Search NASA⌕ Search

SEARCH · Search NASA

Results for “Accelerators, Computing”

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 1,729 records · Page 96

WE-G-BRD-09: Novel MRI Compatible Electron Accelerator for MRI-Linac Radiotherapy

Purpose: MRI guided radiotherapy is a rapidly growing field; however current linacs are not designed to operate in MRI fringe fields. As such, current MRI- Linac systems require magnetic shielding, impairing MR image quality and system flexibility. Here, we present a bespoke electron accelerator concept with robust operation in in-line magnetic fields. Methods: For in-line MRI-Linac systems, electron gun performance is the major constraint on accelerator performance. To overcome this, we propose placing a cathode directly within the first accelerating cavity. Such a configuration is used extensively in high energy particle physics, but not previously for radiotherapy. Benchmarked computational modelling (CST, Darmstadt, Germany) was employed to design and assess a 5.5 cell side coupled accelerator with a temperature limited thermionic cathode in the first accelerating cell. This simulation was coupled to magnetic fields from a 1T MRI model to assess robustness in magnetic fields for Source to Isocenter Distance between 1 and 2 meters. Performance was compared to a conventional electron gun based system in the same magnetic field. Results: A temperature limited cathode (work function 1.8eV, temperature 1245K, emission constant 60A/K/cm{sup 2}) will emit a mean current density of 24mA/mm{sup 2} (Richardson’s Law). We modeled a circular cathode with radius 2mm and mean current 300mA. Capture efficiency of the device was 43%, resulting in target current of 130 mA. The electron beam had a FWHM of 0.2mm, and mean energy of 5.9MeV (interquartile spread of 0.1MeV). Such an electron beam is suitable for radiotherapy, comparing favourably to conventional systems. This model was robust to operation the MRI fringe field, with a maximum current loss of 6% compared to 85% for the conventional system. Conclusion: The bespoke electron accelerator is robust to operation in in-line magnetic fields. This will enable MRI-Linacs with no accelerator magnetic shielding, and minimise painstaking optimisation of the MRI fringe field. This work was supported by US (NIH) and Australian (NHMRC & Cancer Institute NSW) government research funding. In addition, I would like to thank cancer institute NSW and the Ingham Institute for scholarship support.

60 APPLIED LIFE SCIENCES↗

Performance Optimization of a Beam Dynamics PIC Code On Hybrid Computer Architectures

The self-consistent multi-particle tracking based on particle-in-cell method (PIC) has been widely used in particle accelerator beam dynamics study. However, the PIC simulation is time-consuming and needs to use modern parallel computers for high resolution applications. In this paper, we implemented and optimized a parallel beam dynamics PIC code on two types of hybrid parallel computer architectures: one is the GPU and GPU cluster, while the other is the "Knight Landing" CPU cluster.

Liu, Zhicong↗

A chain-of-states acceleration method for the efficient location of minimum energy paths

We describe a robust and efficient chain-of-states method for computing Minimum Energy Paths (MEPs) associated to barrier-crossing events in poly-atomic systems, which we call the acceleration method. The path is parametrized in terms of a continuous variable t ∈ [0, 1] that plays the role of time. In contrast to previous chain-of-states algorithms such as the nudged elastic band or string methods, where the positions of the states in the chain are taken as variational parameters in the search for the MEP, our strategy is to formulate the problem in terms of the second derivatives of the coordinates with respect to t, i.e., the state accelerations. We show this to result in a very simple and efficient method for determining the MEP. We describe the application of the method to a series of test cases, including two low-dimensional problems and the Stone-Wales transformation in C{sub 60}.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Defining quantum-ready primitives for hybrid HPC-QC supercomputing: a case study in Hamiltonian simulation

As computational demands in scientific applications continue to rise, hybrid high-performance computing (HPC) systems integrating classical and quantum computers (HPC-QC) are emerging as a promising approach to tackling complex computational challenges. One critical area of application is Hamiltonian simulation, a fundamental task in quantum physics and other large-scale scientific domains. This paper investigates strategies for quantum-classical integration to enhance Hamiltonian simulation within hybrid supercomputing environments. By analyzing computational primitives in HPC allocations dedicated to these tasks, we identify key components in Hamiltonian simulation workflows that stand to benefit from quantum acceleration. To this end, we systematically break down the Hamiltonian simulation process into discrete computational phases, highlighting specific primitives that could be effectively offloaded to quantum processors for improved efficiency. Our empirical findings provide insights into system integration, potential offloading techniques, and the challenges of achieving seamless quantum-classical interoperability. We assess the feasibility of quantum-ready primitives within HPC workflows and discuss key barriers such as synchronization, data transfer latency, and algorithmic adaptability. These results contribute to the ongoing development of optimized hybrid solutions, advancing the role of quantum-enhanced computing in scientific research.

97 MATHEMATICS AND COMPUTING↗

Bayesian calibration of bubble size dynamics applied to CO 2 gas fermenters

To accelerate the scale-up of gaseous CO 2 fermentation reactors, computational models need to predict gas-to-liquid mass transfer which requires capturing the bubble size dynamics, i.e. bubble breakup and coalescence. However, the applicability of existing models beyond air–water mixtures remains to be established. Here, an inverse modeling approach, accelerated with a neural network surrogate, calibrates the breakup and coalescence closure models, that are used in class methods for population balance modeling (PBM). The calibration is performed based on experimental results obtained in a CO 2 -air–water-coflowing bubble column reactor. Bayesian inference is used to account for noise in the experimental dataset and bias in the simulation results. To accurately capture gas holdup and interphase mass transfer, the results show that the breakage rate needs to be increased by one order of magnitude. In conclusion, the inferred model parameters are then used on a separate configuration and shown to also improve bubble size distribution predictions.

09 BIOMASS FUELS↗

A Benchmark Suite for Evaluating Scientific AI Workloads on GPUs

AI applications have been steadily increasing in the allocation portfolio among leadership computing facilities. These applications depend on deep learning frameworks with hardware acceleration and underlying software systems. With the rapid development of applications, software stacks, and hardware devices, it is essential to evaluate the performance of core operations in AI workloads for direction of optimizations and procurement of next-generation high-performance computing (HPC) infrastructures. Currently, most benchmarks lack scientific AI workloads. So, we present DeepKernelBench and the experimental results of evaluating the benchmark suite for early observations and performance comparisons on datacenter GPUs using representative workloads for scientific AI, including Attentions, General matrix multiplications, Geometrics and Fourier neural operations.

Jin, Zheming [Advanced Micro Devices (AMD)]↗

A time-parallel multiple-shooting method for large-scale quantum optimal control

Quantum optimal control plays a crucial role in quantum computing by providing the interface between compiler and hardware. Solving the optimal control problem is particularly challenging for multi-qubit gates, due to the exponential growth in computational complexity with the system's dimensionality and the deterioration of optimization convergence. To ameliorate the computational complexity of time-integration, this paper introduces a multiple-shooting approach in which the time domain is divided into multiple windows and the intermediate states at window boundaries are treated as additional optimization variables. Further, this enables parallel computation of state evolution across time-windows, significantly accelerating objective function and gradient evaluations. Since the initial state matrix in each window is only guaranteed to be unitary upon convergence of the optimization algorithm, the conventional gate trace infidelity is replaced by a generalized infidelity that is convex for non-unitary state matrices. Continuity of the state across window boundaries is enforced by equality constraints. A quadratic penalty optimization method is used to solve the constrained optimal control problem, and an efficient adjoint technique is employed to calculate the gradients in each iteration. We demonstrate the effectiveness of the proposed method through numerical experiments on quantum Fourier transform gates in systems with 2, 3, and 4 qubits, noting a speedup of 80x for evaluating the gradient in the 4-qubit case, highlighting the method's potential for optimizing control pulses in multi-qubit quantum systems.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Toward performance-portable PETSc for GPU-based exascale systems

The Portable Extensible Toolkit for Scientific computation (PETSc) library delivers scalable solvers for nonlinear time-dependent differential and algebraic equations and for numerical optimization. The PETSc design for performance portability addresses fundamental GPU accelerator challenges and stresses flexibility and extensibility by separating the programming model used by the application from that used by the library, and it enables application developers to use their preferred programming model, such as Kokkos, RAJA, SYCL, HIP, CUDA, or OpenCL, on upcoming exascale systems. Furthermore, a blueprint for using GPUs from PETSc-based codes is provided, and case studies emphasize the flexibility and high performance achieved on current GPU-based systems.

97 MATHEMATICS AND COMPUTING↗

A Case For Intra-rack Resource Disaggregation in HPC

The expected halt of traditional technology scaling is motivating increased heterogeneity in high-performance computing (HPC) systems with the emergence of numerous specialized accelerators. As heterogeneity increases, so does the risk of underutilizing expensive hardware resources if we preserve today’s rigid node configuration and reservation strategies. This has sparked interest in resource disaggregation to enable finer-grain allocation of hardware resources to applications. However, there is currently no data-driven study of what range of disaggregation is appropriate in HPC. To that end, we perform a detailed analysis of key metrics sampled in NERSC’s Cori, a production HPC system that executes a diverse open-science HPC workload. In addition, we profile a variety of deep-learning applications to represent an emerging workload. We show that for a rack (cabinet) configuration and applications similar to Cori, a central processing unit with intra-rack disaggregation has a 99.5% probability to find all resources it requires inside its rack. In addition, ideal intra-rack resource disaggregation in Cori could reduce memory and NIC resources by 5.36% to 69.01% and still satisfy the worst-case average rack utilization.

97 MATHEMATICS AND COMPUTING↗

GPU Acceleration of Large-Scale Full-Frequency GW Calculations

Many-body perturbation theory is a powerful method to simulate electronic excitations in molecules and materials starting from the output of density functional theory calculations. By implementing the theory efficiently so as to run at scale on the latest leadership high-performance computing systems it is possible to extend the scope of GW calculations. Here, we present a GPU acceleration study of the full-frequency GW method as implemented in the WEST code. Excellent performance is achieved through the use of (i) optimized GPU libraries, e.g., cuFFT and cuBLAS, (ii) a hierarchical parallelization strategy that minimizes CPU-CPU, CPU-GPU, and GPU-GPU data transfer operations, (iii) nonblocking MPI communications that overlap with GPU computations, and (iv) mixed precision in selected portions of the code. A series of performance benchmarks has been carried out on leadership high-performance computing systems, showing a substantial speedup of the GPU-accelerated version of WEST with respect to its CPU version. Good strong and weak scaling is demonstrated using up to 25 920 GPUs. Finally, we showcase the capability of the GPU version of WEST for large-scale, full-frequency GW calculations of realistic systems, e.g., a nanostructure, an interface, and a defect, comprising up to 10 368 valence electrons.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Experimental Physical Sciences Vistas: MaRIE (draft)

To achieve breakthrough scientific discoveries in the 21st century, a convergence and integration of world-leading experimental facilities and capabilities with theory, modeling, and simulation is necessary. In this issue of Experimental Physical Sciences Vistas, I am excited to present our plans for Los Alamos National Laboratory's future flagship experimental facility, MaRIE (Matter-Radiation Interactions in Extremes). MaRIE is a facility that will provide transformational understanding of matter in extreme conditions required to reduce or resolve key weapons performance uncertainties, develop the materials needed for advanced energy systems, and transform our ability to create materials by design. Our unique role in materials science starting with the Manhattan Project has positioned us well to develop a contemporary materials strategy pushing the frontiers of controlled functionality - the design and tailoring of a material for the unique demands of a specific application. Controlled functionality requires improvement in understanding of the structure and properties of materials in order to synthesize and process materials with unique characteristics. In the nuclear weapons program today, improving data and models to increase confidence in the stockpile can take years from concept to new knowledge. Our goal with MaRIE is to accelerate this process by enhancing predictive capability - the ability to compute a priori the observables of an experiment or test and pertinent confidence intervals using verified and validated simulation tools. It is a science-based approach that includes the use of advanced experimental tools, theoretical models, and multi-physics codes, simultaneously dealing with multiple aspects of physical operation of a system that are needed to develop an increasingly mature predictive capability. This same approach is needed to accelerate improvements to other systems such as nuclear reactors. MaRIE will be valuable to many national security science challenges. Our first issue of Vistas focused on our current national user facilities (the Los Alamos Neutron Science Center [LANSCE], the National High Magnetic Field Laboratory-Pulsed Field Facility, and the Center for Integrated Nanotechnologies) and the vitality they bring to our Laboratory. These facilities are a magnet for students, postdoctoral researchers, and staff members from all over the world. This, in turn, allows us to continue to develop and maintain our strong staff across the relevant disciplines and conduct world-class discovery science. The second issue of Vistas was devoted entirely to the Laboratory's materials strategy - one of the three strategic science thrusts for the Laboratory. This strategy has helped focus our thinking for MaRIE. We believe there is a bright future in cutting-edge experimental materials research, and that a 21st-century facility with unique capability is necessary to fulfill this goal. The Laboratory has spent the last several years defining MaRIE, and this issue of Vistas presents our current vision of that facility. MaRIE will leverage LANSCE and our other user facilities, as well as our internal and external materials community for decades to come, giving Los Alamos a unique competitive advantage, advancing materials science for the Laboratory's missions and attracting and recruiting scientists of international stature. MaRIE will give the international materials research community a suite of tools capable of meeting a broad range of outstanding grand challenges.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

GPU-resident sparse direct linear solvers for alternating current optimal power flow analysis

Integrating renewable resources within the transmission grid at a wide scale poses significant challenges for economic dispatch as it requires analysis with more optimization parameters, constraints, and sources of uncertainty. This motivates the investigation of more efficient computational methods, especially those for solving the underlying linear systems, which typically take more than half of the overall computation time. In this paper, we present our work on sparse linear solvers that take advantage of hardware accelerators, such as graphical processing units (GPUs), and improve the overall performance when used within economic dispatch computations. We treat the problems as sparse, which allows for faster execution but also makes the implementation of numerical methods more challenging. We present the first GPU-native sparse direct solver that can execute on both AMD and NVIDIA GPUs. We demonstrate significant performance improvements when using high-performance linear solvers within alternating current optimal power flow (ACOPF) analysis. Furthermore, we demonstrate the feasibility of getting significant performance improvements by executing the entire computation on GPU-based hardware. Finally, we identify outstanding research issues and opportunities for even better utilization of heterogeneous systems, including those equipped with GPUs.

24 POWER TRANSMISSION AND DISTRIBUTION↗

3D-ReG: A 3D ReRAM-based Heterogeneous Architecture for Training Deep Neural Networks

Deep neural network (DNN) models are being expanded to a broader range of applications. The computational capability of traditional hardware platforms cannot accommodate the growth of model complexity. Among recent technologies to accelerate DNN, resistive memory (ReRAM)-based processing-in-memory (PIM) emerged as a promising solution for DNN inference due to its high efficiency for matrix-based computation. We face two major technical challenges in extending the use of ReRAM-based accelerators for training: (1) full-precision data is essential in back-propagation; (2) the need to support both feed-forward and back-propagation aggravates the data-movement burden. We propose a heterogeneous architecture named as 3D-ReG, which leverages full-precision GPU to ensure training accuracy and low-overhead 3D integration to provide low-cost data movements. Moreover, we introduce conservative and aggressive task-mapping schemes, which partition the computation phases in different ways to balance execution efficiency and training accuracy. We evaluate 3D-ReG implemented with two 3D integration technologies, through-silicon vias (TSVs) and monolithic inter-tier vias (MIVs), and compare them with GPU-only and PIM-only counterparts. Various GPU-only platforms using two main-memory technologies (DRAM, ReRAM) and three interconnect technologies (2D, TSV, MIV) are evaluated as well. Experimental results show that 3D-ReG can achieve on average 5.64× training speedup and 3.56× higher energy efficiency compared with the GPU with DRAM as main memory, at the cost of 0.05%–3.39% accuracy drop. We define a new metric, gain-loss ratio (GLR), which quantitatively evaluates the capability of a DNN training hardware in terms of the model accuracy and hardware efficiency. The results of our comparison show that the aggressive task-mapping scheme on MIV-based 3D-ReG outperforms the other methods.

Computer Science↗

Study of High-Frequency Impedance of Small-Angle Tapers and Collimators

Collimators and transitions in accelerator vacuum chambers often include small-angle tapering to lower the wakefields generated by the beam. While the low-frequency impedance is well described by Yokoya's formula (for axisymmetric geometry), much less is known about the behavior of the impedance in the high frequency limit. In this paper we develop an analytical approach to the highfrequency regime for round collimators and tapers. Our analytical results are compared with computer simulations using the code ECHO.

43 PARTICLE ACCELERATORS↗

Anderson acceleration and application to the three-temperature energy equations

The Anderson acceleration method is an algorithm for accelerating the convergence of fixed-point iterations, including the Picard method. Anderson acceleration was first proposed in 1965 and, for some years, has been used successfully to accelerate the convergence of self-consistent field iterations in electronic-structure computations. Recently, the method has attracted growing attention in other application areas and among numerical analysts. Compared with a Newton-like method, an advantage of Anderson acceleration is that there is no need to form the Jacobian matrix. Thus the method is easy to implement. In this paper, an Anderson-accelerated Picard method is employed to solve the three-temperature energy equations, which are a type of strong nonlinear radiation-diffusion equations. Two strategies are used to improve the robustness of the Anderson acceleration method. One strategy is to adjust the iterates when necessary to satisfy the physical constraint. Another strategy is to monitor and, if necessary, reduce the matrix condition number of the least-squares problem in the Anderson-acceleration implementation so that numerical stability can be guaranteed. Numerical results show that the Anderson-accelerated Picard method can solve the three-temperature energy equations efficiently. Compared with the Picard method without acceleration, Anderson acceleration can reduce the number of iterations by at least half. A comparison between a Jacobian-free Newton–Krylov method, the Picard method, and the Anderson-accelerated Picard method is conducted in this paper.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Advanced Methods for the Computation of Particle Beam Transport and the Computation of Electromagnetic Fields and Multiparticle Phenomena

Since 1980, under the grant DEFG02-96ER40949, the Department of Energy has supported the educational and research work of the University of Maryland Dynamical Systems and Accelerator Theory (DSAT) Group. The primary focus of this educational/research group has been on the computation and analysis of charged-particle beam transport using Lie algebraic methods, and on advanced methods for the computation of electromagnetic fields and multiparticle phenomena. This Final Report summarizes the accomplishments of the DSAT Group from its inception in 1980 through its end in 2011.

43 PARTICLE ACCELERATORS↗