Hide Sakaguchi

Parallel-vector algorithms for particle simulations on shared-memory multiprocessors

Over the last few decades, the computational demands of massive particle-based simulations for both scientific and industrial purposes have been continuously increasing. Hence, considerable efforts are being made to develop parallel computing techniques on various platforms. In such simulations, particles freely move within a given space, and so on a distributed-memory system, load balancing, i.e., assigning an equal number of particles to each processor, is not guaranteed. However, shared-memory systems achieve better load balancing for particle models, but suffer from the intrinsic drawback of memory access competition, particularly during (1) paring of contact candidates from among neighboring particles and (2) force summation for each particle. Here, novel algorithms are proposed to overcome these two problems. For the first problem, the key is a pre-conditioning process during which particle labels are sorted by a cell label in the domain to which the particles belong. Then, a list of contact candidates is constructed by pairing the sorted particle labels. For the latter problem, a table comprising the list indexes of the contact candidate pairs is created and used to sum the contact forces acting on each particle for all contacts according to Newtons third law. With just these methods, memory access competition is avoided without additional redundant procedures. The parallel efficiency and compatibility of these two algorithms were evaluated in discrete element method (DEM) simulations on four types of shared-memory parallel computers: a multicore multiprocessor computer, scalar supercomputer, vector supercomputer, and graphics processing unit. The computational efficiency of a DEM code was found to be drastically improved with our algorithms on all but the scalar supercomputer. Thus, the developed parallel algorithms are useful on shared-memory parallel computers with sufficient memory bandwidth.

Proton pumping accompanies calcification in foraminifera

Ongoing ocean acidification is widely reported to reduce the ability of calcifying marine organisms to produce their shells and skeletons. Whereas increased dissolution due to acidification is a largely inorganic process, strong organismal control over biomineralization influences calcification and hence complicates predicting the response of marine calcifyers. Here we show that calcification is driven by rapid transformation of bicarbonate into carbonate inside the cytoplasm, achieved by active outward proton pumping. Moreover, this proton flux is maintained over a wide range of pCO2 levels. We furthermore show that a V-type H+ ATPase is responsible for the proton flux and thereby calcification. External transformation of bicarbonate into CO2 due to the proton pumping implies that biomineralization does not rely on availability of carbonate ions, but total dissolved CO2 may not reduce calcification, thereby potentially maintaining the current global marine carbonate production.

Gutenberg‐Richter's law in sliding friction of gels

We report on experimental studies of spatio-temporally heterogeneous stick-slip motions in the sliding friction between a hard polymethyl methacrylate (PMMA, plexiglass) block and a soft poly-dimethyl siloxane (PDMS, silicone) gel plate. We perform experiments on two PDMS gels with different viscoelastic properties. For the less viscous gel, large and rapid events are preceded by an alternation of active and less active periods. For the more viscous gel, successive slow slip events take place continuously. The probability distributions of the force drop, a quantity analogous to seismic moment, obey a power law similar to Gutenberg-Richters empirical law for the frequency-size statistics of earthquakes, and the exponents of the power law vary with the plate velocity and the viscosity of the gel. We propose a simple model to explain the dependence of the power law exponent on the plate velocity, which agrees with experimental results.

Computational performance of a smoothed particle hydrodynamics simulation for shared-memory parallel computing

Abstract The computational performance of a smoothed particle hydrodynamics (SPH) simulation is investigated for three types of current shared-memory parallel computer devices: many integrated core (MIC) processors, graphics processing units (GPUs), and multi-core CPUs. We are especially interested in efficient shared-memory allocation methods for each chipset, because the efficient data access patterns differ between compute unified device architecture (CUDA) programming for GPUs and OpenMP programming for MIC processors and multi-core CPUs. We first introduce several parallel implementation techniques for the SPH code, and then examine these on our target computer architectures to determine the most effective algorithms for each processor unit. In addition, we evaluate the effective computing performance and power efficiency of the SPH simulation on each architecture, as these are critical metrics for overall performance in a multi-device environment. In our benchmark test, the GPU is found to produce the best arithmetic performance as a standalone device unit, and gives the most efficient power consumption. The multi-core CPU obtains the most effective computing performance. The computational speed of the MIC processor on Xeon Phi approached that of two Xeon CPUs. This indicates that using MICs is an attractive choice for existing SPH codes on multi-core CPUs parallelized by OpenMP, as it gains computational acceleration without the need for significant changes to the source code.

Self-organized domain microstructures in a plate-like particle suspension subjected to rapid simple shear

The evolution of the microstructure and rheological properties of plate-like particle suspensions subjected to rapid simple shear is studied numerically. In response to the shear-induced strain, particles in the suspensions rearrange to form a steady-state microstructure, and the suspension viscosity reaches a steady value. Under this condition, the microstructure is composed of two domains having different particle fractions and particle orientations. In the matrix (particle-poor) and cluster (particle-rich) domains, the particles’ long axes are oriented subparallel to the shear plane and normal to the maximum compressive principal direction, respectively. A higher particle concentration and friction coefficient enhance the development of cluster domains relative to matrix domains leading the intensity of the preferred particle orientation to decrease and the number of contacting particles, the aspect ratio of clusters, the inter-particle force, and the suspension viscosity to increase. The domain microstructure is governed by two factors: (1) geometric relations between the particle orientation and the maximum compressive axes and (2) the magnitude of particle–fluid and particle–particle interactions. The first factor results in the coupling of the particle orientation and the local fraction of particles, which is an important character of the domain microstructure. The second factor controls the relative development of the cluster and matrix domains through the change in the particles’ rotational behavior. Our results suggest that the microstructure of plate-like suspensions subjected to rapid shear is predictable in terms of the cluster stability, which has important implications for the kinematics of flow-related microstructures in nature and manufacturing.

3D dynamic simulation of crack propagation in extracorporeal shock wave lithotripsy

Some experimental observations of Shock Wave Lithotripsy(SWL), which include 3D dynamic crack propagation, are simulated with the aim of reproducing fragmentation of kidney stones with SWL. Extracorporeal shock wave lithotripsy (ESWL) is the fragmentation of kidney stones by focusing an ultrasonic pressure pulse onto the stones. 3D models with fine discretization are used to accurately capture the high amplitude shear shock waves. For solving the resulting large scale dynamic crack propagation problem, PDS-FEM is used; it provides numerically efficient failure treatments. With a distributed memory parallel code of PDS-FEM, experimentally observed 3D photoelastic images of transient stress waves and crack patterns in cylindrical samples are successfully reproduced. The numerical crack patterns are in good agreement with the experimental ones, quantitatively. The results shows that the high amplitude shear waves induced in solid, by the lithotriptor generated shock wave, play a dominant role in stone fragmentation.

Evidence for erosion and deposition by the 2011 Tohoku-oki tsunami on the nearshore shelf of Sendai Bay, Japan

Ongoing geological research into processes operating on the nearshore continental shelf and beyond is vital to our understanding of modern tsunami-generated sediment transport and deposition. This paper investigates the southern part of Sendai Bay, Japan, by means of high-resolution seismic surveys, vibracoring, bathymetric data assimilation, and radioisotope analysis of a core. For the first time, it was possible to identify an erosional surface in the shallow subsurface, formed by both seafloor erosion and associated offshore-directed sediment transport caused by the 2011 Tohoku-oki tsunami. The area of erosion and deposition extends at least 1,100 m offshore from the shoreline down to water depths of 16.7 m. The tsunami-generated sedimentological signature reaches up to 1.2 m below the present seafloor, whereas bathymetric changes due to storm-related reworking over a period of 3 years following the tsunami event have been limited to the upper ~0.3 m, despite the fact that the study area is located on an open shelf facing the Pacific Ocean. Tsunami-generated erosion surfaces may thus be preserved for extended periods of time, and may even enter the rock record, because the depth of tsunami erosion can exceed the depth of storm erosion. This finding is also important for interpretation of modern submarine strata, since erosion surfaces in shallow (depths less than ~1 m) seismic records from open coast shelves have generally been interpreted as storm-generated surfaces or transgressive ravinement surfaces.

Development of Viscoelastic Multi-Body Simulation and Impact Response Analysis of a Ballasted Railway Track under Cyclic Loading

Simulation of a large number of deformable bodies is often difficult because complex high-level modeling is required to address both multi-body contact and viscoelastic deformation. This necessitates the combined use of a discrete element method (DEM) and a finite element method (FEM). In this study, a quadruple discrete element method (QDEM) was developed for dynamic analysis of viscoelastic materials using a simpler algorithm compared to the standard FEM. QDEM easily incorporates the contact algorithm used in DEM. As the first step toward multi-body simulation, the fundamental performance of QDEM was investigated for viscoelastic analysis. The amplitude and frequency of cantilever elastic vibration were nearly equal to those obtained by the standard FEM. A comparison of creep recovery tests with an analytical solution showed good agreement between them. In addition, good correlation between the attenuation degree and the real physical viscosity was confirmed for viscoelastic vibration analysis. Therefore, the high accuracy of QDEM in the fundamental analysis of infinitesimal viscoelastic deformations was verified. Finally, the impact response of a ballast and sleeper under cyclic loading on a railway track was analyzed using QDEM as an application of deformable multi-body dynamics. The results showed that the vibration of the ballasted track was qualitatively in good agreement with the actual measurements. Moreover, the ballast layer with high friction reduced the ballasted track deterioration. This study suggests that QDEM, as an alternative to DEM and FEM, can provide deeper insights into the contact dynamics of a large number of deformable bodies.

Shear resistance reduction due to vibration in simulated fault gouge

In order to investigate the mechanisms of dynamic triggering for earthquakes or creep events on natural faults with gouge layers, the frictional behavior of granular materials and its sensitivity to vibrational disturbances were examined by means of a direct shear test with crushed quartz sand and precise measurement equipments. The disturbances were created by light tapping on the lower part of a shear box. No displacement was observed from the tapping without shear load. From a series of experiments we found an acceleration in the horizontal displacement just after small vibration under shear. The acceleration indicates reduced shear resistance in the gouge layer, by about 3 % of the resistance just before the small vibration. This reduction in shear resistance seemed to be recovered soon and did not affect the long-term behavior of the gouge layer. The response of the vertical displacement to the small vibration depends on the amount of accumulated dilatation of the gouge layer. The mechanisms of shear resistance reduction and the variable response of the vertical displacement due to the vibration can be explained by the intrinsic feature of pillar-like structure of a force chain network in the gouge layer. Our results indicate that there might be dynamic triggering of fault motion due to shear resistance reduction of a gouge layer in a natural fault zone.

Experimental Study on the Characteristics of VIV and Whirl Motion of Rotating Drill Pipe

During drilling operations, the scientific drilling vessel Chikyu encounters vortex-induced vibration (VIV) of the riser during riser drilling as well as that of the drill pipe during riserless drilling. The characteristics of the VIV of the drill pipe will differ from those of the riser because the drill pipe rotates during operations. Therefore, investigating the characteristics of the VIV during drill pipe rotation is important. In this paper, the results of model experiments that show the characteristics of the VIV of a rotating drill pipe and drill pipe whirl are described. The authors conducted model experiments on the VIV response of a rotating cylinder in a water tank to observe the motion of the rotating cylinder. The results confirmed that rotating affects the characteristics of the VIV. When the rotating frequency of the drill pipe was low, the VIV resonated with the rotation; as the rotating frequency increased, the frequency of the VIV decreased. In addition, a self-excited whirl motion was also observed during the experiments. This “drill pipe whirl,” which rotated in the direction opposite to the rotation of the drill pipe at a low frequency, was larger in amplitude than the VIV. The results demonstrate that rotation of the drill pipe affects the characteristics of the VIV and generates the whirl motion.© 2013 ASME