Serial Simulations Research Articles

Large-scale three-dimensional phase-field simulations for phase coarsening at ultrahigh volume fraction on high-performance architectures

A parallel algorithm for large-scale three-dimensional phase-field simulations of phase coarsening is developed and implemented on high-performance architectures. From the large-scale simulations, a new kinetics in phase coarsening in the region of ultrahigh volume fraction is found. The parallel implementation is capable of harnessing the greater computer power available from high-performance architectures. The parallelized code enables increase in three-dimensional simulation system size up to a 5123 grid cube. Through the parallelized code, practical runtime can be achieved for three-dimensional large-scale simulations, and the statistical significance of the results from these high resolution parallel simulations are greatly improved over those obtainable from serial simulations. A detailed performance analysis on speed-up and scalability is presented, showing good scalability which improves with increasing problem size. In addition, a model for prediction of runtime is developed, which shows a good agreement with actual run time from numerical tests.

Quality between mechanical compression on reducible stretcher versus manual compression on standard stretcher in small elevator

ObjectivesManual cardiopulmonary resuscitation (CPR) during vertical transport in small elevators using standard stretcher for out-of-hospital cardiac arrest can raise concerns with diminishing quality. Mechanical CPR on a reducible stretcher (RS-CPR) that can be shortened in the length was tested to compare the CPR quality with manual CPR on a standard stretcher (SS-CPR). MethodsA randomized crossover manikin simulation was designed. Three teams of emergency medical technicians were recruited to perform serial CPR simulations using two different protocols (RS-CPR and SS-CPR) according to a randomization; the first 6 minutes of manual CPR at the scene was identical for both scenarios and two different protocols during vertical transport in a small elevator followed on a basis of cross-over assignment. The LUCAS-2 Chest Compression System (Zolife AB, Lund, Sweden) was used for RS-CPR. CPR quality was measured using a resuscitation manikin (Resusci Anne QCPR, Laerdal Medical, Stavanger, Norway) in terms of no flow fraction, compression depth, and rate (median and IQR). ResultsA total of 42 simulations were analyzed. CPR quality did not differ significantly at the scene. No flow fraction (%) was significantly lower when the stretcher was moving in RS-CPR then SS-CPR (36.0 (33.8-38.7) vs 44.0 (36.8-54.4), P< .01). RS-CPR showed significantly better quality than SS-CPR; 93.2 (50.6-95.6) vs 14.8 (0-20.8) for adequate depth (P< 0.01), and 97.5 (96.6-98.2) vs 68.9(43.4-78.5) for adequate rate (P< .01). ConclusionMechanical CPR on a reducible stretcher during vertical transport showed significant improvement in CPR quality in terms of no-flow fraction, compression depth, and rate compared with manual CPR on a standard stretcher.

Estimating numerical error in neural network simulations on Graphics Processing Units

Modern graphics processing units (GPUs) are becoming a popular hardware substrate for spiking neural network simulations [1-3], due to their massive parallelism and impressive cost-to-speed ratio. However, verifying and interpreting the results of a GPU simulation can be difficult because the results are never exactly reproducible, unlike those from an equivalent serial simulation on a CPU; not only is the simulation subject to the usual rounding errors of floating-point arithmetic, but there are also elements of stochasticity due to the non-determinism of the thread scheduling mechanism on the hardware, alongside the non-associativity of floating-point addition and multiplication. Consider, for example, a typical postsynaptic integration step for the summation of incoming synapse currents, executed on a GPU. If there are multiple threads, which are each simulating an incoming synapse, there is no guarantee for the order in which each synapse thread's current will be accumulated into the total current. Therefore, rounding errors will be different depending on this order and the result could be different every time the simulation runs. Such effects would initially be small but can be amplified in unstable or chaotic systems to a degree that the final results appear completely random across different runs (see Figure ​Figure11). Figure 1 In repeated runs, results of numerical simulations on GPUs can vary. Mean, standard deviation and range of observed membrane potential of a neuron in a network of 10,000 Izhikevich neurons, 8,000 excitatory and 2,000 inhibitory, with 1,000 random connections ... When comparing runs between GPU and CPU implementations there are additional sources of divergence. There are subtle differences in the way each architecture implements floating-point arithmetic. For instance, the NVIDIA C2070 GPU tested in this study implements the fused multiply-add (FMA) operation, introduced in the latest IEEE-754-2008 floating-point standard, whereas most Intel CPUs perform the multiplication and addition operations separately, with lower accuracy. Only Intel's most recent Haswell CPU architecture implements the more accurate FMA operation, whilst many current lab workstations contain chips that do not. The aim of the work presented here is to analytically determine the theoretical worst-case and average-case numerical absolute error incurred when simulating neural network models on an NVIDIA CUDA GPU. These error measurements are also compared with the absolute error resulting from the equivalent serial algorithm running on a single CPU core, using standard float32 (float) and float64 (double) precision floating-point arithmetic, to determine a reasonable error margin for verifying the results of parallel GPU simulations against those of equivalent serial CPU simulations. Furthermore, both CPU and GPU implementations are compared against an equivalent simulation using an accurate arbitrary-precision floating-point arithmetic library, to determine how far the CPU and GPU simulation trajectories deviate from the analytically 'correct' trajectory. For illustration, the divergence of a single neuron in a 10,000 neuron Izhikevich network is plotted in the figure. Finally, we also analyse the role of errors originating from approximate integration methods and compare them to the underlying numerical errors discussed thus far.

Simulations in generalized ensembles through noninstantaneous switches.

Generalized-ensemble simulations, such as replica exchange and serial generalized-ensemble methods, are powerful simulation tools to enhance sampling of free energy landscapes in systems with high energy barriers. In these methods, sampling is enhanced through instantaneous transitions of replicas, i.e., copies of the system, between different ensembles characterized by some control parameter associated with thermodynamical variables (e.g., temperature or pressure) or collective mechanical variables (e.g., interatomic distances or torsional angles). An interesting evolution of these methodologies has been proposed by replacing the conventional instantaneous (trial) switches of replicas with noninstantaneous switches, realized by varying the control parameter in a finite time and accepting the final replica configuration with a Metropolis-like criterion based on the Crooks nonequilibrium work (CNW) theorem. Here we revise these techniques focusing on their correlation with the CNW theorem in the framework of Markovian processes. An outcome of this report is the derivation of the acceptance probability for noninstantaneous switches in serial generalized-ensemble simulations, where we show that explicit knowledge of the time dependence of the weight factors entering such simulations is not necessary. A generalized relationship of the CNW theorem is also provided in terms of the underlying equilibrium probability distribution at a fixed control parameter. Illustrative calculations on a toy model are performed with serial generalized-ensemble simulations, especially focusing on the different behavior of instantaneous and noninstantaneous replica transition schemes.

Comparison of serial and parallel simulations of a corridor fire using FDS

Current fire simulators allow to model the course of fire in large areas and its impact on structure and equipment. This paper deals with a comparison of serial and parallel calculations of simulation of a corridor fire by the FDS (Fire Dynamics Simulator) system. In parallel case, the whole computational domain is divided into several computational meshes, the computation on each mesh is considered as a single MPI (Message Passing Interface) process realised on one computational core and communication between MPI processes is provided by MPI. The aim of this paper is to determine the size of error caused by parallelization of computation, which occurs at touches of computational meshes.

Optimizations of the energy grid search algorithm in continuous-energy Monte Carlo particle transport codes

In this work we propose, implement, and test various optimizations of the typical energy grid-cross section pair lookup algorithm in Monte Carlo particle transport codes. The key feature common to all of the optimizations is a reduction in the length of the vector of energies that must be searched when locating the index of a particle’s current energy. Other factors held constant, a reduction in energy vector length yields a reduction in CPU time. The computational methods we present here are physics-informed. That is, they are designed to utilize the physical information embedded in a simulation in order to reduce the length of the vector to be searched. More specifically, the optimizations take advantage of information about scattering kinematics, neutron cross section structure and data representation, and also the expected characteristics of a system’s spatial flux distribution and energy spectrum. The methods that we present are implemented in the OpenMC Monte Carlo neutron transport code as part of this work. The gains in computational efficiency, as measured by overall code speedup, associated with each of the optimizations are demonstrated in both serial and multithreaded simulations of realistic systems. Depending on the system, simulation parameters, and optimization method employed, overall code speedup factors of 1.2–1.5, relative to the typical single-nuclide binary search algorithm, are routinely observed.

Genealogical relationships between early medieval and modern inhabitants of Piedmont.

In the period between 400 to 800 AD, also known as the period of the Barbarian invasions, intense migration is documented in the historical record of Europe. However, little is known about the demographic impact of these historical movements, potentially ranging from negligible to substantial. As a pilot study in a broader project on Medieval Europe, we sampled 102 specimens from 5 burial sites in Northwestern Italy, archaeologically classified as belonging to Lombards or Longobards, a Germanic people ruling over a vast section of the Italian peninsula from 568 to 774. We successfully amplified and typed the mitochondrial hypervariable region I (HVR-I) of 28 individuals. Comparisons of genetic diversity with other ancient populations and haplotype networks did not suggest that these samples are heterogeneous, and hence allowed us to jointly compare them with three isolated contemporary populations, and with a modern sample of a large city, representing a control for the effects of recent immigration. We then generated by serial coalescent simulations 16 millions of genealogies, contrasting a model of genealogical continuity with one in which the contemporary samples are genealogically independent from the medieval sample. Analyses by Approximate Bayesian Computation showed that the latter model fits the data in most cases, with one exception, Trino Vercellese, in which the evidence was compatible with persistence up to the present time of genetic features observed among this early medieval population. We conclude that it is possible, in general, to detect evidence of genealogical ties between medieval and specific modern populations. However, only seldom did mitochondrial DNA data allow us to reject with confidence either model tested, which indicates that broader analyses, based on larger assemblages of samples and genetic markers, are needed to understand in detail the effects of medieval migration.

Interdisciplinary Disaster Drill Simulation: Laying the Groundwork for Further Research.

The aim of this study was to examine the effect of using serial simulations with progression through the nursing curriculum. Simulation provides a way to learn without fear of failure and increase critical thinking and clinical decision-making skills. Learning in an interdisciplinary simulation provides a greater understanding of teamwork and communication skills. The NLN/Jeffries Simulation Framework was used in an interactive disaster drill with role-playing patient actors and manikins. In a debriefing session, nursing and radiology students co-presented scenarios. Students displayed critical thinking and clinical decision-making skills. They reported an increase in self-confidence in caring for patients during a disaster, an increase in empathy, and learning by observing others. This pilot study revealed that an interdisciplinary disaster drill simulation experience was a positive learning experience for both nursing and radiology students.

OpenMP-accelerated SWAT simulation using Intel C and FORTRAN compilers: Development and benchmark

We developed a practical method to accelerate execution of Soil and Water Assessment Tool (SWAT) using open (free) computational resources. The SWAT source code (rev 622) was recompiled using a non-commercial Intel FORTRAN compiler in Ubuntu 12.04LTS Linux platform, and newly named iOMP-SWAT in this study. GNU utilities of make, gprof, and diff were used to develop the iOMP-SWAT package, profile memory usage, and check identicalness of parallel and serial simulations. Among 302SWAT subroutines, the slowest routines were identified using GNU gprof, and later modified using Open Multiple Processing (OpenMP) library in an 8-core shared memory system. In addition, a C wrapping function was used to rapidly set large arrays to zero by cross compiling with the original SWAT FORTRAN package. A universal speedup ratio of 2.3 was achieved using input data sets of a large number of hydrological response units. As we specifically focus on acceleration of a single SWAT run, the use of iOMP-SWAT for parameter calibrations will significantly improve the performance of SWAT optimization.

A parallel computational framework to solve flow and transport in integrated surface–subsurface hydrologic systems

Hydrologic modeling requires the handling of a wide range of highly nonlinear processes from the scale of a hill slope to the continental scale, and thus the computational efficiency of the model becomes a critical issue for water resource management. This work is aimed at implementing and evaluating a flexible parallel computing framework for hydrologic simulations by applying OpenMP in the HydroGeoSphere (HGS) model. HGS is a 3D control-volume finite element model that solves the nonlinear coupled equations describing surface–subsurface water flow, solute migration and energy transport. The computing efficiency of HGS is improved by three parallel computing schemes: 1) parallelization of Jacobian matrix assembly, 2) multi-block node reordering for performing LU solve efficiently, and 3) parameter privatization for reducing memory access latency. Regarding to the accuracy and consistency of the simulation solutions obtained with parallel computing, differences in the solutions are entirely due to use of a finite linear solver iteration tolerance, which produces slightly different solutions which satisfy the convergence tolerance. The maximum difference in the head solution between the serial and parallel simulations is less than 10−3 m, using typical convergence tolerances. Using the parallel schemes developed in this work, three key achievements can be summarized: (1) parallelization of a physically-based hydrologic simulator can be performed in a manner that allows the same code to be executed on various shared memory platforms with minimal maintenance; (2) a general, flexible and robust parallel iterative sparse-matrix solver can be implemented in a wide range of numerical models employing either structured or unstructured mesh; and (3) the methodology is flexible, especially for the efficient construction of the coefficient and Jacobian matrices, compared to other parallelized hydrologic models which use parallel library packages.

Climate change underlies global demographic, genetic, and cultural transitions in pre-Columbian southern Peru.

Several archaeological studies in the Central Andes have pointed at the temporal coincidence of climatic fluctuations (both long- and short-term) and episodes of cultural transition and changes of socioeconomic structures throughout the pre-Columbian period. Although most scholars explain the connection between environmental and cultural changes by the impact of climatic alterations on the capacities of the ecosystems inhabited by pre-Columbian cultures, direct evidence for assumed demographic consequences is missing so far. In this study, we address directly the impact of climatic changes on the spatial population dynamics of the Central Andes. We use a large dataset of pre-Columbian mitochondrial DNA sequences from the northern Rio Grande de Nasca drainage (RGND) in southern Peru, dating from ∼840 BC to 1450 AD. Alternative demographic scenarios are tested using Bayesian serial coalescent simulations in an approximate Bayesian computational framework. Our results indicate migrations from the lower coastal valleys of southern Peru into the Andean highlands coincident with increasing climate variability at the end of the Nasca culture at ∼640 AD. We also find support for a back-migration from the highlands to the coast coincident with droughts in the southeastern Andean highlands and improvement of climatic conditions on the coast after the decline of the Wari and Tiwanaku empires (∼1200 AD), leading to a genetic homogenization in the RGND and probably southern Peru as a whole.

A hybrid artificial bee colony assisted differential evolution algorithm for optimal reactive power flow

Optimal Reactive Power Flow (ORPF) is a branch problem in the gradual development of the optimal power flow problem. Differential Evolution (DE) has been proved to be a promising evolutionary algorithm for solving the ORPF problem, but it requires a relatively large population size to avoid premature convergence, which will increase the algorithm convergence time. On the other hand, Artificial Bee Colony (ABC) algorithm has been proved to have good global search ability. Integrating the respective advantages of DE and ABC, a hybrid ABC assisted DE algorithm, denoted as DE–ABC, is proposed in this study to overcome DE’s disadvantage of requiring large population size and strengthen the global search ability. At the last, the effectiveness of DE–ABC is verified by the serial simulations on the IEEE 14-bus, 30-bus and 57-bus system test cases. The simulation results show, in the case of achieving the same effect, the required population size of DE–ABC hybrid algorithm is greatly less than that of DE algorithm, the algorithm convergence time is less too and the algorithm is robust.

Pre-whaling genetic diversity and population ecology in eastern Pacific gray whales: insights from ancient DNA and stable isotopes.

Commercial whaling decimated many whale populations, including the eastern Pacific gray whale, but little is known about how population dynamics or ecology differed prior to these removals. Of particular interest is the possibility of a large population decline prior to whaling, as such a decline could explain the ∼5-fold difference between genetic estimates of prior abundance and estimates based on historical records. We analyzed genetic (mitochondrial control region) and isotopic information from modern and prehistoric gray whales using serial coalescent simulations and Bayesian skyline analyses to test for a pre-whaling decline and to examine prehistoric genetic diversity, population dynamics and ecology. Simulations demonstrate that significant genetic differences observed between ancient and modern samples could be caused by a large, recent population bottleneck, roughly concurrent with commercial whaling. Stable isotopes show minimal differences between modern and ancient gray whale foraging ecology. Using rejection-based Approximate Bayesian Computation, we estimate the size of the population bottleneck at its minimum abundance and the pre-bottleneck abundance. Our results agree with previous genetic studies suggesting the historical size of the eastern gray whale population was roughly three to five times its current size.

Real-world hydrologic assessment of a fully-distributed hydrological model in a parallel computing environment

► We quantify the performance of a new parallelization method for a distributed model. ► Parallel model performance is found to be dependent on the hydrologic variability. ► Load balancing methods improve parallel performance in particular for large domains. ► A wider range of distributed applications is now possible through parallelization. A major challenge in the use of fully-distributed hydrologic models has been the lack of computational capabilities for high-resolution, long-term simulations in large river basins. In this study, we present the parallel model implementation and real-world hydrologic assessment of the Triangulated Irregular Network (TIN)-based Real-time Integrated Basin Simulator (tRIBS). Our parallelization approach is based on the decomposition of a complex watershed using the channel network as a directed graph. The resulting sub-basin partitioning divides effort among processors and handles hydrologic exchanges across boundaries. Through numerical experiments in a set of nested basins, we quantify parallel performance relative to serial runs for a range of processors, simulation complexities and lengths, and sub-basin partitioning methods, while accounting for inter-run variability on a parallel computing system. In contrast to serial simulations, the parallel model speed-up depends on the variability of hydrologic processes. Load balancing significantly improves parallel speed-up with proportionally faster runs as simulation complexity (domain resolution and channel network extent) increases. The best strategy for large river basins is to combine a balanced partitioning with an extended channel network, with potential savings through a lower TIN resolution. Based on these advances, a wider range of applications for fully-distributed hydrologic models are now possible. This is illustrated through a set of ensemble forecasts that account for precipitation uncertainty derived from a statistical downscaling model.

Reordering and incomplete preconditioning in serial and parallel adaptive mesh refinement and coarsening flow solutions

SUMMARYThe effects of reordering the unknowns on the convergence of incomplete factorization preconditioned Krylov subspace methods are investigated. Of particular interest is the resulting preconditioned iterative solver behavior when adaptive mesh refinement and coarsening (AMR/C) are utilized for serial or distributed parallel simulations. As representative schemes, we consider the familiar reverse Cuthill–McKee and quotient minimum degree algorithms applied with incomplete factorization preconditioners to CG and GMRES solvers. In the parallel distributed case, reordering is applied to local subdomains for block ILU preconditioning, and subdomains are repartitioned dynamically as mesh adaptation proceeds. Numerical studies for representative applications are conducted using the object‐oriented AMR/C software system libMesh linked to the PETSc solver library. Serial tests demonstrate that global unknown reordering and incomplete factorization preconditioning can reduce the number of iterations and improve serial CPU time in AMR/C computations. Parallel experiments indicate that local reordering for subdomain block preconditioning associated with dynamic repartitioning because of AMR/C leads to an overall reduction in processing time. Copyright © 2011 John Wiley &amp; Sons, Ltd.

Development and implementation of a polydispersed multiphase flow model in OpenFOAM

Abstract The two-phase flow solver implemented in the open-source OpenFOAM code was extended to a multiphase flow formulation ( n dispersed and one continuous phases) and then coupled to the population balance equation (PBE) solution by the Direct Quadrature Method of Moments (DQMOM), originating a polydispersed multiphase flow solver. Although each dispersed phase has its own velocity field, the present implementation considers only the interfacial momentum exchange between the continuous and the dispersed phases. The multiphase flow formulation was described and the details of the PBE–CFD coupling algorithms in OpenFOAM were provided. The implementation of the multiphase flow code was verified and evaluated against the original OpenFOAM two-phase flow solver for flow through a 2D backward facing step, using simplified breakage and aggregation kernels. The computational cost of both codes were compared for serial and parallel simulations.

Serial tempering without exchange

Serial tempering is a computational method that turns the temperature T (or more generally any independent λ parameter) into a dynamical variable. It is shown that, under conditions for which this variable is fast, serial tempering is equivalent to the umbrella sampling method with a single effective potential. This equivalence is demonstrated using both a small one-dimensional system and a small solvated peptide. The suggestion is then made to replace the serial tempering protocol with the equivalent umbrella sampling calculation. This approach, serial tempering without exchange (STeWiE), has the same performance as serial tempering in the limit that exchanges are frequent, is simpler to implement, and has fewer adjustable parameters than conventional serial tempering. The equivalence of serial tempering and STeWiE also provides a convenient route for estimating and optimizing the performance of serial tempering simulations and other generalized-ensemble methods.

Optimal Weights in Serial Generalized-Ensemble Simulations.

In serial generalized-ensemble simulations, the sampling of a collective coordinate of a system is enhanced through non-Boltzmann weighting schemes. A popular version of such methods is certainly the simulated tempering technique, which is based on a random walk in temperature ensembles to explore the phase space more thoroughly. The most critical aspect of serial generalized-ensemble methods with respect to their parallel counterparts, such as replica exchange, is the difficulty of weight determination. Here we propose an adaptive approach to update the weights on the fly during the simulation. The algorithm is based on generalized forms of the Bennett acceptance ratio and of the free energy perturbation. It does not require intensive communication between processors and, therefore, is prone to be used in distributed computing environments with modest computational cost. We illustrate the method in a series of molecular dynamics simulations of a model system and compare its performances to two recent approaches, one based on adaptive Bayesian-weighted histogram analysis and the other based on initial estimates of weight factors obtained by potential energy averages.

Interactions Between a Voltage Sensor and a Toxin via Multiscale Simulations

Gating-modifier toxins inhibit voltage-gated ion channels by binding the voltage sensors (VS) and altering the energetics of voltage-dependent gating. These toxins are thought to gain access to the VS via the membrane (i.e., by partitioning from water into the membrane before binding the VS). We used serial multiscale molecular-dynamics (MD) simulations, via a combination of coarse-grained (CG) and atomistic (AT) simulations, to study how the toxin VSTx1, which inhibits the archeabacterial voltage-gated potassium channel KvAP, interacts with an isolated membrane-embedded VS domain. In the CG simulations, VSTx1, which was initially located in water, partitioned into the headgroup/water interface of the lipid bilayer before binding the VS. The CG configurations were used to generate AT representations of the system, which were subjected to AT-MD to further evaluate the stability of the complex and refine the predicted VS/toxin interface. VSTx1 interacted with a binding site on the VS formed by the C-terminus of S1, the S1-S2 linker, and the N-terminus of S4. The predicted VS/toxin interactions are suggestive of toxin-mediated perturbations of the interaction between the VS and the pore domain of Kv channels, and of the membrane. Our simulations support a membrane-access mechanism of inhibition of Kv channels by VS toxins. Overall, the results show that serial multiscale MD simulations may be used to model a two-stage process of protein-bilayer and protein-protein interactions within a membrane.

A Parallel Numerical Study of Transient Heat Transfer and Fluid Flow of Weld Pool during Laser Keyhole Welding

Numerical simulation provides a way to improve our understandings of the heat transfer and fluid flow behaviors of the weld pool during laser keyhole welding. However, current numerical studies are only limited to serial simulations which running on a single CPU. In this study, a parallel numerical study of the heat transfer and fluid flow of the weld pool is presented. A mathematical model considering the effect of Marangoni force, buoyancy force, friction force of the mushy zone region and the effect of keyhole is presented. A combined keyhole volume and surface heat source model is also developed. The coupled transient heat transfer and Navier-stokes equations are solved with a high order accuracy parallel projection method. The simulation code is parallelized with the OpenMP language. It is shown that 200% speedup can be achieved on a shared memory quad-core CPU using the presented parallel simulation system. The simulation results agree well with the in-situ high speed CCD video imaging experiments and the literature results.