Giuseppe Bilotta

SPH on GPU with CUDA

A Smoothed Particle Hydrodynamics (SPH) method for free surface flows has been implemented on a graphical processing unit (GPU) using the Compute Unified Device Architecture (CUDA) developed by Nvidia, resulting in tremendous speed-ups. The entire SPH code, with its three main components: neighbor list construction, force computation, and integration of the equation of motion, is computed on the GPU, fully exploiting its computational power. The simulation speed achieved is one to two orders of magnitude faster than the equivalent CPU code. Example applications are shown for paddle-generated waves in a basin and a dam-break wave impact on a structure. GPU implementation of SPH permits high resolution SPH modeling in hours and days rather than weeks and months on inexpensive and readily available hardware.

Lava flow hazards at Mount Etna: constraints imposed by eruptive history and numerical simulations

Improving lava flow hazard assessment is one of the most important and challenging fields of volcanology, and has an immediate and practical impact on society. Here, we present a methodology for the quantitative assessment of lava flow hazards based on a combination of field data, numerical simulations and probability analyses. With the extensive data available on historic eruptions of Mt. Etna, going back over 2000 years, it has been possible to construct two hazard maps, one for flank and the other for summit eruptions, allowing a quantitative analysis of the most likely future courses of lava flows. The effective use of hazard maps of Etna may help in minimizing the damage from volcanic eruptions through correct land use in densely urbanized area with a population of almost one million people. Although this study was conducted on Mt. Etna, the approach used is designed to be applicable to other volcanic areas.

Advances in Multi-GPU Smoothed Particle Hydrodynamics Simulations

We present a multi-GPU version of GPUSPH, a CUDA implementation of fluid-dynamics models based on the smoothed particle hydrodynamics (SPH) numerical method. The SPH is a well-known Lagrangian model for the simulation of free-surface fluid flows; it exposes a high degree of parallelism and has already been successfully ported to GPU. We extend the GPU-based simulator to run simulations on multiple GPUs simultaneously, to obtain a gain in speed and overcome the memory limitations of using a single device. The computational domain is spatially split with minimal overlapping and shared volume slices are updated at every iteration of the simulation. Data transfers are asynchronous with computations, thus completely covering the overhead introduced by slice exchange. A simple yet effective load balancing policy preserves the performance in case of unbalanced simulations due to asymmetric fluid topologies. The obtained speedup factor (up to 4.5x for 6 GPUs) closely follows the expected one (5x for 6 GPUs) and it is possible to run simulations with a higher number of particles than would fit on a single device. We use the Karp-Flatt metric to formally estimate the overall efficiency of the parallelization.

Smoothed Particle Hydrodynamics Simulations on Multi-GPU Systems

We present a multi-GPU version of GPUSPH, a CUDA implementation of fluid-dynamics models based on the Smoothed Particle Hydrodynamics (SPH) numerical method. SPH is a well-known lagrangian model to simulate fluid flows, it exposes a high degree of parallelism and has already been successfully ported to GPU. We extend the GPU-based simulator to run simulations on multiple GPUs simultaneously, to obtain a gain in speed and overcome the memory limitations of using a single device. The computational domain is spatially split with minimal overlapping and shared volume slices are updated at every iteration of the simulation. Data transfers are asynchronous with computations, thus completely covering the overhead introduced by slice exchange. The obtained speedup factor differs from the ideal one by a small cost function linear in the number of devices, and it is possible to run simulations with a higher number of particles than would fit on a single device.

Simulation of Nearshore Tsunami Breaking by Smoothed Particle Hydrodynamics Method

AbstractThis study applies the numerical model GPUSPH, an implementation of the weakly compressible smoothed particle hydrodynamics (SPH) method on graphics processing units, to simulate nearshore tsunami processes. Two sets of laboratory experiments that involve violent wave breaking are simulated by the three-dimensional numerical model. The first set of experiments addresses tsunamilike solitary wave breaking on and overtopping an impermeable seawall. Comparison with free-surface profiles and laboratory images shows that GPUSPH satisfactorily reproduces the complicated wave processes involving wave plunging, collapsing, splash-up, and overtopping. The other set of experiments investigates tsunamilike solitary wave breaking and inundation over shallow water reefs. The performance of GPUSPH is evaluated by comparing its results with (1) experimental data including free-surface measurements and cross-shore velocity profiles, and (2) published numerical results obtained in two mesh-based wave models: the n...

Three-Dimensional SPH Modeling of a Bar/Rip Channel System

AbstractA Lagrangian numerical method called smoothed particle hydrodynamics (SPH) is used to analyze a simplified bar/rip channel system on a beach. Prior studies have shown that SPH models propagate water waves well, including breaking waves; here, it is shown that SPH also models the mean wave-induced nearshore circulation created by breaking waves. The model predictions are compared with the previous laboratory measurements and show good agreement, including the mean velocity profiles, mean surface elevation, and cross-shore velocity components over the rip channel. The alongshore variation of various components of radiation stress and the resulting alongshore force that acts as a feeder for the rip current are obtained from the numerical results.

MAGFLOW: a physics-based model for the dynamics of lava-flow emplacement

Abstract The MAGFLOW model for lava-flow simulations is based on the cellular automaton (CA) approach, and uses a physical model for the thermal and rheological evolution of the flowing lava. We discuss the potential of MAGFLOW to improve our understanding of the dynamics of lava-flow emplacement and our ability to assess lava-flow hazards. Sensitivity analysis of the input parameters controlling the evolution function of the automaton demonstrates that water content and solidus temperatures are the parameters to which MAGFLOW is most sensitive. Additional tests also indicate that temporal changes in effusion rate strongly influence the accuracy of the predictive modelling of lava-flow paths. The parallel implementation of MAGFLOW on graphic processing units (GPUs) can achieve speed-ups of two orders of magnitude relative to the corresponding serial implementation, providing a lava-flow simulation spanning several days of eruption in just a few minutes. We describe and demonstrate the operation of MAGFLOW using two case studies from Mt Etna: one is a reconstruction of the detailed chronology of the lava-flow emplacement during the 2006 flank eruption; and the other is the production of the lava-flow hazard map of the persistent eruptive activity at the summit craters.

GPUSPH: a Smoothed Particle Hydrodynamics model for the thermal and rheological evolution of lava flows

Abstract GPUSPH is a fully three-dimensional model for the simulation of the thermal and rheological evolution of lava flows that relies on the Smoothed Particle Hydrodynamics (SPH) numerical method. Thanks to the Lagrangian, meshless nature of SPH, the model incorporates a more complete physical description of the emplacement process and rheology of lava that considers the free surface, the irregular boundaries represented by the topography, the solidification fronts and the non-Newtonian rheology with temperature-dependent parameters. GPUSPH follows the very general Herschel–Bulkley rheological model, which encompasses Newtonian, power-law and Bingham flow behaviours, with both constant and temperature-dependent parameters, and can thus be used to explore in detail the impact of rheology on the behaviour of lava flows and on their emplacement. To illustrate this possibility, we present some preliminary applications of the model for studying the rheology of lava flows with different constitutive relationships and thermal regimes using the real topography of the Mt Etna volcano.

Optimizing Satellite Monitoring of Volcanic Areas Through GPUs and Multi-Core CPUs Image Processing: An OpenCL Case Study

Satellite image processing algorithms often offer a very high degree of parallelism (e.g., pixel-by-pixel processing) that make them optimal candidates for execution on high-performance parallel computing hardware such as modern graphic processing units (GPUs) and multicore CPUs with vector processing capabilities. By using the OpenCL computing standard, a single implementation of a parallel algorithm can be deployed on a wide range of hardware platforms. However, achieving the best performance on each individual platform may still require a custom implementation. We show some possible approaches to the optimization of satellite image processing algorithms on a range of different platforms, discussing the implementation in OpenCL of the classic Brightness Temperature Difference ash-cloud detection algorithm.

HOTSAT: a multiplatform system for the thermal monitoring of volcanic activity using satellite data

Abstract The HOTSAT multiplatform system for the analysis of infrared data from satellites provides a framework that allows the detection of volcanic hotspots and an output of their associated radiative power. This multiplatform system can operate on both Moderate Resolution Imaging Spectroradiometer and Spinning Enhanced Visible and Infrared Imager data. The new version of the system is now implemented on graphics processing units and its interface is available on the internet under restricted access conditions. Combining the estimation of time-varying discharge rates using HOTSAT with the MAGFLOW physics-based model to simulate lava flow paths resulted in the first operational system in which satellite observations drive the modelling of lava flow emplacement. This allows the timely definition of the parameters and maps essential for hazard assessment, including the propagation time of lava flows and the maximum run-out distance. The system was first used in an operational context during the paroxysmal episode at Mt Etna on 12–13 January 2011, when we produced real-time predictions of the areas likely to be inundated by lava flows while the eruption was still ongoing. This allowed key at-risk areas to be rapidly and appropriately identified.