Engineering Papers⌕ Search

SEARCH · Engineering Papers

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 397 records · Page 22

Wakefield Computations for the CLIC PETS using the Parallel Finite Element Time-Domain Code T3P

In recent years, SLAC's Advanced Computations Department (ACD) has developed the high-performance parallel 3D electromagnetic time-domain code, T3P, for simulations of wakefields and transients in complex accelerator structures. T3P is based on advanced higher-order Finite Element methods on unstructured grids with quadratic surface approximation. Optimized for large-scale parallel processing on leadership supercomputing facilities, T3P allows simulations of realistic 3D structures with unprecedented accuracy, aiding the design of the next generation of accelerator facilities. Applications to the Compact Linear Collider (CLIC) Power Extraction and Transfer Structure (PETS) are presented.

43 PARTICLE ACCELERATORS↗

Efficient Minimum-Polynomial And Reduced-Rank Extrapolation

MPERRE computer program accelerates convergence of sequence of vectors by use of minimum-polynomial extrapolation (MPE) and reduced-rank extrapolation (RRE). Effective in accelerating convergences of such sequences of vectors as those obtained from iterative solution of systems of linear and nonlinear equations. In conjunction with various iterative techniques, successfully employed in finite-difference solution of large-scale elliptic boundary-value problems and problems in computational-fluid-dynamics. Only input required is sequence of vectors, convergence of which is accelerated. Program economical and easy to use. Written in FORTRAN 77.

Sidi, Avram↗

Improvements of the boundary projection acceleration technique applied to the discrete-ordinates transport solver in XYZ geometries

In this paper we discuss the latest upgrades of the Boundary Projection Acceleration (BPA) applied to the XYZ transport solver of APOLLO3, namely IDT. The acceleration method is a well-known effective technique for the speed-up of the source iterations of the discrete-ordinates method. The BPA in IDT has been improved in three aspects: the taking into account of the residue on boundary conditions as a boundary source for the acceleration problem, the extension of the method to higher order angular moments in the case of anisotropic scattering and, finally, the application of the method to the multigroup iterations for the acceleration of the fission source and k-effective. The spectrum of the method has been Fourier-analyzed to explore the effectiveness. The 3D mock-up geometry of the ZPPR is presented as final study to test the performances of the acceleration on a realistic whole-core 3D calculation. (authors)

97 MATHEMATICS AND COMPUTING↗

Multigrid acceleration and turbulence models for computations of 3D turbulent jets in crossflow

A multigrid method is presented for the calculation of three-dimensional turbulent jets in crossflow. Turbulence closure is achieved with either the standard k-epsilon model or a Reynolds Stress Model (RSM). Multigrid acceleration enables convergence rates which are far superior to that for a single grid method. With the k-epsilon model the rate approaches that for laminar flow, but with RSM it is somewhat slower. The increased stiffness of the system of equations in the latter may be responsible. Computed results with both turbulence models are compared with experimental data for a pair of opposed jets in crossflow. Both models yield reasonable agreement with mean flow velocity but RSM yields better prediction of the Reynolds stresses.

Demuren, A. O.↗

Cold Optical Injection Producing Monoenergetic, Multi-GeV Electron Bunches

A cold optical injection mechanism for a laser-plasma accelerator is described. It relies on a short, circularly polarized, low-energy laser pulse counterpropagating to and colliding with a circularly polarized main pulse in a low density plasma. Contrary to previously published optical injection schemes, injection is not caused here by electron heating. Instead, the collision between the pulses creates a spatially periodic and time-independent beat force. This force can block the longitudinal electron motion, leading to their entry and injection into the propagating wake. In a specific setup, we compute after acceleration over 0.6 mm, a 60 MeV, 50 pC electron bunch with 0.7 MeV rms energy spread, proving the interest of this scheme to inject electron bunches with a narrow absolute energy spread. Acceleration to 3 GeV with a rms spread smaller than 1% is computed after propagation over 3.8 cm in a plasma channel.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Evaluation of a multigrid-based Navier-Stokes solver for aerothermodynamic computations

A multigrid acceleration technique developed for solving the three-dimensional Navier-Stokes equations is used for computing high Mach number flows over configurations of practical interest. An explicit multistage Runge-Kutta type of time-stepping scheme is used as the basic algorithm. Solutions are presented for a spherically blunted cone at Mach 10 and a modified shuttle orbiter at Mach 6. The computed surface heat-transfer distributions are shown to compare favorably with the experimental data. Effect of grid-refinement on computed heat-transfer distributions is also examined to assess the numerical accuracy of the computed solutions. The rapid convergence rate associated with multigrid schemes in previous applications at transonic speeds is observed at the higher Mach number flows investigated here.

Vatsa, Veer N.↗

Multigrid acceleration and turbulence models for computations of 3D turbulent jets in crossflow

A multigrid method is presented for the calculation of three-dimensional turbulent jets in crossflow. Turbulence closure is achieved with either the standard k-epsilon model or a Reynolds stress model (RSM). Multigrid acceleration enables convergence rates which are far superior to that for a single grid method to be obtained with both turbulence models. With the k-epsilon model the rate approaches that for laminar flow, but with RSM it is somewhat slower. The increased stiffness of the system of equation in the latter may be responsible. Computed results with both turbulence models are compared to experimental data for a pair of opposed jets in crossflow. Both models yield reasonable agreement for the mean flow velocity, but RSM yields better predictions of the Reynolds stresses.

Demuren, A. O.↗

Revolutionizing Neuromorphic Computing for Science (Brochure on the 2024 ASCR Workshop on Neuromorphic Computing for Science)

The ASCR basic research needs for Neuromorphic Computing for Science workshop was held in September 2024. The workshop brochure and report aim to inform and draft a set of grand challenges for advancing the field of neuromorphic computing and developing proof of principle neuromorphic circuits applicable for High Performance Computer (HPC) acceleration for scientific discovery, and brainstorm ideas needed for a successful, robust, and world leading basic research program. The resulting priority research directions are: (1) Neuromorphic computing circuit primitives; (2) Connectivity, communication, and hardware integration; (3) Neuroscience-derived dynamics and algorithms; and (4) Ecosystem for scalable neuromorphic co-design. Breakthroughs in understanding, designing, and prototyping the circuitry and simulation capabilities for a truly neuromorphic computer are essential to enable progress in the field.

97 MATHEMATICS AND COMPUTING↗

Revolutionizing Neuromorphic Computing for Science (Report for the 2024 ASCR Workshop on Neuromorphic Computing for Science)

The ASCR basic research needs for Neuromorphic Computing for Science workshop was held in September 2024. The workshop brochure and report aim to inform and draft a set of grand challenges for advancing the field of neuromorphic computing and developing proof of principle neuromorphic circuits applicable for High Performance Computer (HPC) acceleration for scientific discovery, and brainstorm ideas needed for a successful, robust, and world leading basic research program. The resulting priority research directions are: (1) Neuromorphic computing circuit primitives; (2) Connectivity, communication, and hardware integration; (3) Neuroscience-derived dynamics and algorithms; and (4) Ecosystem for scalable neuromorphic co-design. Breakthroughs in understanding, designing, and prototyping the circuitry and simulation capabilities for a truly neuromorphic computer are essential to enable progress in the field.

97 MATHEMATICS AND COMPUTING↗

ComPort: Rigorous Testing Methods to Safeguard Software Porting (Final Technical Report)

This is a technical report from the lead institution – University of Utah, Kahlert School of Computing – funded under the Department of Energy, Office of Science, Office of Advanced Scientific Computing Research under award number DE-SC0022252. We summarize our work done over the three years of funding received. The relevant papers and software have already been uploaded at the DOE site.

97 MATHEMATICS AND COMPUTING↗

Convergence acceleration of iterative solutions of Euler equations for transonic flow computations

In this paper, two acceleration techniques for Euler calculations are investigated. The first technique is an extrapolation procedure based on the Power Method; it is applicable when the iterative matrix has dominant eigenvalues. Both real and complex conjugate roots are allowed. The second technique is a generalization of the Minimal Residual Method, where the extrapolation step consists of a weighted combination of the corrections at different iteration levels and the weights are chosen to minimize the Lz norm of the residual. Numerical results, using Jameson's Runge-Kutta Multigrid Code, are presented. The extra computational work to apply either technique is negligible and the extra storage is not a problem on current supercomputers.

Hafez, M.↗

Visual and motion cueing in helicopter simulation

Early experience in fixed-cockpit simulators, with limited field of view, demonstrated the basic difficulties of simulating helicopter flight at the level of subjective fidelity required for confident evaluation of vehicle characteristics. More recent programs, utilizing large-amplitude cockpit motion and a multiwindow visual-simulation system have received a much higher degree of pilot acceptance. However, none of these simulations has presented critical visual-flight tasks that have been accepted by the pilots as the full equivalent of flight. In this paper, the visual cues presented in the simulator are compared with those of flight in an attempt to identify deficiencies that contribute significantly to these assessments. For the low-amplitude maneuvering tasks normally associated with the hover mode, the unique motion capabilities of the Vertical Motion Simulator (VMS) at Ames Research Center permit nearly a full representation of vehicle motion. Especially appreciated in these tasks are the vertical-acceleration responses to collective control. For larger-amplitude maneuvering, motion fidelity must suffer diminution through direct attenuation through high-pass filtering washout of the computer cockpit accelerations or both. Experiments were conducted in an attempt to determine the effects of these distortions on pilot performance of height-control tasks.

Bray, R. S.↗

Visual and motion cueing in helicopter simulation

The visual cues presented in the simulator are compared with those of flight in an attempt to identify deficiencies. For the low-amplitude maneuvering tasks normally associated with the hover mode, the unique motion capabilities of the Vertical Motion Simulator (VMS) at Ames Research Center permit nearly a full representation of vehicle motion. Especially appreciated in these tasks are the vertical-acceleration responses to collective control. For larger-amplitude maneuvering, motion fidelity must suffer diminution through direct attenuation through high-pass filtering washout of the computer cockpit accelerations or both. Experiments were conducted in an attempt to determine the effects of these distortions on pilot performance of height-control tasks.

Bray, Richard S.↗

Collaborating CPU and GPU for large-scale high-order CFD simulations with complex grids on the TianHe-1A supercomputer

Programming and optimizing complex, real-world CFD codes on current many-core accelerated HPC systems is very challenging, especially when collaborating CPUs and accelerators to fully tap the potential of heterogeneous systems. In this paper, with a tri-level hybrid and heterogeneous programming model using MPI + OpenMP + CUDA, we port and optimize our high-order multi-block structured CFD software HOSTA on the GPU-accelerated TianHe-1A supercomputer. HOSTA adopts two self-developed high-order compact definite difference schemes WCNS and HDCS that can simulate flows with complex geometries. We present a dual-level parallelization scheme for efficient multi-block computation on GPUs and perform particular kernel optimizations for high-order CFD schemes. The GPU-only approach achieves a speedup of about 1.3 when comparing one Tesla M2050 GPU with two Xeon X5670 CPUs. To achieve a greater speedup, we collaborate CPU and GPU for HOSTA instead of using a naive GPU-only approach. We present a novel scheme to balance the loads between the store-poor GPU and the store-rich CPU. Taking CPU and GPU load balance into account, we improve the maximum simulation problem size per TianHe-1A node for HOSTA by 2.3×, meanwhile the collaborative approach can improve the performance by around 45% compared to the GPU-only approach. Further, to scale HOSTA on TianHe-1A, we propose a gather/scatter optimization to minimize PCI-e data transfer times for ghost and singularity data of 3D grid blocks, and overlap the collaborative computation and communication as far as possible using some advanced CUDA and MPI features. Scalability tests show that HOSTA can achieve a parallel efficiency of above 60% on 1024 TianHe-1A nodes. With our method, we have successfully simulated an EET high-lift airfoil configuration containing 800M cells and China's large civil airplane configuration containing 150M cells. To our best knowledge, those are the largest-scale CPU–GPU collaborative simulations that solve realistic CFD problems with both complex configurations and high-order schemes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Analysis of the Cause of High External Q Modes in the JLab High Gradient Prototype Cryomodule Renascence

The Renascence cryomodule [1] installed in CEBAF in 2007 consists of 8 cavities as shown in Figure 1. The first three cavities (No.1-No.3) in the upstream end are of the Low Loss (LL) shape design, and the remaining 5 cavities (No.4-No.8) on the beam downstream end are the High Gradient (HG) shape design. The fundamental power couplers (FPCs) are the rectangular waveguides, and the little cylindrical structures are the HOM couplers. The locations of the FPC in the last four cavities are mirrored about the beam z axis. Cavities No.4 and No.5 form a back-to-back cavity pair. Among the HG cavities installed in the Renascence cryomodule, the only identifiable difference from their fabrication documentation is that cavity No.5 received an extra EBW pass on one equator weld, specifically cell 5. The non-uniform mechanical tuning required to compensate the fundamental mode tune and flatness for the extra shrinkage of this cell is believed to contribute the most significant differences from the other HG cavities. Beam based instability studies on this cryomodule in CEBAF have shown a significant beam breakup (BBU) threshold current reduction, well below design value. Frequency spectrum peaked by the off-sided beam power indicated the cause is due to abnormal high Q modes in the cavity No.5. Measured beam off-axis position at the cavity No.5 does not correspond to the shunt impedances calculated for an ideal cavity. Low power RF measurements have identified that the problematic modes are in the second dipole band (TM110 like). Three of the modes have external Qs two orders magnitude higher than the others, while the rest of modes in the first two dipole bands are normal in terms of the design values. The cause of this abnormality and the future impact on the BBU was not able to be resolved due to the limitations of information that can be obtained from the measurements. It is important to understand the cause of this abnormality so that effective QA/QC measures can be implemented to avoid such problem in the final upgrade design and manufacture. The goal of this work is to utilize advanced simulation tools to understand the high external Q (Q{sub ext}) problem observed in the Renascence cryomodule. In the past years, SLAC has built a set of state-of-the-art advanced simulation tools based on finite-element unstructured meshes and parallel computation implementations on supercomputers [2, 3]. The codes are capable of simulating large complex RF systems with unprecedented resolution and turnaround time. They have been successfully applied to many existing and future accelerator R&D projects to improve the machine performance and to optimize the designs. These tools are essential to perform accurate full system analyses such as the JLab's SRF cavities. We will use the simulation results and the data from the RF measurements to gain a better understanding of the cavity performance and tolerance issues and provide a solid foundation to do the BBU simulation and prediction for the 12GeV Upgrade project by using JLab's BBU codes. In this report, we will focus on the following two main tasks: (1) Ideal cavity simulation--to evaluate the effectiveness of the damping by the higher-order-mode (HOM) couplers, and search for possible trapped modes in a back-to-back cavity pair (e.g. cavity No.4 & No.5). (2) Abnormal cavity study--to understand the cause of the high Q{sub ext} modes in cavity No.5 using an advanced Shape Determination Tool.

43 PARTICLE ACCELERATORS↗

GPU-Accelerated Denoising in 3D (GD3D)

The raw computational power GPU Accelerators enables fast denoising of 3D MR images using bilateral filtering, anisotropic diffusion, and non-local means. This software addresses two facets of this promising application: what tuning is necessary to achieve optimal performance on a modern GPU? And what parameters yield the best denoising results in practice? To answer the first question, the software performs an autotuning step to empirically determine optimal memory blocking on the GPU. To answer the second, it performs a sweep of algorithm parameters to determine the combination that best reduces the mean squared error relative to a noiseless reference image.

Howison, Mark↗

GPU-Accelerated Long-Term Simulations of Beam-Beam Effects in Colliders

We present an update on the development of the new code for long-term simulation of beam-beameffects in particle colliders. The underlying physical model relies on a matrix-based arbitrary-order particle tracking (including a symplectic option) for beam transport and the generalized Bassetti-Erskine approximation for beam-beam interaction. The computations are accelerated through a parallel implementation on a hybrid GPU/CPU platform. With the new code, previously computationally prohibitive longterm simulations become tractable. The new code will be used to model the proposed Medium-energy Electron-Ion Collider (MEIC) at Jefferson Lab.

Terzic, B.↗

Hands-on Performance Tuning of 3D Finite Difference Earthquake Simulation on GPU Fermi Chipset

3D simulation of earthquake ground motion is one of the most challenging computational problems in science. The emergence of graphic processing units (GPU) as an effective alternative to traditional general purpose processors has become increasingly capable in terms of accelerating scientific computing research. In this paper, we describe our experiences in porting AWP-ODC, a 3D finite difference seismic wave propagation code, to the latest GPU Fermi chipset. We completely rewrote this Fortran-based 13-point asymmetric stencil computation code in C and MPI-CUDA in order to take advantage of the powerful GPU computing capabilities. Our new CUDA code implemented the asymmetric 3D stencil on Fermi to make the best use of GPU on-chip memory for an aggressive parallel efficiency. Benchmark on NVIDIA Tesla M2090 demonstrated 10x speedup versus the original fully optimized AWP-ODC FORTRAN MPI code running on a single Intel Nehalem 2.4 GHz CPU socket (4 cores/CPU), and 15x speedup versus the same MPI code running on a single AMD Istanbul 2.6 GHz CPU socket (6 cores/CPU). Sustained single-GPU performance of 143.8 GFLOPS in single precision is benchmarked for the testing case of 128x128x960 mesh size.

3D Stencil Computation↗