GX is a code for solving the nonlinear gyrokinetic system for low-frequency turbulence in magnetized plasmas, particularly tokamaks and stellarators. In GX, our primary motivation and target is a fast gyrokinetic solver that can be used for fusion reactor design and optimization along with wide-ranging physics exploration. This has led to several code and algorithm design decisions, specifically chosen to prioritize time to solution. First, we have used a discretization algorithm that is pseudo-spectral in the entire phase-space, including a Laguerre-Hermite pseudo-spectral formulation of velocity space, which allows for smooth interpolation between coarse gyrofluid-like resolutions and finer conventional gyrokinetic resolutions and efficient evaluation of a model collision operator. Additionally, we have built GX to natively target graphics processors (GPUs), which are among the fastest computational platforms available today. Finally, we have taken advantage of the reactor-relevant limit of small $\rho_*$ by using the radially-local flux-tube approach. In this paper we present details about the gyrokinetic system and the numerical algorithms used in GX to solve the system. We then present several numerical benchmarks against established gyrokinetic codes in both tokamak and stellarator magnetic geometries to verify that GX correctly simulates gyrokinetic turbulence in the small $\rho_*$ limit. Moreover, we show that the convergence properties of the Laguerre-Hermite spectral velocity formulation are quite favorable for nonlinear problems of interest. Coupled with GPU acceleration, which we also investigate with scaling studies, this enables GX to be able to produce useful turbulence simulations in minutes on one (or a few) GPUs. GX is open-source software that is ready for fusion reactor design studies.

One of the most promising applications of machine learning (ML) in computational physics is to accelerate the solution of partial differential equations (PDEs). The key objective of ML-based PDE solvers is to output a sufficiently accurate solution faster than standard numerical methods, which are used as a baseline comparison. We first perform a systematic review of the ML-for-PDE solving literature. Of articles that use ML to solve a fluid-related PDE and claim to outperform a standard numerical method, we determine that 79% (60/76) compare to a weak baseline. Second, we find evidence that reporting biases, especially outcome reporting bias and publication bias, are widespread. We conclude that ML-for-PDE solving research is overoptimistic: weak baselines lead to overly positive results, while reporting biases lead to underreporting of negative results. To a large extent, these issues appear to be caused by factors similar to those of past reproducibility crises: researcher degrees of freedom and a bias towards positive results. We call for bottom-up cultural changes to minimize biased reporting as well as top-down structural reforms intended to reduce perverse incentives for doing so.

The EPED model [P.B. Snyder et al 2011 Nucl. Fusion 51 103016] had success in describing type-I ELM and QH-mode pedestals in conventional tokamaks, by combining kinetic ballooning mode (KBM) and peeling-ballooning (PB) constraints. Within EPED, the KBM constraint is usually approximated by the ideal ballooning mode (IBM) stability threshold. It has been noted that quantitative differences between local ideal MHD and gyro-kinetic (GK) ballooning stability can be larger at low aspect ratio. KBM critical pedestals are consistent with observation in initial studies on conventional and spherical tokamaks. In this work, the application of a reduced model for the calculation of the kinetic ballooning stability boundary is presented based on a novel and newly developed Gyro-Fluid System (GFS) code [G.M. Staebler et al 2023 Phys. Plasmas 30 102501]. GFS is observed to capture KBMs in DIII-D as well as the NSTX(-U) pedestals, opening the route for the integration of this model into EPED. Finally, high-n global ballooning modes are observed to limit local 2nd stability access and thus provide a transport mechanism that constrains the width evolution with beta_p,ped. The high-n global ballooning stability is approximated by its ideal MHD analogue using ELITE. It is shown that nearly local high-n with k_y*rho_s~1/2 modes can provide a proxy for the critical beta_p,ped when 2nd stable access exists on DIII-D plasmas. The use of GFS and ELITE scaling in EPED provided improved agreement in comparison to EPED1 with DIII-D pedestal data.

We study the propagation of cosmic rays (CRs) through a simulation of magnetohydrodynamic (MHD) turbulence at unprecedented resolution of $10{,}240^3$. We drive turbulence that is subsonic and super-Alfvénic, characterized by $\delta B_{\rm rms}/B_0=2$. The high resolution enables an extended inertial range such that the Alfvén scale $l_A$, where $\delta B (l_A)\approx B_0$, is well resolved. This allows us to properly capture how the cascade transitions from large amplitudes on large scales to small amplitudes on small scales. We find that sharp bends in the magnetic field are key mediators of particle transport even on small scales via resonant curvature scattering. We further find that particle scattering in the turbulence shows strong hints of self-similarity: (1) the diffusion has weak energy dependence over almost two decades in particle energy and (2) the particles' random walk exhibits a broad power-law distribution of collision times such that the diffusion is dominated by the rarest, long-distance excursions. Our results suggest that large-amplitude MHD turbulence can provide efficient scattering over a wide range of CR energies and may help explain many CR observations above a $\sim$TeV: the flattening of the B/C spectrum, the hardening of CR primary spectra and the weak dependence of arrival anisotropy on CR energy.

Deep learning algorithms provide a new paradigm to study high-dimensional dynamical behaviors, such as those in fusion plasma systems. Development of novel model reduction methods, coupled with detection of abnormal modes with plasma physics, opens a unique opportunity for building efficient models to identify plasma instabilities for real-time control. Our Fusion Transfer Learning (FTL) model demonstrates success in reconstructing nonlinear kink mode structures by learning from a limited amount of nonlinear simulation data. The knowledge transfer process leverages a pre-trained neural encoder-decoder network, initially trained on linear simulations, to effectively capture nonlinear dynamics. The low-dimensional embeddings extract the coherent structures of interest, while preserving the inherent dynamics of the complex system. Experimental results highlight FTL's capacity to capture transitional behaviors and dynamical features in plasma dynamics -- a task often challenging for conventional methods. The model developed in this study is generalizable and can be extended broadly through transfer learning to address various magnetohydrodynamics (MHD) modes.

Magnetohydrodynamic turbulence regulates the transfer of energy from large to small scales in many astrophysical systems, including the solar atmosphere. We perform three-dimensional magnetohydrodynamic simulations with unprecedentedly large magnetic Reynolds number to reveal how rapid reconnection of magnetic field lines changes the classical paradigm of the turbulent energy cascade. By breaking elongated current sheets into chains of small magnetic flux ropes (or plasmoids), magnetic reconnection leads to a new range of turbulent energy cascade, where the rate of energy transfer is controlled by the growth rate of the plasmoids. As a consequence, the turbulent energy spectra steepen and attain a spectral index of -2.2 that is accompanied by changes in the anisotropy of turbulence eddies. The omnipresence of plasmoids and their consequences on, e.g., solar coronal heating, can be further explored with current and future spacecraft and telescopes.

Designing magnets for three-dimensional plasma confinement is a key task for advancing the stellarator as a fusion reactor concept. Stellarator magnets must produce an accurate field while leaving adequate room for other components and being reasonably simple to construct and assemble. In this paper, a framework for coil design and optimization is introduced that enables the attainment of sparse magnet solutions with arbitrary restrictions on where coils may be located. The solution space is formulated as a "wireframe" consisting of a mesh of interconnected wire segments enclosing the plasma. Two methods are developed for optimizing the current distribution on a wireframe: Regularized Constrained Least Squares (RCLS), which uses a linear least-squares approach to optimize the currents in each segment, and Greedy Stellarator Coil Optimization (GSCO), a fully discrete procedure in which loops of current are added to the mesh one by one to achieve the desired magnetic field on the plasma boundary. Examples are presented of solutions obtainable with each method, some of which achieve high field accuracy while obeying spatial constraints that permit easy assembly.

Magnetic reconnection is a fundamental and omnipresent energy conversion process in plasma physics. Novel observations of fields and particles from Parker Solar Probe (PSP) have shown the absence of reconnection in a large number of current sheets in the near-Sun solar wind. Using near-Sun observations from PSP Encounters 4 to 11 (Jan 2020 to March 2022), we investigate whether reconnection onset might be suppressed by velocity shear. We compare estimates of the tearing mode growth rate in the presence of shear flow for time periods identified as containing reconnecting current sheets versus non-reconnecting times, finding systematically larger growth rates for reconnection periods. Upon examination of the parameters associated with reconnection onset, we find that 85% of the reconnection events are embedded in slow, non-Alfvenic wind streams. We compare with fast, slow non-Alfvenic, and slow Alfvenic streams, finding that the growth rate is suppressed in highly Alfvenic fast and slow wind and reconnection is not seen in these wind types, as would be expected from our theoretical expressions. These wind streams have strong Alfvenic flow shear, consistent with the idea of reconnection suppression by such flows. This could help explain the frequent absence of reconnection events in the highly Alfvenic, near-Sun solar wind observed by PSP. Finally, we find a steepening of both the trace and magnitude magnetic field spectra within reconnection periods in comparison to ambient wind. We tie this to the dynamics of relatively balanced turbulence within these reconnection periods and the potential generation of compressible fluctuations.

We present first-principles numerical calculations of the depolarization rate of spin-polarized deuterium and tritium nuclei in realistic tokamak plasmas, driven by resonant interactions with plasma waves. Backed up by first-of-a-kind linear and nonlinear simulations, we find that alpha particle-driven Alfvénic modes cause only negligible depolarization, which is contrary to expectations in prior literature. Other Alfvénic instabilities can in principle degrade polarization, but only under conditions unlikely to be realized on transport timescales. By combining full-orbit particle tracing with a dedicated depolarization solver, we demonstrate that wave-driven depolarization is surprisingly weak in SPARC and ITER-scale devices. These results provide strong evidence that spin-polarized fuel can maintain its polarization long enough to boost fusion reactivity, opening a viable path toward substantially enhanced performance in magnetic confinement fusion power plants.

Quantum computing is gaining increased attention as a potential way to speed up simulations of physical systems, and it is also of interest to apply it to simulations of classical plasmas. However, quantum information science is traditionally aimed at modeling linear Hamiltonian systems of a particular form that is found in quantum mechanics, so extending the existing results to plasma applications remains a challenge. Here, we report a preliminary exploration of the long-term opportunities and likely obstacles in this area. First, we show that many plasma-wave problems are naturally representable in a quantumlike form and thus are naturally fit for quantum computers. Second, we consider more general plasma problems that include non-Hermitian dynamics (instabilities, irreversible dissipation) and nonlinearities. We show that by extending the configuration space, such systems can also be represented in a quantumlike form and thus can be simulated with quantum computers too, albeit that requires more computational resources compared to the first case. Third, we outline potential applications of hybrid quantum--classical computers, which include analysis of global eigenmodes and also an alternative approach to nonlinear simulations.

Machine learning algorithms often struggle to control complex real-world systems. In the case of nuclear fusion, these challenges are exacerbated, as the dynamics are notoriously complex, data is poor, hardware is subject to failures, and experiments often affect dynamics beyond the experiment's duration. Existing tools like reinforcement learning, supervised learning, and Bayesian optimization address some of these challenges but fail to provide a comprehensive solution. To overcome these limitations, we present a multi-scale Bayesian optimization approach that integrates a high-frequency data-driven dynamics model with a low-frequency Gaussian process. By updating the Gaussian process between experiments, the method rapidly adapts to new data, refining the predictions of the less reliable dynamical model. We validate our approach by controlling tearing instabilities in the DIII-D nuclear fusion plant. Offline testing on historical data shows that our method significantly outperforms several baselines. Results on live experiments on the DIII-D tokamak, conducted under high-performance plasma scenarios prone to instabilities, shows a 50% success rate, marking a 117% improvement over historical outcomes.

We present laboratory results from supercritical, magnetized collisionless shock experiments ($M_A \lesssim 10$, $\beta\sim 1$). We report the first observation of fully-developed shocks ($R=4$ compression ratio and a downstream region decoupled from the piston) after seven upstream ion gyration periods. A foot ahead of the shock exhibits super-adiabatic electron and ion heating. We measure the electron temperature $T_e = 115$ eV and ion temperature $T_i = 15$ eV upstream of the shock; whereas, downstream, we measure $T_e=390$ eV and infer $T_i=340$ eV, consistent with both Thomson scattering ion-acoustic wave spectral broadening and Rankine-Hugoniot conditions. The downstream electron temperature has a $30$-percent excess from adiabatic and collisional electron-ion heating, implying significant collisionless anomalous electron heating. Furthermore, downstream electrons and ions are in equipartition, with a unity electron-ion temperature ratio $T_e/T_i = 1.2$.

Extended periods of radio pulsations have been observed for six magnetars, displaying characteristics different from those of ordinary pulsars. In this Letter, we argue that radio emission is generated in a closed, twisted magnetic flux bundle originating near the magnetic pole and extending beyond 100 km from the magnetar. The electron-positron flow in the twisted bundle has to carry electric current and, at the same time, experiences a strong drag by the radiation field of the magnetar. This combination forces the plasma into a ``radiatively locked'' state with a sustained two-stream instability, generating radio emission. We demonstrate this mechanism using novel first-principles simulations that follow the plasma behavior by solving the relativistic Vlasov equation with the discontinuous Galerkin method. First, using one-dimensional simulations, we demonstrate how radiative drag induces the two-stream instability, sustaining turbulent electric fields. When extended to two dimensions, the system produces electromagnetic waves, including superluminal modes capable of escaping the magnetosphere. We measure their frequency and emitted power, and incorporate the local simulation results into a global magnetospheric model. The model explains key features of observed radio emission from magnetars: its appearance after an X-ray outburst, wide pulse profiles, luminosities $\sim 10^{30}{\rm{erg/s}}$, and a broad range of frequencies extending up to $\sim 100\, \mathrm{GHz}$.

In addressing the Department of Energy's April, 2022 announcement of a Bold Decadal Vision for delivering a Fusion Pilot Plant by 2035, associated software tools need to be developed for the integration of real world engineering and supply chain data with advanced science models that are accelerated with Machine Learning. An associated research and development effort has been introduced here with promising early progress on the delivery of a realistic Digital Twin Tokamak that has benefited from accelerated advances by the Princeton University AI Deep Learning innovative near-real-time simulators accompanied by technological capabilities from the NVIDIA Omniverse, an open computing platform for building and operating applications that connect with leading scientific computing visualization software. Working with the CAD files for the GA/DIII-D tokamak including equilibrium evolution as an exemplar tokamak application using Omniverse, the Princeton-NVIDIA collaboration has integrated modern AI/HPC-enabled near-real-time kinetic dynamics to connect and accelerate state-of-the-art, synthetic, HPC simulators to model fusion devices and control systems. The overarching goal is to deliver an interactive scientific digital twin of an advanced MFE tokamak that enables near-real-time simulation workflows built with Omniverse to eventually help open doors to new capabilities for generating clean power for a better future.

We consider the novel problem of optimizing a large set of passive superconducting coils (PSCs) with currents induced by a background magnetic field rather than power supplies. In the nuclear fusion literature, such coils have been proposed to partially produce the 3D magnetic fields for stellarators and provide passive stabilization. We perform the first optimizations of PSC arrays with respect to the orientation, shape, and location of each coil, jointly minimized with the background fields. We conclude by generating passive coil array solutions for four stellarators.

Entity is a new-generation, fully open-source particle-in-cell (PIC) code developed to overcome key limitations in astrophysical plasma modeling, particularly the extreme separation of scales and the performance challenges associated with evolving, GPU-centric computing infrastructures. It achieves hardware-agnostic performance portability across various GPU and CPU architectures using the Kokkos library. Crucially, Entity maintains a high standard for usability, clarity, and customizability, offering a robust and easy-to-use framework for developing new algorithms and grid geometries, which allows extensive control without requiring edits to the core source code. This paper details the core general-coordinate special-relativistic module. Entity is the first PIC code designed to solve the Vlasov-Maxwell system in general coordinates, enabling a coordinate-agnostic framework that provides the foundational structure for straightforward extension to arbitrary coordinate geometries. The core methodology achieves numerical stability by solving particle equations of motion in the global orthonormal Cartesian basis, despite using generalized coordinates like Cartesian, axisymmetric spherical, and quasi-spherical grids. Charge conservation is ensured via a specialized current deposition technique using conformal currents. The code exhibits robust scalability and performance portability on major GPU platforms (AMD MI250X, NVIDIA A100, and Intel Max Series), with the 3D particle pusher and the current deposition operating efficiently at about 2 nanoseconds per particle per timestep. Functionality is validated through a comprehensive suite of standard Cartesian plasma tests and the accurate modeling of relativistic magnetospheres in curvilinear axisymmetric geometries.

Computer aided engineering of multi-time-scale plasma systems which exhibit a quasi-steady state solution are challenging due to the large number of time steps required to reach convergence. Machine learning techniques combined with traditional first-principles simulations and high-performance computing offer many interesting pathways towards resolving this challenge. We consider acceleration of kinetic plasma simulations via machine learning generated initial conditions. The approach is demonstrated through modeling of capacitively coupled plasma discharges relevant to the microelectronics industry. Three models are trained on simulations across a parameter space of device driving frequency and operating pressure. The models incorporate elements of a multi-layer perceptron, principal component analysis, and convolutional neural networks to predict the final time-averaged profiles of ion-density and velocity distribution functions. These data-driven initial condition generators (ICGs) provide a mean speedup of 17.1x in convergence time, when measured using an offline procedure, or a 4.4x speedup with an online procedure, with convolutional neural networks leading to the best performance. The paper also outlines a workflow for continuous data-driven model improvement and simulation speedup, with the aim of generating sufficient data for full device digital twins.

The Vera C. Rubin Observatory recently released Data Preview 1 (DP1) in advance of the upcoming Legacy Survey of Space and Time (LSST), which will enable boundless discoveries in time-domain astronomy over the next ten years. DP1 provides an ideal sandbox for validating innovative data analysis approaches for the LSST mission, whose scale challenges established software infrastructure paradigms. This note presents a pair of such pipelines for variability-finding using powerful software infrastructure suited to LSST data, namely the HATS (Hierarchical Adaptive Tiling Scheme) format and the LSDB framework, developed by the LSST Interdisciplinary Network for Collaboration and Computing (LINCC) Frameworks team. This article presents a pair of variability-finding pipelines built on LSDB, the HATS catalog of DP1 data, and preliminary results of detected variable objects, two of which are novel discoveries.

An improved understanding of Field Reversed Configuration (FRC) merging and stability in high acceleration and compression magnetic fields is needed to speed up the development of the pulsed fusion concept developed at Helion Energy. All previous theoretical and simulation work on FRC merging and compression was performed using 2D MHD models. The results of novel 2D hybrid simulations (fluid electrons and full-orbit kinetic ions) of FRC merging and compression are presented. Results of kinetic and MHD simulations, computed using the HYM code, are compared and analyzed. In cases without axial magnetic compression, both the MHD and hybrid simulations show a high sensitivity to the initial parameters (i.e. FRC separation, velocity, normalized separatrix radius, and plasma viscosity), showing that FRCs with large elongation and separatrix radius either do not merge or merge partially, forming a doublet FRC. Application of a mirror coil field at the FRC ends with increasing strength is shown to lead to fast and complete merging of the FRCs in MHD and kinetic simulations.