Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userAn Efficient Particle Tracking Algorithm for Large-Scale Parallel Pseudo-Spectral Simulations of Turbulence
139
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
Particle tracking in large-scale numerical simulations of turbulent flows presents one of the major bottlenecks in parallel performance and scaling efficiency. Here, we describe a particle tracking algorithm for large-scale parallel pseudo-spectral simulations of turbulence which scales well up to billions of tracer particles on modern high-performance computing architectures. We summarize the standard parallel methods used to solve the fluid equations in our hybrid MPI/OpenMP implementation. As the main focus, we describe the implementation of the particle tracking algorithm and document its computational performance. To address the extensive inter-process communication required by particle tracking, we introduce a task-based approach to overlap point-to-point communications with computations, thereby enabling improved resource utilization. We characterize the computational cost as a function of the number of particles tracked and compare it with the flow field computation, showing that the cost of particle tracking is very small for typical applications.
rate research
Read More
We propose a one-way coupled model that tracks individual primary particles in a conceptually simple cellular flow setup to predict flocculation in turbulence. This computationally efficient model accounts for Stokes drag, lubrication, cohesive and direct contact forces on the primary spherical particles and allows for a systematic simulation campaign that yields the transient mean floc size as a function of the governing dimensionless parameters. The simulations reproduce the growth of the cohesive flocs with time, and the emergence of a log-normal equilibrium distribution governed by the balance of aggregation and breakage. Flocculation proceeds most rapidly when the Stokes number of the primary particles is textit{O}(1). Results from this simple computational model are consistent with experimental observations, thus allowing us to propose a new analytical flocculation model that yields improved agreement with experimental data, especially during the transient stages.
Clustering of inertial particles is important for many types of astrophysical and geophysical turbulence, but it has been studied predominately for incompressible flows. Here we study compressible flows and compare clustering in both compressively (irrotationally) and vortically (solenoidally) forced turbulence. Vortically and compressively forced flows are driven stochastically either by solenoidal waves or by circular expansion waves, respectively. For compressively forced flows, the power spectra of the density of inertial particles are a particularly sensitive tool for displaying particle clustering relative to the density enhancement. We use both Lagrangian and Eulerian descriptions for the particles. Particle clustering through shock interaction is found to be particularly prominent in turbulence driven by spherical expansion waves. It manifests itself through a double-peaked distribution of spectral power as a function of Stokes number. The two peaks are associated with two distinct clustering mechanisms; shock interaction for smaller Stokes numbers and the centrifugal sling effect for larger values. The clustering of inertial particles is associated with the formation of caustics. Such caustics can only be captured in the Lagrangian description, which allows us to assess the relative importance of caustics in vortically and irrotationally forced turbulence. We show that the statistical noise resulting from the limited number of particles in the Lagrangian description can be removed from the particle power spectra, allowing us a more detailed comparison of the residual spectra. We focus on the Epstein drag law relevant for rarefied gases, but show that our findings apply also to the usual Stokes drag.
We examine the accuracy of spatial derivatives computed from numerical simulations of supersonic turbulence. Two sets of simulations, carried out using a finite-volume code that evolves the hydrodynamic equations with an approximate Riemann solver and a finite-difference code that solves the Navier-Stokes equations, are tested against a number of criteria based on the continuity equation, including exact results at statistically steady state. We find that the spatial derivatives in the Navier-Stokes runs are accurate and satisfy all the criteria. In particular, they satisfy our exact results that the conditional mean velocity divergence, $langle abla cdot {bs u}|srangle$, where $s$ is the logarithm of density, and the conditional mean of the advection of $s$, $langle {bs u} cdot abla s|srangle$, vanish at steady state for all density values, $s$. On the other hand, the Riemann solver simulations fail all the tests that require accurate evaluation of spatial derivatives, resulting in apparent violation of the continuity equation, even if the solver enforces mass conservation. In particular, analysis of the Riemann simulations may lead to the incorrect conclusion that the $pdv$ work tends to preferentially convert kinetic energy into thermal energy, inconsistent with the exact result that the energy exchange by $pdv$ work is symmetric in barotropic supersonic turbulence at steady state. The inaccuracy of spatial derivatives is a general problem in the post-processing of simulations of supersonic turbulence with Riemann solvers. Solutions from such simulations must be used with caution in post-processing studies concerning the spatial gradients.
Magnetohydrodynamical (MHD) dynamos emerge in many different astrophysical situations where turbulence is present, but the interaction between large-scale (LSD) and small-scale dynamos (SSD) is not fully understood. We performed a systematic study of turbulent dynamos driven by isotropic forcing in isothermal MHD with magnetic Prandtl number of unity, focusing on the exponential growth stage. Both helical and non-helical forcing was employed to separate the effects of LSD and SSD in a periodic domain. Reynolds numbers (Rm) up to $approx 250$ were examined and multiple resolutions used for convergence checks. We ran our simulations with the Astaroth code, designed to accelerate 3D stencil computations on graphics processing units (GPUs) and to employ multiple GPUs with peer-to-peer communication. We observed a speedup of $approx 35$ in single-node performance compared to the widely used multi-CPU MHD solver Pencil Code. We estimated the growth rates both from the averaged magnetic fields and their power spectra. At low Rm, LSD growth dominates, but at high Rm SSD appears to dominate in both helically and non-helically forced cases. Pure SSD growth rates follow a logarithmic scaling as a function of Rm. Probability density functions of the magnetic field from the growth stage exhibit SSD behaviour in helically forced cases even at intermediate Rm. We estimated mean-field turbulence transport coefficients using closures like the second-order correlation approximation (SOCA). They yield growth rates similar to the directly measured ones and provide evidence of $alpha$ quenching. Our results are consistent with the SSD inhibiting the growth of the LSD at moderate Rm, while the dynamo growth is enhanced at higher Rm.
Population annealing Monte Carlo is an efficient sequential algorithm for simulating k-local Boolean Hamiltonians. Because of its structure, the algorithm is inherently parallel and therefore well suited for large-scale simulations of computationally hard problems. Here we present various ways of optimizing population annealing Monte Carlo using 2-local spin-glass Hamiltonians as a case study. We demonstrate how the algorithm can be optimized from an implementation, algorithmic accelerator, as well as scalable parallelization points of view. This makes population annealing Monte Carlo perfectly suited to study other frustrated problems such as pyrochlore lattices, constraint-satisfaction problems, as well as higher-order Hamiltonians commonly found in, e.g., topological color codes.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
