Hybrid OpenMP/MPI parallelization of the charge deposition step in the global gyrokinetic Particle-In-Cell code ORB5

Gyrokinetic simulations are computationally extremely demanding due to the high dimensionality of the physical phase space and the interplay between plasma particles and electromagnetic fields. It is thus essential to make full use of the available numerical resources to be able to simulate more complex physical problems. With the aim of optimizing the gyrokinetic Particle-In-Cell code ORB5 towards exascale computing, a particle sorting method is implemented to increase data locality. Furthermore, different algorithms are used to improve vectorization, and the MPI parallelization is complemented with OpenMP. More specifically, we shall focus on the particle to grid operations involved in the PIC charge deposition step. The latter is critical to parallelize using a shared memory paradigm due to the scatter operations involved. We will present the different algorithms and parallelization schemes implemented in the ORB5 charge deposition step and how they affect the speedup compared to the base MPI case.

Optical Mirror from Laser-Trapped Mesoscopic Particles

Jean-Marc Fournier, Johann Rohner

Trapping of mesoscopic particles by optical forces usually relies on the gradient force, whereby particles are attracted into optical wells formed by landscaping the intensity of an optical field. This is most often achieved by optical Gaussian beams, interference patterns, general phase contrast methods, or other mechanisms. Hence, although the simultaneous trapping of several hundreds of particles can be achieved, these particles remain mostly independent with negligible interaction. Optical matter, however, relies on close packing and binding forces, with fundamentally different electrodynamic properties. In this Letter, we build ensembles of optically bound particles to realize a reflective surface that can be used to image an object or to focus a light beam. To our knowledge, this is the first experimental proof of the creation of a mirror by optical matter, and represents an important step toward the realization of a laser-trapped mirror (LTM) in space. From a theoretical point of view, optically bound close packing requires an exact solver of Maxwell's equations in order to precisely compute the field scattered by the collection of particles. Such rigorous calculations have been developed and are used here to study the focusing and resolving power of an LTM.

Amer Physical Soc2014

A Fast Ion Loss Detector for the TCV Tokamak

Lorenzo Stipani

A Fast Ion Loss Detector (FILD) was designed, assembled, installed and commissioned for TCV. This is a radially positionable, scintillator based detector that provides information on the 2D fast-ion velocity space lost at the probe's location. The collected particles are collimated inside the probe head and impinge upon a plate coated with a scintillator material that emits light. The photon flux is relayed to two acquisition systems: a sCMOS camera for high spatial resolution measurements of the emission locations in the scintillator, and fast photo-multipliers that allow time-correlation studies of fast-ion losses with high-frequency electromagnetic fluctuations. From its position with respect to the confined plasma and the fast-ion sources on TCV, FILD probes a small portion of the lost fast-ion phase space with high velocity-space and temporal resolution. Therefore, it naturally complements other available diagnostics such as neutral particle analysers, spectroscopy techniques for light emission from energetic particles following CX reactions, or neutron counters, which feature a broader, but less resolved, spatial coverage of fast-ion dynamics.The TCV-FILD design introduces some novelties for exploring new ranges of operation that may be adopted in similar systems for other Tokamaks, such as ITER. Two entrance slits can collect particles that circulate in co- and cntr-plasma current directions. This will be particularly useful when the second NBH system, injecting in the opposite toroidal direction, with particle energies that can excite strong Alfvénic modes, will be operated on TCV. A controlled pneumatic linear actuator radially positions the detector to expose the slits to the particle flux up to 9mm inward of the vessel wall. A plug-in design was conceived to facilitate diagnostic installation. The diagnostic was installed and commissioned during TCV experiments in 2020. The sensitivity of the detector to the local magnetic field line direction was investigated. The direction of the plasma current was found to select which of the two slits may be traversed by lost particles. The operational limits in discharges with NBH with total delivered energies up to 1MJ were assessed with the help of sensors monitoring the temperature of the probe head and cameras detecting visible light emissions resulting from the graphite shield heating by particle fluxes. Using FILD, fast-ion losses were detected for the first time on TCV in plasma discharges exhibiting strong MHD modes. Ejections of energetic ions were found correlated in time with Sawtooth crashes, as a sequence of individual events, and in strong phase coherence, as a continuous loss in time, with magnetic perturbations of a saturated and toroidally rotating magnetic island of a NTM. These observations, in addition to providing initial results on the relevant physics phenomena, stimulating further experiments and comparisons with theory, demonstrate the ability of TCV-FILD to provide valuable information in plasma discharges of interest for fast-ion studies. These measurements and additional information from other TCV diagnostics may now be combined to reconstruct, with tomographic inversion techniques, more of the fast-ion phase space. This will be used to identify the conditions for the excitation/suppression of magnetic instabilities, develop methods for their real-time control with heating and/or shaping actuators and investigate their dependencies on plasma parameters.

EPFL2021

Schwinger production of scalar particles during and after inflation from the first principles

Oleksandr Sobol

By using the first-principles approach, we derive a system of three quantum kinetic equations governing the production and evolution of charged scalar particles by an electric field in an expanding universe. Analyzing the ultraviolet asymptotic behavior of the kinetic functions, we found the divergent parts of the electric current and the energy-momentum tensor of the produced particles and determined the corresponding counterterms. The renormalized system of equations is used to study the generation of electromagnetic fields during and after inflation in the kinetic coupling model L-EM = -(1/4)f(2)(phi)F-mu nu F-mu nu with the Ratra coupling function f = exp(beta phi/M-p). It is found that the electric current of created particles is retarded with respect to the electric field. This leads to an oscillatory behavior of both quantities in agreement with the results obtained previously in phenomenological kinetic and hydrodynamical approaches.

2020

A Fast Ion Loss Detector for the TCV Tokamak

Lorenzo Stipani

EPFL2021