Massive Parallel Computing Research Articles

THC-MP: High performance numerical simulation of reactive transport and multiphase flow in porous media

The numerical simulation of multiphase flow and reactive transport in the porous media on complex subsurface problem is a computationally intensive application. To meet the increasingly computational requirements, this paper presents a parallel computing method and architecture. Derived from TOUGHREACT that is a well-established code for simulating subsurface multi-phase flow and reactive transport problems, we developed a high performance computing THC-MP based on massive parallel computer, which extends greatly on the computational capability for the original code. The domain decomposition method was applied to the coupled numerical computing procedure in the THC-MP. We designed the distributed data structure, implemented the data initialization and exchange between the computing nodes and the core solving module using the hybrid parallel iterative and direct solver. Numerical accuracy of the THC-MP was verified through a CO2 injection-induced reactive transport problem by comparing the results obtained from the parallel computing and sequential computing (original code). Execution efficiency and code scalability were examined through field scale carbon sequestration applications on the multicore cluster. The results demonstrate successfully the enhanced performance using the THC-MP on parallel computing facilities.

A novel locally guided genome reassembling technique using an artificial ant system

DNA reassembling is an NP-hard problem (Brun, Theor Comput Sci 395:31---46, 2008; Medvedev et al 2007; Ma and Lombardi 2008). The present article presents a locally guided global learning system to solve the problem of genome reassembling. We have used a reference DNA sequence which is 99 % similar to an unknown DNA sequence. Two different sequences from the same organism generally have around 99 % similarity (Wei et al 2007). We have considered different DNA sequences from NCBI website (http://www.ncbi.nlm.nih.gov). Then we have simulated the tasks of cloning the sequence, followed by shearing the clones to a number of short reads. In our algorithm, we have introduced a new concept in the task of DNA reassembling using Ant Colony Optimization, where pheromone concentration is proportional to the score of assembled DNA fragments with some known reference sequences within the same organism. Unlike local overlapping, we have used here local alignment score of short reads with some known local reference region as the heuristic information. The result shows that our algorithm is capable of reassembling at par with the state-of-the-art. DNA reassembling techniques may need a massive parallel computation and huge memory space (Kurniawan et al 2008) because of size ~109bp of DNA sequences of mammals (Miller et al, Genomics 95:315---327, 2010; Blazewicz et al, Comput Biol Chem 33:224---230, 2009; Butler et al, Genome Res 18:810---820, 2008; Joshi et al 2011; Stupar et al, Arch Oncol 19:3---4, 2011; Quail et al, BMC Genomics 13:1471---2164, 2012), and ACO is inherently concurrent in nature (Dorigo and Stutzle 2004). Due to lack of appropriate computational resources, we had to confine ourselves to deal with the sequences of length up to ~105bp. We have considered 22 sequences of different organism, including Homo sapiens BRCA1 (127429bp) gene. For large sequences, we have applied hierarchical BAC-by-BAC sequencing (Fig. 2) (Myers, Comput Sci Eng 1:33---43, 1999), to stitch the individual segments to retrieve the original DNA sequence.

A GPU-Accelerated Parallel Shooting Algorithm for Analysis of Radio Frequency and Microwave Integrated Circuits

This paper presents a new parallel shooting-Newton method based on a graphic processing unit (GPU)-accelerated periodic Arnoldi shooting solver (GAPAS) for fast periodic steady-state analysis of radio frequency/millimeter-wave integrated circuits. The new algorithm first explores a periodic structure of the state matrix by using a periodic Arnoldi algorithm for computing the resulting structured Krylov subspace in the generalized minimal residual (GMRES) solver. The resulting periodic Arnoldi shooting method is very amenable for massive parallel computing, such as GPUs. Second, the periodic Arnoldi-based GMRES solver in the shooting-Newton method is parallelized on the recent NVIDIA Tesla GPU platforms. We further explore CUDA GPUs features, such as coalesced memory access and overlapping transfers with computation to boost the efficiency of the resulting parallel GAPAS method. Experimental results from several industrial examples show that when compared with the state-of-the-art implicit GMRES method under the same accuracy, the new parallel shooting-Newton method can lead up to $8\times$ speedup.

Guest Editorial Special Issue on Cloud Robotics and Automation

The articles in this special section focus on the use of cloud computing in the robotics industry. The Internet and the availability of vast computational resources, ever-growing data and storage capacity have the potential to define a new paradigm for robotics and automation. An intelligent system connected to the Internet can expand its onboard local data, computation and sensors with huge data repositories from similar and very different domains, massive parallel computation from server farms and sensor/actuator streams from other robots and automata. It is the potential and also the research challenges of the field that become the focus on this special section. The goal is to group together and to show the state-of-the-art of this newly emerged field, identify the relevant advances and topics, point out the current lines of research and potential applications, and discuss the main research challenges and future work directions.

Searching for Sinks for the Hénon Map using a Multipleprecision GPU Arithmetic Library

Today, GPUs represent an important hardware development platform for many problems in dynamical systems, where massive parallel computations are needed. Beside that, many numerical studies of chaotic dynamical systems require a computing precision higher than common oating point (FP) formats. One such application is locating invariant sets for chaotic dynamical systems. In particular, we focus on rigorously proving the existence of stable periodic orbits for the Hénon map for parameter values close to the classical ones. For that, we present a multiple-precision floating-point arithmetic library in CUDA programming language for the NVIDIA GPU platform. Our library extends the precision using so-called FP expansions, where a number is represented as the unevaluated sum of standard machine precision FP numbers. This format offers the advantage of using directly available and highly optimized hardware FP operations. We generalize algorithms used by multiple-precisions libraries such as Bailey's QD, or the analogue GPU version, GQD.

고차 정확도 수치기법의 GPU 계산을 통한 효율적인 압축성 유동 해석

The present paper deals with the efficient computation of higher-order CFD methods for compressible flow using graphics processing units (GPU). The higher-order CFD methods, such as discontinuous Galerkin (DG) methods and correction procedure via reconstruction (CPR) methods, can realize arbitrary higher-order accuracy with compact stencil on unstructured mesh. However, they require much more computational costs compared to the widely used finite volume methods (FVM). Graphics processing unit, consisting of hundreds or thousands small cores, is apt to massive parallel computations of compressible flow based on the higher-order CFD methods and can reduce computational time greatly. Higher-order multi-dimensional limiting process (MLP) is applied for the robust control of numerical oscillations around shock discontinuity and implemented efficiently on GPU. The program is written and optimized in CUDA library offered from NVIDIA. The whole algorithms are implemented to guarantee accurate and efficient computations for parallel programming on shared-memory model of GPU. The extensive numerical experiments validates that the GPU successfully accelerates computing compressible flow using higher-order method.

Real‐Time Frequency‐Domain Decomposition for Structural Health Monitoring Using General‐Purpose Graphic Processing Unit

AbstractFrequency‐domain decomposition (FDD) is used in civil engineering to identify the modal properties of structures by analyzing the data output of structural health monitoring (SHM) systems. However, because FDD is computationally expensive, it prevents CPUs from achieving real‐time performance. A CPU takes seconds to perform FDD of 16 input signals but minutes to perform FDD of hundreds of input signals; and the deployed SHM systems are becoming larger and larger. Instead, a supercomputer can achieve real‐time performance but it cannot be installed near a civil structure because it is bulky, expensive, and requires constant maintenance. In this study, FDD is performed using general‐purpose graphic processor unit (GPGPU). A GPU is capable of massive parallel computing. The developed parallel FDD algorithm is up to hundreds of times faster than its serial version on CPU. For SHM of civil structures, where natural frequencies are less than 20 Hz parallel FDD on a single GPU achieves real‐time performance. The use of GPGPU offers many advantages. The modal properties are tracked in real time. A GPU can be installed inside the base station at a structure site. A GPU is energy efficient and does not require the maintenance of a supercomputer.

Scalable rank-mapping algorithm for an icosahedral grid system on the massive parallel computer with a 3-D torus network

Abstract In this paper, we develop a rank-mapping algorithm for an icosahedral grid system on a massive parallel computer with the 3-D torus network topology, specifically on the K computer. Our aim is to improve the weak scaling performance of the point-to-point communications for exchanging grid-point values between adjacent grid regions on a sphere. We formulate a new rank-mapping algorithm to reduce the maximum number of hops for the point-to-point communications. We evaluate both the new algorithm and the standard ones on the K computer, using the communication kernel of the Nonhydrostatic Icosahedral Atmospheric Model (NICAM), a global atmospheric model with an icosahedral grid system. We confirm that, unlike the standard algorithms, the new one achieves almost perfect performance in the weak scaling on the K computer, even for 10,240 nodes. Results of additional experiments imply that the high scalability of the new rank-mapping algorithm on the K computer is achieved by reducing network congestion in the links between adjacent nodes.

SU-E-T-180: Fano Cavity Test of Proton Transport in Monte Carlo Codes Running On GPU and Xeon Phi

Purpose: In proton dose calculation, clinically compatible speeds are now achieved with Monte Carlo codes (MC) that combine 1) adequate simplifications in the physics of transport and 2) the use of hardware architectures enabling massive parallel computing (like GPUs). However, the uncertainties related to the transport algorithms used in these codes must be kept minimal. Such algorithms can be checked with the so-called “Fano cavity test”. We implemented the test in two codes that run on specific hardware: gPMC on an nVidia GPU and MCsquare on an Intel Xeon Phi (60 cores). Methods: gPMC and MCsquare are designed for transporting protons in CT geometries. Both codes use the method of fictitious interaction to sample the step-length for each transport step. The considered geometry is a water cavity (2×2×0.2 cm3, 0.001 g/cm3) in a 10×10×50 cm3 water phantom (1 g/cm3). CPE in the cavity is established by generating protons over the phantom volume with a uniform momentum (energy E) and a uniform intensity per unit mass I. Assuming no nuclear reactions and no generation of other secondaries, the computed cavity dose should equal IE, according to Fano's theorem. Both codes were tested for initial proton energies of 50, 100, and 200 MeV. Results: For all energies, gPMC and MCsquare are within 0.3 and 0.2 % of the theoretical value IE, respectively (0.1% standard deviation). Single-precision computations (instead of double) increased the error by about 0.1% in MCsquare. Conclusion: Despite the simplifications in the physics of transport, both gPMC and MCsquare successfully pass the Fano test. This ensures optimal accuracy of the codes for clinical applications within the uncertainties on the underlying physical models. It also opens the path to other applications of these codes, like the simulation of ion chamber response.

Scalable Evaluation of Polarization Energy and Associated Forces in Polarizable Molecular Dynamics: II.Towards Massively Parallel Computations using Smooth Particle Mesh Ewald.

In this paper, we present a scalable and efficient implementation of point dipole-based polarizable force fields for molecular dynamics (MD) simulations with periodic boundary conditions (PBC). The Smooth Particle-Mesh Ewald technique is combined with two optimal iterative strategies, namely, a preconditioned conjugate gradient solver and a Jacobi solver in conjunction with the Direct Inversion in the Iterative Subspace for convergence acceleration, to solve the polarization equations. We show that both solvers exhibit very good parallel performances and overall very competitive timings in an energy-force computation needed to perform a MD step. Various tests on large systems are provided in the context of the polarizable AMOEBA force field as implemented in the newly developed Tinker-HP package which is the first implementation for a polarizable model making large scale experiments for massively parallel PBC point dipole models possible. We show that using a large number of cores offers a significant acceleration of the overall process involving the iterative methods within the context of spme and a noticeable improvement of the memory management giving access to very large systems (hundreds of thousands of atoms) as the algorithm naturally distributes the data on different cores. Coupled with advanced MD techniques, gains ranging from 2 to 3 orders of magnitude in time are now possible compared to non-optimized, sequential implementations giving new directions for polarizable molecular dynamics in periodic boundary conditions using massively parallel implementations.

Sampling the Denatured State of Polypeptides in Water, Urea, and Guanidine Chloride to Strict Equilibrium Conditions with the Help of Massively Parallel Computers.

The denatured state of polypeptides and proteins, stabilized by chemical denaturants like urea and guanidine chloride, displays residual secondary structure when studied by nuclear-magnetic-resonance spectroscopy. However, these experimental techniques are weakly sensitive, and thus molecular-dynamics simulations can be useful to complement the experimental findings. To sample the denatured state, we made use of massively-parallel computers and of a variant of the replica exchange algorithm, in which the different branches, connected with unbiased replicas, favor the formation and disruption of local secondary structure. The algorithm is applied to the second hairpin of GB1 in water, in urea, and in guanidine chloride. We show with the help of different criteria that the simulations converge to equilibrium. It results that urea and guanidine chloride, besides inducing some polyproline-II structure, have different effect on the hairpin. Urea disrupts completely the native region and stabilizes a state which resembles a random coil, while guanidine chloride has a milder effect.

Modeling of Parallel Computers Based on Network of Computing

The optimal resource allocation to satisfy such demands and the proper settlement of contention when demands exceed the capacity of the resources, constitute the problem of being able to understand and to predict system behavior. To this analysis we can use both analytical and simulation methods. Modeling and simulation are methods, which are commonly used by performance analysts to represent constraints and to optimize performance. Principally analytical methods represented first of all by queuing theory belongs to the preferred method in comparison to the simulation method, because of their potential ability of general analysis and also of their ability to potentially analyze also massive parallel computers. But these arguments supposed to develop and to verify suggested analytical models. This article goes further in applying the achieved analytical results in queuing theory for complex performance evaluation in parallel computing [9, 14]. The extensions are mainly in extending derived analytical models to whole range of parallel computers including massive parallel computers (Grid, meta computer). The article therefore describes standard analytical model based on M/M/m, M/D/m and M/M/1, M/D/1 queuing theory systems. Then the paper describes derivation of the correction factor for standard analytical model, based on M/M/m and M/M/1 queuing systems, to study more precise their basic performance parameters (overhead latencies, throughput etc.). All the derived analytical models were compared with performed simulation results in order to estimate the magnitude of improvement. Likewise they were tested under various ranges of parameters, which influence the architecture of the parallel computers and its communication networks too. These results are very important in practical use.

A Scalable 3D Heterogeneous Multi-Core Processor with Inductive-Coupling ThruChip Interface

A scalable heterogeneous multi-core processor is developed. 3D heterogeneous chip stacking of a general-purpose CPU and reconfigurable multi-core accelerators improves computational energy efficiency by proper task assignment and massive parallel computing. The stacked chips interconnect through a scalable 3D Network on Chip (NoC). By simply changing the number of stacked accelerator chips, processor parallelism can be widely scaled. In combination with Dynamic Voltage and Frequency Scaling (DVFS), the energy efficiency can be optimized for various performance requirements. No design change is needed, and hence no additional Non-Recurring Engineering (NRE) cost. An inductive-coupling ThruChip Interface (TCI) is applied to stacked-chip communications, forming a low-cost and robust high-speed 3D NoC. A prototype demonstration system has been developed with 65nm CMOS test chips. Successful system operations including 10-hours continuous Linux OS operation are confirmed for the first time.

Modeling of Communication Complexity in Parallel Computing

Parallel principles are the most effective way how to increase parallel computer performance and parallel algorithms (PA) too. Parallel using of more computing nodes (processors, cores), which have to cooperate each other in solving complex problems in a parallel way, opened imperative problem of modeling communication complexity so in symmetrical multiprocessors (SMP) based on motherboard as in other asynchronous parallel computers (computer networks, cluster etc.). In actually dominant parallel computers based on NOW and Grid (network of NOW networks) [31] there is necessary to model communication latency because it could be dominant at using massive (number of processors more than 100) parallel computers [17]. In this sense the paper is devoted to modeling of communication complexity in parallel computing (parallel computers and algorithms). At first the paper describes very shortly various used communication topologies and networks and then it summarized basic concepts for modeling of communication complexity and latency too. To illustrate the analyzed modeling concepts the paper considers in its experimental part the results for real analyzed examples of abstract square matrix and its possible decomposition models. These illustration examples we have chosen first due to wide matrix application in scientific and engineering fields and second from its typical exemplary representation for any other PA.

Unified Analytical Models of Parallel and Distributed Computing

The optimal resource allocation satisfies the needed capacity of the used resources. To such analysis we can use both analytical and simulation methods. Principally analytical methods (AM) belong to the preferred method in comparison to the simulation method, because of their potential ability of more general analysis and also of ability to analyze massive parallel computers. This article goes further in developing AM based on queuing theory results in relation to our published paper in [9]. The extensions are in extending derived AM to whole range of parallel computers and also to sum up public acceptance of our published paper. The article therefore describes deriving of correction factor of standard AM based on M/M/m and M/M/1queuing theory systems. In detail the paper describes derivation of a correction factor for standard AM to study more precise their performance. The paper contributions are in unified AM and in deriving correction factor in order to take into account real non-exponential nature of the inputs to the computing nodes and node’s communication channels. The derived analytical results were compared with performed simulation results in order to estimate the magnitude of improvement. Likewise the corrected AM were tested under various ranges of parameters, which influence the architecture of the parallel computers and its communication networks too. These results are very important in practical use.

Large Scale Joint Inversion of Geophysical Data using the Finite Element Method in escript

Inversion is finding the minimum value of a cost function measuring data defect and physical property variation subject to constraints defined by a partial differential equation (PDE). In general, this problem is equivalent to the solution of a system for coupled PDEs for three unknowns -- the physical property, the observation and a Lagrangean multiplier. Therefore it is appropriate to use established PDE solution methods such as the finite element method (FEM) and solver software systems such as escript (see https://launchpad.net/escript-finley) to tackle inversion problems. Besides the ability to handle non-linearity in the physical model (e.g. required for high susceptibility) as well as in the regularisation term or in the cross-gradient terms the PDE approach provides a number of computational advantages in comparison to traditional, linear algebra based solution approaches. The method is data sparse by nature and does not require compression or sparsification on the sensitivity matrix when solving large scale inversion problems. Moreover, domain decomposition can be applied to run the inversion across processors in a parallel computer. In contrast to traditional tiling, domain decomposition is not imposed on the inversion from the outside but applied on the lowest level which makes the approach more computaional more efficient. In the presentation we will introduce the concept of PDE based, joint inversion of gravity and magnetic data. We will outline the appropriate solution methods when using the finite element method, outline the implementation strategy using the PDE solver escript and show results from joint inversion runs of field data using massive parallel computer.

Accelerating Pattern Matching Using a Novel Parallel Algorithm on GPUs

Graphics processing units (GPUs) have attracted a lot of attention due to their cost-effective and enormous power for massive data parallel computing. In this paper, we propose a novel parallel algorithm for exact pattern matching on GPUs. A traditional exact pattern matching algorithm matches multiple patterns simultaneously by traversing a special state machine called an Aho-Corasick machine. Considering the particular parallel architecture of GPUs, in this paper, we first propose an efficient state machine on which we perform very efficient parallel algorithms. Also, several techniques are introduced to do optimization on GPUs, including reducing global memory transactions of input buffer, reducing latency of transition table lookup, eliminating output table accesses, avoiding bank-conflict of shared memory, coalescing writes to global memory, and enhancing data transmission via peripheral component interconnect express. We evaluate the performance of the proposed algorithm using attack patterns from Snort V2.8 and input streams from DEFCON. The experimental results show that the proposed algorithm performed on NVIDIA GPUs achieves up to 143.16-Gbps throughput, 14.74 times faster than the Aho-Corasick algorithm implemented on a 3.06-GHz quad-core CPU with the OpenMP. The library of the proposed algorithm is publically accessible through Google Code.

Deep Web and MapReduce

This invited paper introduces results on Web science and technology obtained during work with the Korea Advanced Institute of Science and Technology. In the first part, we discuss algorithms for exploring the deep Web, which refers to the collection of Web pages that cannot be reached by conventional Web crawlers. In the second part, we discuss sorting algorithms on the MapReduce system, which has become a dominant paradigm for massive parallel computing.

A parallel Jacobian-free Newton–Krylov solver for a coupled sea ice-ocean model

The most common representation of sea ice dynamics in climate models assumes that sea ice is a quasi-continuous non-normal fluid with a viscous-plastic rheology. This rheology leads to non-linear sea ice momentum equations that are notoriously difficult to solve. Recently a Jacobian-free Newton–Krylov (JFNK) solver was shown to solve the equations accurately at moderate costs. This solver is extended for massive parallel architectures and vector computers and implemented in a coupled sea ice–ocean general circulation model for climate studies. Numerical performance is discussed along with numerical difficulties in realistic applications with up to 1920 CPUs. The parallel JFNK-solverʼs scalability competes with traditional solvers although the collective communication overhead starts to show a little earlier. When accuracy of the solution is required (i.e. reduction of the residual norm of the momentum equations of more that one or two orders of magnitude) the JFNK-solver is unrivalled in efficiency. The new implementation opens up the opportunity to explore physical mechanisms in the context of large scale sea ice models and climate models and to clearly differentiate these physical effects from numerical artifacts.

SpiNNaker: Fault tolerance in a power- and area- constrained large-scale neuromimetic architecture

SpiNNaker is a biologically-inspired massively-parallel computer designed to model up to a billion spiking neurons in real-time. A full-fledged implementation of a SpiNNaker system will comprise more than 105 integrated circuits (half of which are SDRAMs and half multi-core systems-on-chip). Given this scale, it is unavoidable that some components fail and, in consequence, fault-tolerance is a foundation of the system design. Although the target application can tolerate a certain, low level of failures, important efforts have been devoted to incorporate different techniques for fault tolerance. This paper is devoted to discussing how hardware and software mechanisms collaborate to make SpiNNaker operate properly even in the very likely scenario of component failures and how it can tolerate system-degradation levels well above those expected.