Multiprocessor Computer Systems Research Articles

Computational aspects in controlling filtration of fluids and gases in porous media

To control filtration of gases, liquid hydrocarbons, oil, and water in homogeneous and heterogeneous porous media, it is necessary to obtain a multivariate solution to systems of nonlinear or quasi-linear equations in partial differentials of the parabolic type that define hydrodynamic (mathematical) models of control objects. In this paper, we propose hierarchical multigrid variants of balance and variational methods together with methods of domain decomposition, splitting with respect to physical processes and spatial coordinates. The paper consists of two parts. In the first part we consider a model of single-phase filtration of gas in gas-field developing; in the second, a model of two-phase filtration of oil and water, i.e., oil fields. This statement proves the universal character of the proposed results. Due to the multilevel partition of the original initial boundary value problem, we can construct economical solution algorithms using multiprocessor computer systems of the cluster type of parallel action for equations of the model.

Numerical studies of flow around aircraft using the multiprocessor computer systems

The topical problems of the computational aerodynamics were described. Consideration was given to the realization of algorithms on the multiprocessor computer systems. Difficulties arising with such problems were described. Examples of numerical modeling of some problems were presented.

Topological optimization of the large-scale data transmission networks

Consideration was given to the design of the topological structure of the doubly connected large-scale networks by the cost criterion and constraints on the diameter. An exact combinatorial algorithm to solve the problem and a parallel algorithm to carry out calculations on the multiprocessor data-flow computer system were proposed. Parallelization was shown to provide a substantial gain in the time of computations in the optimal topological structure of the large-scale networks.

Numerical simulation of X-ray scattering processes during radiographic inspection of materials

A technique for numerical simulation of the processes of forming X-ray radiography images of complex multicomponent objects with allowance for the penetrating-radiation absorption and scattering processes is proposed. The technique proposed is based on an efficient description of complex 3D objects with piecewise-homogeneous structures and uses precisely specified shells separating homogeneous components of an object. Being combined with a developed version of the Monte Carlo method, this approach yields an efficient computational apparatus for analyzing the regularities of the formation of radiographic images of objects’ internal structures. This apparatus is intended for state-of-the-art multiprocessor computing systems. Having been obtained via the technique developed, the results of the numerical analysis of the effect of different mechanisms of interaction between photons and matter (including coherent and incoherent scattering with allowance for bonds of electrons in atoms) on the formed radiographic image of the object’s internal structure are discussed. The advantage of the described technique over the known MCNP program complex is shown.

3D Unsteady RANS Modeling of Complex Hydraulic Engineering Flows. I: Numerical Model

A general-purpose numerical method is developed for solving the full three-dimensional (3D), incompressible, unsteady Reynolds-averaged Navier-Stokes (URANS) equations in natural river reaches containing complex hydraulic structures at full-scale Reynolds numbers. The method adopts body-fitted, chimera overset grids in conjunction with a grid-embedding strategy to accurately and efficiently discretize arbitrarily complex, multiconnected flow domains. The URANS and turbulence closure equations are discretized using a second-order accurate finite-volume approach. The discrete equations are integrated in time via a dual-time-stepping, artificial compressibility method in conjunction with an efficient coupled, block-implicit, approximate factorization iterative solver. The computer code is parallelized to take full advantage of multiprocessor computer systems so that unsteady solutions on grids with 106 nodes can be obtained within reasonable computational time. The power of the method is demonstrated by appl...

Kinetic schemes and quasi-gas-dynamic system of equations

This paper deals with the description of kinetic schemes and the quasi-gas-dynamic system of equations, which is a promising trend for modelling complex gas dynamic flows. The possibilities of kinetic schemes and the quasi-gas-dynamic system as a model for describing a viscous gas flow are discussed. Particular attention is given to the possibility of their application to the computation of gas dynamics problems on high-performance multiprocessor computer systems. The possibilities to use the specified approaches for the solution of other problems of continuum mechanics are considered.

Static Forecasting of the Execution Times of Complexes of Interrelated Jobs in the Multiprocessor Computer Systems

Consideration was given to a procedure of approximate forecasting of the execution times of the complexes of interrelated jobs in the multiprocessor computer systems with the Erlangian execution time of each job. This distribution is used as a tool for modeling the normal distribution. The order and parameter of the approximating Erlangian distribution are chosen from the parameters of the normal distribution.

Dynamic modeling of Tampa Bay urban development using parallel computing

Urban land use and land cover has changed significantly in the environs of Tampa Bay, Florida, over the past 50 years. Extensive urbanization has created substantial change to the region's landscape and ecosystems. This paper uses a dynamic urban-growth model, SLEUTH, which applies six geospatial data themes (slope, land use, exclusion, urban extent, transportation, hillside), to study the process of urbanization and associated land use and land cover change in the Tampa Bay area. To reduce processing time and complete the modeling process within an acceptable period, the model is recoded and ported to a Beowulf cluster. The parallel-processing computer system accomplishes the massive amount of computation the modeling simulation requires. SLEUTH calibration process for the Tampa Bay urban growth simulation spends only 10 h CPU time. The model predicts future land use/cover change trends for Tampa Bay from 1992 to 2025. Urban extent is predicted to double in the Tampa Bay watershed between 1992 and 2025. Results show an upward trend of urbanization at the expense of a decline of 58% and 80% in agriculture and forested lands, respectively.

Shaped lattice graphs as models of the multiprocessor computer systems

A new family of the uniform ordinary graphs named the shaped lattice graphs was proposed and studied. As applied to the structural modeling of the multiprocessor computer systems, they have two useful characteristics such as high flexibility at choosing their size (number of vertices) and possibility of obtaining small (down to d = 2) diameters independently of the graph size. The method of cycle definition and transformation by the ring sequences of edge weights, which was previously proposed for the binary hypercubes and other Cayley graphs, was extended to the shaped lattice graphs.

Kinetic schemes and quasi-gas-dynamic system of equations

This paper deals with the description of kinetic schemes and the quasi-gas-dynamic system of equations, which is a promising trend for modelling complex gas dynamic flows. The possibilities of kinetic schemes and the quasi-gas-dynamic system as a model for describing a viscous gas flow are discussed. Particular attention is given to the possibility of their application to the computation of gas dynamics problems on high-performance multiprocessor computer systems. The possibilities to use the specified approaches for the solution of other problems of continuum mechanics are considered.

Minimized Embedding of Arbitrary Hamiltonian Graphs in Fault-tolerant Graph and Reconfiguration at Faults. I. One-fault-tolerant Structures

The paper proposed approaches to minimized embedding of the Hamiltonian graphs in the enveloping fault-tolerant graph representing the structural model of a fault-tolerant multiprocessor computer system. Failures are regarded as faults of vertices and/or connections between the graph vertices. Mathematical studies rely on the group-theoretical analysis of the characteristics of system structure. It underlies the proposed unique approach to designing the one-fault-tolerant and k-fault-tolerant structures retaining after reconfiguration the logical structure of the original target graph and, therefore, the compiled code of system tasks. The minimum fault-tolerant solutions were obtained for one-fault-tolerant and k-fault-tolerant cycles, simple and diagonal grids, and other popular structures, including arbitrary Hamiltonian graphs for which solutions are of minimized nature. Consideration was given to the algorithms of reconfiguration after arbitrary single and multiple faults. Restoration after faults is very simple; it is based on small tables of the group of system automorphisms which enable correct restoration of the system “at the level of theorems” without either static or dynamic additional verification of the reconfiguration process.

Rigid-body dynamics in the isothermal-isobaric ensemble: a test on the accuracy and computational efficiency.

We have developed a time-reversible rigid-body (rRB) molecular dynamics algorithm in the isothermal-isobaric (NPT) ensemble. The algorithm is an extension of rigid-body dynamics [Matubayasi and Nakahara, J Chem Phys 1999, 110, 3291] to the NPT ensemble on the basis of non-Hamiltonian statistical mechanics [Martyna, G. J. et al., J Chem Phys 1994, 101, 4177]. A series of MD simulations of water as well as fully hydrated lipid bilayer systems have been undertaken to investigate the accuracy and efficiency of the algorithm. The rRB algorithm was shown to be superior to the state-of-the-art constraint-dynamics algorithm SHAKE/RATTLE/ROLL, with respect to computational efficiency. However, it was revealed that both algorithms produced accurate trajectories of molecules in the NPT as well as NVT ensembles, as long as a reasonably short time step was used. A couple of multiple time-step (MTS) integration schemes were also examined. The advantage of the rRB algorithm for computational efficiency increased when the MD simulation was carried out using MTS on parallel processing computer systems; total computer time for MTS-MD of a lipid bilayer using 64 processors was reduced by about 40% using rRB instead of SHAKE/RATTLE/ROLL.

The GA-based approach to optimization of parallel calculations in large physics problems

Parallelization of computational algorithms in large physics problems for multiprocessor computer system containing thousand and more of processor elements requires special tools. In the present work, the approach based on using Genetic Algorithm is proposed. The experiment results and the influence of GA parameters on convergence of the method are given. The approach may be used for real-time parallelization.

Fresh Breeze

It is well-known that multiprocessor systems are vastly more difficult to program than systems that support sequential programming models. In a 1998 paper[11] this author argued that six important principles for supporting modular software construction are often violated by the architectures proposed for multiprocessor computer systems. The Fresh Breeze project concerns the architecture and design of a multiprocessor chip that can achieve superior performance while honoring these six principles.The envisioned multiprocessor chip will incorporate three ideas that are significant departures from mainstream thinking about multiprocessor architecture: (1) Simultaneous multithreading has been shown to have performance advantages relative to contemporary superscalar designs. This advantage can be exploited through use of a programming model that exposes parallelism in the form of multiple threads of computation. (2) The value of a shared address space is widely appreciated. Through the use of 64-bit pointers, the conventional distinction between "memory" and the file system can be abolished. This can provide a superior execution environment in support of program modularity and software reuse, as well as supporting multi-user data protection and security that is consistent with modular software structure. (3) No memory update; cycle-free heap. Data items are created, used, and released, but never modified once created. The allocation, release, and garbage collection of fixed-size chunks of memory will be implemented by efficient hardware mechanisms. A major benefit of this choice is that the multiprocessor cache coherence problem vanishes: any object retrieved from the memory system is immutable. In addition, it is easy to prevent the formation of pointer cycles, simplifying the design of memory management support.

Efficiency of the Methods of Scheduling Complex Sets of Tasks in Nonuniform Multiprocessor Computer Systems

In order to minimize in a nonuniform multiprocessor computer system the time of execution of each particular set of user-defined tasks having random execution times, the methodology of mathematical forecasting of the execution times of complex sets of interrelated tasks was used for the first time to examine the problem of formal estimation and choice of a criterion (discipline) from their predefined set for dynamic scheduling of tasks to the processors of nonuniform multiprocessor computer systems.

Real-time concurrency control in a multiprocessor environment

Although many high-performance computer systems are now multiprocessor-based, little work has been done in real-time concurrency control of transaction executions in a multiprocessor environment. Real-time concurrency control protocols designed for uniprocessor or distributed environments may not fit the needs of multiprocessor-based real-time database systems because of a lower concurrency degree of transaction executions and a larger number of priority inversions. This paper proposes the concept of a priority cap to bound the maximum number of priority inversions in multiprocessor-based real-time database systems to meet transaction deadlines. We also explore the concept of two-version data to increase the system concurrency level and to explore the abundant computing resources of multiprocessor computer systems. The capability of the proposed methodology is evaluated in a multiprocessor real-time database system under different workloads, database sizes and processor configurations. It is shown that the benefits of the priority cap in reducing the blocking time of urgent transactions are far greater than the losses involved in committing less urgent transactions. The idea of two-version data also greatly improves the system performance because of a much higher concurrency degree in the system.

Calculation of the Flow of a Weakly Compressible Viscous Gas Using a Multiprocessor Computer System

A method to calculate the flow of a weakly compressible gas on a multiprocessor computer system using the (α − β) iteration algorithm has been proposed. The problem of gas flow in a cavity with a moving cover has been considered as an example. The flow is described using Navier–Stokes equations in natural variables; MAC‐type spaced grids are employed. Calculations with the use of a monoprocessor version of the (α − β) algorithm and using its parallel implementation for the case of four processors have been carried out; the issue of the efficiency of multisequencing has been studied. Certain aspects of organization of remote access to the multiprocessor system have been discussed.

Dynamic on-line task scheduling on parallel processors

The problem of scheduling on-line a stream of multi-tasked jobs on a set of parallel processors is considered. The goal is to minimize the average sojourn time experienced by each individual job in the system. The arriving jobs are comprised of parallel applications, each consisting of multiple independent tasks that are instantaneously assigned to processor queues upon its arrival to the system. The processors independently and concurrently service the tasks in their local queues which are first-come-first-served (FCFS). A job can depart from the system only when all its tasks have been executed and reassembled into the original single-job structure; until then, completed tasks are placed in a reassembly queue. All tasks have i.i.d. memoryless exponential service times and any task can be processed by any processor. Migration of tasks between queues — after their initial allocation — is not allowed. This model captures the main features of multiprocessor computer systems executing parallel programs. The key scheduling issue is as follows. When some queue backlogs are small, an incoming job should spread its tasks to those lightly loaded queues in order to take advantage of the parallel processing gain and lower its processing delay. On the other hand, when all queues are fairly congested, the job should schedule all its tasks sequentially in a single queue to avoid excessive reassembly delay (due to backlog fluctuations) and lower its task synchronization delay. In this paper, the trade-off between these two objectives is quantified and it is shown that the optimal schedule’s structure is based on backlog thresholds. Moreover, an off-line mechanism for determining these thresholds is provided, and further characteristics of the optimal scheduling policy under special cases is discussed. Finally, asymptotics and approximations for systems comprised of a large number of processors are considered.

Performance evaluation for CC-NUMA multiprocessors using an OLTP workload

The CC-NUMA is a promising architecture for multiprocessor computer systems because of its scalability and ease of programming. However, little is known about the collective performance of the cache coherency protocol, the interconnection network, and the I/O subsystem of the CC-NUMA multiprocessor. In this paper, we analyzed the collective performance of various configurations of the CC-NUMA multiprocessor using commercial workload targeted for popular OLTP applications. The simulation results showed that the bottleneck on the ring, the switch-based tree, and the I/O subsystem could be identified and effectively removed by changing the configurations.

System Diagnosis of the Failed Modules and Connections in Hypercube Structures of the Multiprocessor Computer Systems

Combinations of failed components (modules and communication lines) for which the i>n-cube is not an i>n-singlefold diagnosable system were shown to be possible in the hypercube structures. A system of i>n chains of a certain structure constructed to check each of the i>N(= 2i>n) hypercube modules was used to transform the initial information about the state of the hypercube. A two-stage procedure for diagnosis of i>p(≤ i>n) failed components was substantiated. Its application to some failure situations in the vertex 3-cube was illustrated.