 20 possible pentad comonomer sequences. Using the experimental pentad sequence prevalences, a computer code was developed to simulate chains with microstructures typical of each sample as a means to visually represent the copolymer blockiness with quantitative precision. Based on the microstructure and distribution of run lengths in these chains, the simulations revealed that the Blocky copolymers contain a high degree of blockiness. By comparing the run lengths in the simulated chains to the average number of styrene units in a crystalline segment of sPS (found by small-angle X-ray scattering), copolymer crystallizability was predicted. For the simulated Blocky B-21% (21 mol % Br-Sty) chain, the probability of randomly selecting a styrene unit in a crystallizable block was 25.8%, while that in the simulated Random R-18% was zero, in excellent agreement with the experimental crystallization behavior measured by differential scanning calorimetry. Additionally, these predictions confirmed that the simulated chains accurately represent the ensemble of chains in their respective copolymer samples. Furthermore, each simulated Blocky chain contained one or more long sPS blocks that paralleled the measured 38-40 styrene units spanning a crystalline segment within the sPS/CCl4 gel. This finding affirmed that the long sPS segments originated from the precise lamellar structure within the heterogeneous gel morphology (i.e., block length is correlated with lamellar thickness). Overall, the ability to tailor the copolymer microstructure through control of the semicrystalline gel morphology opens the door to synthesizing ordered copolymers by postpolymerization functionalization processes with unprecedented levels of compositional control.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Scale-up Unlearnable Examples Learning with High-performance Computing

Recent advancements in AI models, like ChatGPT, are structured to retain user interactions, which could inadvertently include sensitive healthcare data. In the healthcare field, particularly when radiologists use AI-driven diagnostic tools hosted on online platforms, there is a risk that medical imaging data may be repurposed for future AI training without explicit consent, spotlighting critical privacy and intellectual property concerns around healthcare data usage. Addressing these privacy challenges, a novel approach known as Unlearnable Examples (UEs) has been introduced, aiming to make data unlearnable to deep learning models. A prominent method within this area, called Unlearnable Clustering (UC), has shown improved UE performance with larger batch sizes but was previously limited by computational resources (e.g., a single workstation). To push the boundaries of UE performance with theoretically unlimited resources, we scaled up UC learning across various datasets using Distributed Data Parallel (DDP) training on the Summit supercomputer. Our goal was to examine UE efficacy at high-performance computing (HPC) levels to prevent unauthorized learning and enhance data security, particularly exploring the impact of batch size on UE’s unlearnability. Utilizing the robust computational capabilities of the Summit, extensive experiments were conducted on diverse datasets such as Pets, MedMNist, Flowers, and Flowers102. Our findings reveal that both overly large and overly small batch sizes can lead to performance instability and affect accuracy. However, the relationship between batch size and unlearnability varied across datasets, highlighting the necessity for tailored batch size strategies to achieve optimal data protection. The use of Summit’s high-performance GPUs, along with the efficiency of the DDP framework, facilitated rapid updates of model parameters and consistent training across nodes. Our results underscore the critical role of selecting appropriate batch sizes based on the specific characteristics of each dataset to prevent learning and ensure data security in deep learning applications. The source code is publicly available at https: // github. com/ hrlblab/ UE_ HPC .

Zhu, Yanfan [Vanderbilt University, Nashville, TN,↗
DOE OSTI2022

The Astronomy Commons Platform: A Deployable Cloud-based Analysis Platform for Astronomy

Abstract We present a scalable, cloud-based science platform solution designed to enable next-to-the-data analyses of terabyte-scale astronomical tabular data sets. The presented platform is built on Amazon Web Services (over Kubernetes and S3 abstraction layers), utilizes Apache Spark and the Astronomy eXtensions for Spark for parallel data analysis and manipulation, and provides the familiar JupyterHub web-accessible front end for user access. We outline the architecture of the analysis platform, provide implementation details and rationale for (and against) technology choices, verify scalability through strong and weak scaling tests, and demonstrate usability through an example science analysis of data from the Zwicky Transient Facility’s 1Bn+ light-curve catalog. Furthermore, we show how this system enables an end user to iteratively build analyses (in Python) that transparently scale processing with no need for end-user interaction. The system is designed to be deployable by astronomers with moderate cloud engineering knowledge, or (ideally) IT groups. Over the past 3 yr, it has been utilized to build science platforms for the DiRAC Institute, the ZTF partnership, the LSST Solar System Science Collaboration, and the LSST Interdisciplinary Network for Collaboration and Computing, as well as for numerous short-term events (with over 100 simultaneous users). In a live demo instance, the deployment scripts, source code, and cost calculators are accessible. 4 4 http://hub.astronomycommons.org/

79 ASTRONOMY AND ASTROPHYSICS↗