Search NASA⌕ Search

SEARCH · Search NASA

Results for “hardware accelerator”

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 973 records · Page 54
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

A hybrid neural architecture: Online attosecond x-ray characterization

The emergence of high-repetition-rate x-ray free-electron lasers (XFELs), such as SLAC’s LCLS-II, serves as our canonical example for autonomous controls that necessitate high-throughput diagnostics paired with streaming computational pipelines capable of single-shot analysis with extremely low latency. We present the deterministic characterization with an integrated parallelizable hybrid resolver architecture, a hybrid machine learning framework designed for fast, accurate analysis of XFEL diagnostics using angular streaking-based sinogram images. This architecture integrates convolutional neural networks and bidirectional long short-term memory models to denoise input, identify x-ray sub-spike features, and extract sub-spike relative delays with sub-30 attosecond temporal resolution. Deployed on low-latency hardware, it achieves over 10 kHz throughput with 168.3 μs inference latency, indicating scalability to 14 kHz with field-programmable gate array integration. By transforming regression tasks into classification problems and leveraging optimized error encoding, we achieve high precision with low-latency performance that is critical for real-time streaming event selection and experimental control feedback signals. This represents a key development in real-time control pipelines for next-generation autonomous science, generally, and high repetition-rate x-ray experiments in particular.

Accelerator Physics (physics.acc-ph)↗
NASA NTRS2010

Rocket Cratering in Simulated Lunar and Martian Environments

With NASA's planned return to the moon and possibly with lunar outposts being formed, repeated landings at the same site will be necessary. Understanding rocket plume interaction with lunar and Martian surfaces is of paramount importance in order to safely land and protect hardware surrounding the landing site. This work will report on results of three small experiments intended to explore plume impingement onto lunar and Martian surfaces: Handheld Observation of Scour Holes (HOOSH), Handheld Angle of Repose Measurements of Lunar Simulants (HARMLuS), and Mars Architecture Team study (MATS). The first two experiments were performed during two sorties of reduced gravity flights. HOOSH was designed to investigate crater formation as a function of gravitational level (lunar and Martian gravity). HARMLuS was designed to measure the Angle of Failure (related to the angle of repose) at lunar and Martian gravity. Both experiments have complex findings indicative of the hysteretic behavior of granular materials, especially resulting from reduced gravity. The MATS experiment was designed to investigate the effects of regolith compaction on the granular mechanics of crater formation . In general, the granular mechanics is a much stronger function of compaction than gravitation acceleration. Crater formation is greatly enhanced at reduced gravity (resulting in much larger craters). The angle of failure of the lunar simulants increases with decreasing gravitational acceleration, and occasionally becomes infinite for some compactions at lunar gravity. The angle of failure also increases with increasing compaction. While compaction does play a role in the time development of crater formation, the asymptotic behavior is largely unaffected.

Immer, Christopher↗
DOE OSTI2018

Development of Optimized Controls and Hardware for Synchrotron Hard X-ray Microscopes with Advanced Spectroscopic Capabilities

The technical objectives of the program were to develop an integrated, fully automated command and data interface between the TXM-Wizard developed at SLAC and the Xradia control software that is used on the Xradia Nano-TXM x-ray microscope with the goal of allowing non-experts including industry users to routinely access the advanced imaging modes that employ collection of images with spectroscopic information. The major accomplishment associated with this goal was the development of a plan to convert the TXM-Wizard software from MatLab to Python and the hiring of an expert Python programmer under contract to perform the conversion.

43 PARTICLE ACCELERATORS↗
NASA NTRS2016

Post-Test Inspection of NASA's Evolutionary Xenon Thruster Long Duration Test Hardware: Ion Optics

A Long Duration Test (LDT) was initiated in June 2005 as a part of NASA's Evolutionary Xenon Thruster (NEXT) service life validation approach. Testing was voluntarily terminated in February 2014, with the thruster accumulating 51,184 hours of operation, processing 918 kg of xenon propellant, and delivering 35.5 MN-s of total impulse. The post-test inspection objectives for the ion optics were derived from the original NEXT LDT test objectives, such as service life model validation, and expanded to encompass other goals that included verification of in situ measurements, test issue root causes, and past design changes. The ion optics cold grid gap had decreased only by an average of 7% of pretest center grid gap, so efforts to stabilize NEXT grid gap were largely successful. The upstream screen grid surface exhibited a chamfered erosion pattern. Screen grid thicknesses were ≥ 86% of the estimated pretest thickness, indicating that the screen grid has substantial service life remaining. Deposition was found on the screen aperture walls and downstream surfaces that was primarily composed of grid material and back-sputtered carbon, and this deposition likely caused the minor decreases in screen grid ion transparency during the test. Groove depths had eroded through up to 35% of the accelerator grid thickness. Minimum accelerator aperture diameters increased only by about 5-7% of the pretest values and downstream surface diameters increased by about 24-33% of the pretest diameters. These results suggest that increasing the accelerator aperture diameters, improving manufacturing tolerances, and masking down the perforated diameter to 36 cm were successful in reducing the degree of accelerator aperture erosion at larger radii.

Ion Thruster↗
NASA NTRS2021

Post-test Inspection of NASA’s Evolutionary Xenon Thruster Long Duration Test Hardware: Ion Optics

A Long Duration Test (LDT) was initiated in June 2005 as a part of NASA’s Evolutionary Xenon Thruster (NEXT) service life validation approach. Testing was voluntarily terminated in February 2014, with the thruster accumulating 51,184 hours of operation, processing 918 kg of xenon propellant, and delivering 35.5 MN-s of total impulse. The post-test inspection objectives for the ion optics were derived from the original NEXT LDT test objectives, such as service life model validation, and expanded to encompass other goals that included verification of in situ measurements, test issue root causes, and past design changes. The ion optics cold grid gap had decreased only by an average of 7% of pretest center grid gap, so efforts to stabilize NEXT grid gap were largely successful. The upstream screen grid surface exhibited a chamfered erosion pattern. Screen grid thicknesses were ≥ 86% of the estimated pretest thickness, indicating that the screen grid has substantial service life remaining. Deposition was found on the screen aperture walls and downstream surfaces that was primarily composed of grid material and back-sputtered carbon, and this deposition likely caused the minor decreases in screen grid ion transparency during the test. Groove depths had eroded through up to 35% of the accelerator grid thickness. Minimum accelerator aperture diameters increased only by about 5-7% of the pretest values and downstream surface diameters increased by about 24-33% of the pretest diameters. These results suggest that increasing the accelerator aperture diameters, improving manufacturing tolerances, and masking down the perforated diameter to 36 cm were successful in reducing the degree of accelerator aperture erosion at larger radii.

George C. Soulas↗
DOE OSTI2025

Simplifying activations with linear approximations in neural networks

A key step in Neural Networks is activation. Among the different types of activation functions, sigmoid, tanh, and others involve the usage of exponents for calculation. From a hardware perspective, exponential implementation implies the usage of Taylor series or repeated methods involving many addition, multiplication, and division steps, and as a result are power-hungry and consume many clock cycles. We implement a piecewise linear approximation of the sigmoid function as a replacement for standard sigmoid activation libraries. This approach provides a practical alternative by leveraging piecewise segmentation, which simplifies hardware implementation and improves computational efficiency. In this paper, we detail piecewise functions that can be implemented using linear approximations and their implications for overall model accuracy and performance gain. Our results show that for the DenseNet, ResNet, and GoogLeNet architectures, the piecewise linear approximation of the sigmoid function provides faster execution times compared to the standard TensorFlow sigmoid implementation while maintaining comparable accuracy. Specifically, for MNIST with DenseNet, accuracy reaches 99.91% (Piecewise) vs. 99.97% (Base) with up to 1.31x speedup in execution time. For CIFAR-10 with DenseNet, accuracy improves to 98.97% (Piecewise) vs. 99.40% (Base) while achieving 1.24x faster execution. Similarly, for CIFAR-100 with DenseNet, the accuracy is 97.93% (Piecewise) vs. 98.39% (Base), with a 1.18x execution time reduction. These results confirm the proposed method’s capability to efficiently process large-scale datasets and computationally demanding tasks, offering a practical means to accelerate deep learning models, including LSTMs, without compromising accuracy.

Activation function↗
DOE OSTI2024

Analyzing inference workloads for spatiotemporal modeling

Ensuring power grid resiliency, forecasting climate conditions, and optimization of transportation infrastructure are some of the many application areas where data is collected in both space and time. Spatiotemporal modeling is about modeling those patterns for forecasting future trends and carrying out critical decision-making by leveraging machine learning/deep learning. Once trained offline, field deployment of trained models for near real-time inference could be challenging because performance can vary significantly depending on the environment, available compute resources and tolerance to ambiguity in results. Users deploying spatiotemporal models for solving complex problems can benefit from analytical studies considering a plethora of system adaptations to understand the associated performance-quality trade-offs. To facilitate the co-design of next-generation hardware architectures for field deployment of trained models, it is critical to characterize the workloads of these deep learning (DL) applications during inference and assess their computational patterns at different levels of the execution stack. In this paper, we develop several variants of deep learning applications that use spatiotemporal data from dynamical systems. We study the associated computational patterns for inference workloads at different levels, considering relevant models (Long short-term Memory, Convolutional Neural Network and Spatio-Temporal Graph Convolution Network), DL frameworks (Tensorflow and PyTorch), precision (FP16, FP32, AMP, INT16 and INT8), inference runtime (ONNX and AI Template), post-training quantization (TensorRT) and platforms (Nvidia DGX A100 and Sambanova SN10 RDU). Overall, our findings indicate that although there is potential in mixed-precision models and post-training quantization for spatiotemporal modeling, extracting efficiency from contemporary GPU systems might be challenging. Instead, co-designing custom accelerators by leveraging optimized High Level Synthesis frameworks (such as SODA High-Level Synthesizer for customized FPGA/ASIC targets) can make workload-specific adjustments to enhance the efficiency.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2013

Orion MPCV Touchdown Detection Threshold Development and Testing

A robust method of detecting Orion Multi ]Purpose Crew Vehicle (MPCV) splashdown is necessary to ensure crew and hardware safety during descent and after touchdown. The proposed method uses a triple redundant system to inhibit Reaction Control System (RCS) thruster firings, detach parachute risers from the vehicle, and transition to the post ]landing segment of the Flight Software (FSW). The vehicle crew is the prime input for touchdown detection, followed by an autonomous FSW algorithm, and finally a strictly time based backup timer. RCS thrusters must be inhibited before submersion in water to protect against possible damage due to firing these jets under water. In addition, neglecting to declare touchdown will not allow the vehicle to transition to post ]landing activities such as activating the Crew Module Up ]righting System (CMUS), resulting in possible loss of communication and difficult recovery. A previous AIAA paper gAssessment of an Automated Touchdown Detection Algorithm for the Orion Crew Module h concluded that a strictly Inertial Measurement Unit (IMU) based detection method using an acceleration spike algorithm had the highest safety margins and shortest detection times of other methods considered. That study utilized finite element simulations of vehicle splashdown, generated by LS ]DYNA, which were expanded to a larger set of results using a Kriging surface fit. The study also used the Decelerator Systems Simulation (DSS) to generate flight dynamics during vehicle descent under parachutes. Proto ]type IMU and FSW MATLAB models provided the basis for initial algorithm development and testing. This paper documents an in ]depth trade study, using the same dynamics data and MATLAB simulations as the earlier work, to further develop the acceleration detection method. By studying the combined effects of data rate, filtering on the rotational acceleration correction, data persistence limits and values of acceleration thresholds, an optimal configuration was determined. The lever arm calculation, which removes the centripetal acceleration caused by vehicle rotation, requires that the vehicle angular acceleration be derived from vehicle body rates, necessitating the addition of a 2nd order filter to smooth the data. It was determined that using 200 Hz data directly from the vehicle IMU outperforms the 40 Hz FSW data rate. Data persistence counter values and acceleration thresholds were balanced in order to meet desired safety and performance. The algorithm proved to exhibit ample safety margin against early detection while under parachutes, and adequate performance upon vehicle splashdown. Fall times from algorithm initiation were also studied, and a backup timer length was chosen to provide a large safety margin, yet still trigger detection before CMUS inflation. This timer serves as a backup to the primary acceleration detection method. Additionally, these parameters were tested for safety on actual flight test data, demonstrating expected safety margins.

Daum, Jared↗
DOE OSTI2020

Field Programmable Gate Arrays for Enhancing the Speed and Energy Efficiency of Quantum Dynamics Simulations

We present the first application of field programmable gate arrays (FPGAs) as new, customizable hardware architectures for carrying out fast and energy-efficient quantum dynamics simulations of large chemical/material systems. Instead of tailoring the software to fixed hardware, which is the typical case for writing quantum chemistry code for central processing units (CPUs) and graphics processing units (GPUs), FPGAs allow us to directly customize the underlying hardware (even at the level of specific electrical signals in the circuit) to give a truly optimized computational performance for quantum dynamics calculations. By offloading the most intensive and repetitive calculations onto an FPGA, we show that the computational performance of our real-time electron dynamics calculations can even exceed that of optimized commercial mathematical libraries running on high-performance GPUs. In addition to this impressive computational speedup, we show that FPGAs are immensely energy-efficient and consume 4 times less energy than modern GPU or CPU architectures. These energy savings are a practical and important metric for supercomputing centers (many of which exceed over $1 million in power costs alone), as exascale computing capabilities become more widespread and commonplace. Taken together, the implementation techniques and performance metrics of our study demonstrate that FPGAs could play a promising role in upcoming quantum chemistry and materials science applications, particularly for the acceleration and energy-efficient execution of quantum dynamics calculations.

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

Preliminary Plan to Inform Testing of a Heat Exchanger Test Article

This report presents a preliminary plan to guide the qualification testing of advanced heat exchanger (HX) components for nuclear-to-industrial heat transfer applications. The objective is to establish a defensible, physics-based methodology that integrates computational modeling, targeted experimentation, and in-service inspection considerations to demonstrate component performance and reliability under representative reactor conditions. The analysis identifies Sodium-cooled Fast Reactor (SFR) and High-Temperature Gas-cooled Reactor (HTGR) systems as reference configurations in terms of temperature, pressure, and chemical environment. Within these operating envelopes, dominant degradation mechanisms— including creep–fatigue interaction, flow-induced vibration, corrosion, and diffusion-bond deterioration—were evaluated to define test requirements. A comprehensive computationalexperimental framework is proposed to support life prediction and qualification activities. The framework couples high-fidelity structural-mechanics, thermal-hydraulic, and fluid-structure interaction models with accelerated degradation testing to produce a traceable linkage between microstructural evolution, mechanical performance, and remaining useful life (RUL). The approach adheres to established Verification, Validation, and Uncertainty Quantification (VVUQ) standards (ASME V&V 10/20; NUREG-2152) and incorporates a digital-twin architecture for continuous model refinement through data assimilation. The plan further outlines testing methodologies, including pre-test analyses, test-loop design parameters, and sensor placement strategies that maximize information yield while maintaining mechanistic fidelity. Complementary sections describe in-service inspection (ISI), on-line monitoring (OLM), and structural-health-monitoring (SHM) techniques applicable to compact HX geometries typical of advanced reactors. Collectively, these activities establish the technical foundation for demonstrating 40-60-year equivalent service life of advanced heat exchangers in support of the U.S. Department of Energy’s Advanced Reactor and Integrated Energy Systems programs. The forthcoming phase will execute the defined pre-test analyses, initiate hardware fabrication, and implement the integrated testing campaign to validate the proposed qualification methodology.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Artificial Intelligence for Autonomous Molecular Design: A Perspective

Domain-aware artificial intelligence has been increasingly adopted in recent years to expedite molecular design in various applications, including drug design and discovery. Recent advances in areas such as physics-informed machine learning and reasoning, software engineering, high-end hardware development, and computing infrastructures are providing opportunities to build scalable and explainable AI molecular discovery systems. This could improve a design hypothesis through feedback analysis, data integration that can provide a basis for the introduction of end-to-end automation for compound discovery and optimization, and enable more intelligent searches of chemical space. Several state-of-the-art ML architectures are predominantly and independently used for predicting the properties of small molecules, their high throughput synthesis, and screening, iteratively identifying and optimizing lead therapeutic candidates. However, such deep learning and ML approaches also raise considerable conceptual, technical, scalability, and end-to-end error quantification challenges, as well as skepticism about the current AI hype to build automated tools. To this end, synergistically and intelligently using these individual components along with robust quantum physics-based molecular representation and data generation tools in a closed-loop holds enormous promise for accelerated therapeutic design to critically analyze the opportunities and challenges for their more widespread application. This article aims to identify the most recent technology and breakthrough achieved by each of the components and discusses how such autonomous AI and ML workflows can be integrated to radically accelerate the protein target or disease model-based probe design that can be iteratively validated experimentally. Taken together, this could significantly reduce the timeline for end-to-end therapeutic discovery and optimization upon the arrival of any novel zoonotic transmission event. Our article serves as a guide for medicinal, computational chemistry and biology, analytical chemistry, and the ML community to practice autonomous molecular design in precision medicine and drug discovery.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Prime factorization using quantum variational imaginary time evolution

The road to computing on quantum devices has been accelerated by the promises that come from using Shor’s algorithm to reduce the complexity of prime factorization. However, this promise hast not yet been realized due to noisy qubits and lack of robust error correction schemes. Here we explore a promising, alternative method for prime factorization that uses well-established techniques from variational imaginary time evolution. We create a Hamiltonian whose ground state encodes the solution to the problem and use variational techniques to evolve a state iteratively towards these prime factors. We show that the number of circuits evaluated in each iteration scales as \(O(n^{5}d)\) , where n is the bit-length of the number to be factorized and d is the depth of the circuit. We use a single layer of entangling gates to factorize 36 numbers represented using 7, 8, and 9-qubit Hamiltonians. We also verify the method’s performance by implementing it on the IBMQ Lima hardware to factorize 55, 65, 77 and 91 which are greater than the largest number (21) to have been factorized on IBMQ hardware.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2000

Otoliths developed in microgravity

Little is known about mechanisms that regulate the development of the otoliths in the gravity-sensing organs. Several reported experiments suggest that the growth of the otoliths is adjusted to produce a test mass of the appropriate weight. If this is the case, larger than normal otoliths would be expected in animals reared in reduced gravity and a reduced mass, relative to 1-g controls, would be expected in animals reared at elevated g. In gastropod mollusks, the gravity-sensing organ is the statocyst, a spherical organ whose wall is made largely of sensory receptor cells with motile cilia facing the lumen. Dense statoconia in the cyst lumen interact with cilia of receptor cells at the bottom of the cyst and action potentials in their axons carry information on direction and magnitude of gravity and linear acceleration. In the marine mollusk, Aplysia californica, larvae reared at 2 to 5-g, the volume of statoconia was reduced in a graded manner, compared to 1-g control animals. In the statocyst of the fresh-water pond snail, Biomphalaria glabrata, reared in space in the Closed Equilibrated Biological Aquatic System (CEBAS), the number and total volume of statoconia was increased approximately 50%, relative to ground-reared controls. Lychakov found the utricular otolith to be 30% larger in space-reared Xenopus, whereas we found the saccular otolith to be significantly larger in newt larvae reared in space. In cichlid fish reared on a centrifuge, the saccular otolith was smaller than in 1-g controls. Here, we demonstrate that the otoliths of late-stage embryos of the swordtail fish, Xiphophorus helleri, reared in space on STS-89 and STS-90 (Neurolab) were significantly larger than those of ground-controls reared in functionally identical hardware.

short duration↗
NASA NTRS2000

Mechanically Induced g-Jitter from Space Station Rotary Joints

The mission of the International Space Station is to provide a working laboratory in orbit for research in engineering, life sciences, and microgravity. Among the microgravity disciplines that are preparing to utilize this international resource are materials processing, combustion, fluid dynamics, biotechnology, and fundamental physics. The Station promises to enable significant advances in each of these areas by making available a research facility in which gravitational and other accelerations, and their corresponding buoyancy and diffusion effects on various physical processes, are orders of magnitude lower than they are on Earth. In order to fulfill this promise, it is not enough for the Space Station to simply replicate a typical terrestrial scientific laboratory in orbit. Although an orbiting laboratory is free of most of the effects of gravitational acceleration by virtue of its free fall condition, it also produces structural vibration or jitter that can interfere with the processes under study. To ensure the quality of the acceleration environment and enable a successful mission, the Space Station Program has limited potential disturbances in two ways: first, by isolating the most sensitive payloads from the vehicle structure, and second, by quieting major disturbances at their sources. The first area, payload isolation, is implemented inside the pressurized modules at the rack level. Sub-rack level isolators have also been developed. This paper addresses the second area, disturbance source limits, for one of the major sources of mechanical noise on the Space Station: the Solar Alpha Rotary Joints. Due to the potential for large disturbances to the microgravity environment, an initial analytical prediction of rotary joint vibration output was made. Key components were identified and tested to validate the analytical predictions. Based on the component test results, the final vibration output of the joints was verified by a test on each fully assembled flight unit. This paper describes the Space Station microgravity requirements, the rotary joint hardware, and the disturbance producing aspects of joint operation. The test setup, instrumentation, test conditions, and results for the component level and system level measurements are described. An overall forcing function that describes the maximum torque imparted to the Station is created based on the test results, and these disturbances are shown to meet the applicable Space Station microgravity requirements.

Boucher, Robert L.↗
DOE OSTI2018

CBETA Technical Report (CBETA Note #31)

This document includes twelve sections related to the previous Fractional Arc Test (FAT), CBETA final installation plan has which follows the milestones signed in the contract with the New York State Energy Development Authority (NYSERDA). In the second part we provide a plan for final installation, hardware and software preparation and the beam commissioning.

43 PARTICLE ACCELERATORS↗
DOE OSTI2022

Quantum machine learning for chemistry and physics

Machine learning (ML) has emerged as a formidable force for identifying hidden but pertinent patterns within a given data set with the objective of subsequent generation of automated predictive behavior. In recent years, it is safe to conclude that ML and its close cousin, deep learning (DL), have ushered in unprecedented developments in all areas of physical sciences, especially chemistry. Not only classical variants of ML, even those trainable on near-term quantum hardwares have been developed with promising outcomes. Such algorithms have revolutionized materials design and performance of photovoltaics, electronic structure calculations of ground and excited states of correlated matter, computation of force-fields and potential energy surfaces informing chemical reaction dynamics, reactivity inspired rational strategies of drug designing and even classification of phases of matter with accurate identification of emergent criticality. In this review we shall explicate a subset of such topics and delineate the contributions made by both classical and quantum computing enhanced machine learning algorithms over the past few years. We shall not only present a brief overview of the well-known techniques but also highlight their learning strategies using statistical physical insight. The objective of the review is not only to foster exposition of the aforesaid techniques but also to empower and promote cross-pollination among future research in all areas of chemistry which can benefit from ML and in turn can potentially accelerate the growth of such algorithms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2017

Co-Optima Project E2.2.2: Accelerate Development of ACI/LTC Fuel Effects on RCCI Combustion

Many advanced combustion approaches have demonstrated potential for achieving diesel-like thermal efficiency but with much lower pollutant emissions of particulate matter (PM) and nitrogen oxides (NOx). RCCI is one advanced combustion concept, which makes use of in-cylinder blending of two fuels with differing reactivity for improved control of the combustion phasing and rate (Reitz et al., 2015). Previous research and development at ORNL has demonstrated successful implementation of RCCI on a light-duty multi-cylinder engine over a wide range of operating conditions (Curran et al., 2015). Several challenges were encountered when extending the research to practical applications, including limits to the operating range, both for high and low loads. Co-optimizing the engine and fuel aspects of the RCCI approach might allow these operating limits to be overcome. The in-cylinder mechanisms by which fuel properties interact with engine operating condition variables is not well understood, however, in part because RCCI is a new combustion concept that is still being developed, and limited data have been acquired to date, especially using in-cylinder optical/imaging diagnostics. The objective of this work is to use in-cylinder diagnostics in a heavy-duty single-cylinder optical engine at SNL to understand the interplay between fuel properties and engine hardware and operating conditions for RCCI in general, and in particular for the light-duty multi-cylinder all-metal RCCI engine experiments at ORNL.

02 PETROLEUM↗
DOE OSTI2025

Accelerating Neutrino Event Generation in MARLEY Using CUDA-Based RNG and GPU Parallelization

MARLEY is a simulation tool that helps scientists study how low-energy neutrinos interact with matter. To work properly, MARLEY uses random numbers thousands of times in each simulation. These random numbers are important for modeling things like how neutrinos collide with atoms and what particles they produce. Right now, MARLEY runs on a regular computer processor (CPU) and uses a built-in random number generator called the Mersenne Twister. This setup works, but it can be slow, especially when trying to simulate many events. This research focuses on making MARLEY run faster by moving the random number generation and some of the repetitive calculations from the CPU to a graphics processing unit (GPU), which can handle many tasks at the same time. We use CUDA (a tool for programming NVIDIA GPUs) and cuRAND (a GPU-based random number library) to test faster alternatives to the current random number system. We compare different GPU-based generators, like curand_mtgp32, xorwow, and philox, to see which ones are the quickest and still give reliable results. Early tests show that using the GPU can make MARLEY simulations much faster. This project not only helps improve current simulation performance but also moves closer to a full simulation chain where all stages can run on modern GPU hardware.

Dunkley, Kimieka [Florida A-M]↗