Showing 1–6 of 6 results for author: Beidler, M

A divergence-free constrained magnetic field interpolation method for scattered data

Authors: Minglei Yang, Diego del-Castillo-Negrete, Guannan Zhang, Matthew Beidler

Abstract: An interpolation method to evaluate magnetic fields given unstructured, scattered magnetic data is presented. The method is based on the reconstruction of the global magnetic field using a superposition of orthogonal functions. The coefficients of the expansion are obtained by minimizing a cost function defined as the L^2 norm of the difference between the ground truth and the reconstructed magnet… ▽ More An interpolation method to evaluate magnetic fields given unstructured, scattered magnetic data is presented. The method is based on the reconstruction of the global magnetic field using a superposition of orthogonal functions. The coefficients of the expansion are obtained by minimizing a cost function defined as the L^2 norm of the difference between the ground truth and the reconstructed magnetic field evaluated on the training data. The divergence-free condition is incorporated as a constrain in the cost function allowing the method to achieve arbitrarily small errors in the magnetic field divergence. An exponential decay of the approximation error is observed and compared with the less favorable algebraic decay of local splines. Compared to local methods involving computationally expensive search algorithms, the proposed method exhibits a significant reduction of the computational complexity of the field evaluation, while maintaining a small error in the divergence even in the presence of magnetic islands and stochasticity. Applications to the computation of Poincaré sections using data obtained from numerical solutions of the magnetohydrodynamic equations in toroidal geometry are presented and compared with local methods currently in use. △ Less

Submitted 25 July, 2022; originally announced July 2022.

Answer Fast: Accelerating BERT on the Tensor Streaming Processor

Authors: Ibrahim Ahmed, Sahil Parmar, Matthew Boyd, Michael Beidler, Kris Kang, Bill Liu, Kyle Roach, John Kim, Dennis Abts

Abstract: Transformers have become a predominant machine learning workload, they are not only the de-facto standard for natural language processing tasks, but they are also being deployed in other domains such as vision and speech recognition. Many of the transformer-based applications are real-time systems such as machine translation and web search. These real time systems often come with strict end-to-end… ▽ More Transformers have become a predominant machine learning workload, they are not only the de-facto standard for natural language processing tasks, but they are also being deployed in other domains such as vision and speech recognition. Many of the transformer-based applications are real-time systems such as machine translation and web search. These real time systems often come with strict end-to-end inference latency requirements. Unfortunately, while the majority of the transformer computation comes from matrix multiplications, transformers also include several non-linear components that tend to become the bottleneck during an inference. In this work, we accelerate the inference of BERT models on the tensor streaming processor. By carefully fusing all the nonlinear components with the matrix multiplication components, we are able to efficiently utilize the on-chip matrix multiplication units resulting in a deterministic tail latency of 130 $μ$s for a batch-1 inference through BERT-base, which is 6X faster than the current state-of-the-art. △ Less

Submitted 22 June, 2022; originally announced June 2022.

Spatially-dependent modeling and simulation of runaway electron mitigation in DIII-D

Authors: M. T. Beidler, D. del-Castillo-Negrete, L. R. Baylor, D. Shiraki, D. A. Spong

Abstract: New simulations with the Kinetic Orbit Runaway electron (RE) Code KORC show RE deconfinement losses to the wall during plasma scrape off are the primary current dissipation mechanism in DIII-D experiments with high-Z impurity injection, and not collisional slowing down. The majority of simulations also exhibit an increase in the RE beam energy due to acceleration by the induced toroidal electric f… ▽ More New simulations with the Kinetic Orbit Runaway electron (RE) Code KORC show RE deconfinement losses to the wall during plasma scrape off are the primary current dissipation mechanism in DIII-D experiments with high-Z impurity injection, and not collisional slowing down. The majority of simulations also exhibit an increase in the RE beam energy due to acceleration by the induced toroidal electric field, even while the RE beam current is decreasing. In this study, KORC integrates RE orbits using the relativistic guiding center equations of motion, and incorporates time-sequenced, experimental reconstructions of the magnetic and electric fields and line integrated electron density to construct spatiotemporal models of electron and partially-ionized impurity transport in the companion plasma. Comparisons of experimental current evolution and KORC results demonstrate the importance of including Coulomb collisions with partially-ionized impurity physics, initial RE energy, pitch angle, and spatial distributions, and spatiotemporal electron and partially-ionized impurity transport. This research provides an initial quantification of the efficacy of RE mitigation via injected impurities, and identification of the critical role played by loss of confinement due to plasma scrape off on the inner wall as compared to the relatively slow collisional dam**. △ Less

Submitted 6 November, 2020; v1 submitted 30 July, 2020; originally announced July 2020.

Comments: Forthcoming in Physics of Plasmas as of 2020/11/06

A sparse-grid probabilistic scheme for approximation of the runaway probability of electrons in fusion tokamak simulation

Authors: Minglei Yang, Guannan Zhang, Diego del-Castillo-Negrete, Miroslav Stoyanov, Matthew Beidler

Abstract: Runaway electrons (RE) generated during magnetic disruptions present a major threat to the safe operation of plasma nuclear fusion reactors. A critical aspect of understanding RE dynamics is to calculate the runaway probability, i.e., the probability that an electron in the phase space will runaway on, or before, a prescribed time. Such probability can be obtained by solving the adjoint equation o… ▽ More Runaway electrons (RE) generated during magnetic disruptions present a major threat to the safe operation of plasma nuclear fusion reactors. A critical aspect of understanding RE dynamics is to calculate the runaway probability, i.e., the probability that an electron in the phase space will runaway on, or before, a prescribed time. Such probability can be obtained by solving the adjoint equation of the underlying Fokker-Planck equation that controls the electron dynamics. In this effort, we present a sparse-grid probabilistic scheme for computing the runaway probability. The key ingredient of our approach is to represent the solution of the adjoint equation as a conditional expectation, such that discretizing the differential operator reduces to the approximation of a set of integrals. Adaptive sparse grid interpolation is utilized to approximate the map from the phase space to the runaway probability. The main novelties of this effort are the integration of the sparse-grid method into the probabilistic numerical scheme for computing escape probability, as well as the demonstration in computing RE probabilities. Two numerical examples are given to illustrate that the proposed method can achieve $\mathcal{O}(Δt)$ convergence, as well as the adaptive refinement strategy can effectively handle the sharp transition layer between the runaway and non-runaway regions. △ Less

Submitted 17 January, 2020; v1 submitted 16 January, 2020; originally announced January 2020.

Fast Magnetic Reconnection Due to Anisotropic Electron Pressure

Authors: P. A. Cassak, R. N. Baylor, R. L. Fermo, M. T. Beidler, M. A. Shay, M. Swisdak, J. F. Drake, H. Karimabadi

Abstract: A new regime of fast magnetic reconnection with an out-of-plane (guide) magnetic field is reported in which the key role is played by an electron pressure anisotropy described by the Chew-Goldberger-Low gyrotropic equations of state in the generalized Ohm's law, which even dominates the Hall term. A description of the physical cause of this behavior is provided and two-dimensional fluid simulation… ▽ More A new regime of fast magnetic reconnection with an out-of-plane (guide) magnetic field is reported in which the key role is played by an electron pressure anisotropy described by the Chew-Goldberger-Low gyrotropic equations of state in the generalized Ohm's law, which even dominates the Hall term. A description of the physical cause of this behavior is provided and two-dimensional fluid simulations are used to confirm the results. The electron pressure anisotropy causes the out-of-plane magnetic field to develop a quadrupole structure of opposite polarity to the Hall magnetic field and gives rise to dispersive waves. In addition to being important for understanding what causes reconnection to be fast, this mechanism should dominate in plasmas with low plasma beta and a high in-plane plasma beta with electron temperature comparable to or larger than ion temperature, so it could be relevant in the solar wind and some tokamaks. △ Less

Submitted 5 February, 2015; originally announced February 2015.

Comments: 5 pages, 2 figures, accepted for publication as a letter in Physics of Plasmas

Model for Incomplete Reconnection in Sawtooth Crashes

Authors: Matthew T. Beidler, Paul A. Cassak

Abstract: A model for incomplete reconnection in sawtooth crashes is presented. The reconnection inflow during the crash phase of sawteeth self-consistently convects the high pressure core toward the reconnection site, raising the pressure gradient there. Reconnection shuts off if the diamagnetic drift speed at the reconnection site exceeds a threshold, which may explain incomplete reconnection. The relaxat… ▽ More A model for incomplete reconnection in sawtooth crashes is presented. The reconnection inflow during the crash phase of sawteeth self-consistently convects the high pressure core toward the reconnection site, raising the pressure gradient there. Reconnection shuts off if the diamagnetic drift speed at the reconnection site exceeds a threshold, which may explain incomplete reconnection. The relaxation of magnetic shear after reconnection stops may explain the destabilization of ideal interchange instabilities reported previously. Proof-of-principle two-fluid simulations confirm this basic picture. Predictions of the model compare favorably to data from the Mega Ampere Spherical Tokamak. Applications to transport modeling of sawteeth are discussed. The results should apply across tokamaks, including ITER. △ Less

Submitted 2 November, 2011; originally announced November 2011.

Comments: 5 pages, 3 figures, accepted for publication in PRL