Parallel Virtual Machine Research Articles

Parallel program debugging by specification

AbstractMost message passing parallel programs employ logical process topologies with regular characteristics to support their computation. Since process topologies define the relationship between processes, they present an excellent opportunity for debugging. The primary benefit is that process behaviours can be correlated, allowing expected behaviour to be abstracted and identified, and undesirable behaviour reported. However, topology support is inadequate in most message passing parallel programming environments, including the popular Message Passing Interface (MPI) and the Parallel Virtual Machine (PVM). Programmers are forced to implement topology support themselves, increasing the possibility of introducing errors.This paper proposes a trace‐ and topology‐based approach to parallel program debugging, driven by four distinct types of specifications. Trace specifications allow trace data from a variety of sources and message passing libraries to be interpreted in an abstract manner, and topology specifications address the lack of explicit topology knowledge, whilst also facilitating the construction of user‐consistent views of the debugging activity. Loop specifications express topology‐consistent patterns of expected trace events, allowing conformance testing of associated trace data, and error specifications specify undesirable event interactions, including mismatched message sizes and mismatched communication pairs. Both loop and error specifications are simplified by having knowledge of the actual topologies being debugged.The proposed debugging framework enables a wealth of potential debugging views and techniques. Copyright © 2004 John Wiley & Sons, Ltd.

Consistent Comparison of the Codes RELAP5/PARCS and TRAC-M/PARCS for the OECD MSLB Coupled Code Benchmark

A generalized interface module was developed for coupling any thermal-hydraulic code to any spatial kinetic code. In the design used here the thermal-hydraulic and spatial kinetic codes function as independent processes and communicate using the Parallel Virtual Machine software. This approach helps maximize flexibility while minimizing modifications to the respective codes. Using this interface, the U.S. Nuclear Regulatory Commission (NRC) three-dimensional neutron kinetic code, Purdue Advanced Reactor Core Simulator (PARCS), has been coupled to the NRC system analysis codes RELAP5 and Modernized Transient Reactor Analysis Code (TRAC-M). Consistent comparison of code results for the Organization for Economic Cooperation and Development/Nuclear Energy Agency main steam line break benchmark problem using RELAP5/PARCS and TRAC-M/PARCS was made to assess code performance.

Research on Magnetic Abrasive Finishing for the Free Form Surface of Mould by 5-DOF Parallel Virtual Axis Machine Tool

Research on Magnetic Abrasive Finishing for the Free Form Surface of Mould by 5-DOF Parallel Virtual Axis Machine Tool

Solving Linear Programming Problems on the Parallel Virtual Machine Environment

This study developed a parallel algorithm to efficiently solve linear programming models. The proposed algorithm utilizes the Dantzig-Wolfe Decomposition Principle and can be easily implemented in a general distributed computing environment. The analytical performance of the well-known method, including the speedup upper bound and lower bound limits, was derived. Numerical experiments are also provided in order to verify the complexity of the proposed algorithm. The empirical results demonstrate that the speedup of this parallel algorithm approaches linearity, which means that it can take full advantage of the distributed computing power as the size of the problem increases.

Developing SPPVM modules with visual basic

The development of true Parallel Machines of which literatures are scanty is still algorithmitic. Since the need for increasing power in computations is the wish of every programmer or designer, PVMs become more of a choice. PVMs enable large computational problems to be solved more cost effectively by using aggregated power and memory of many computers. This paper develops a SPPVM (Single Processor Parallel Virtual Machine) with Visual Basic. It employs the VB Shell command to cause a single program or instruction to execute in two different shells in same memory space. A quick sort program is implemented on two shells. Each shell sorts 500 elements. The machine used for the execution is: Intel (r) Celeron ™ Processor – Genuine Intel ~ 600 Mhz, with Total Physical Mem. = 30.45MB, Available Physical Mem. = 232KB, Total Virtual Mem. = 2GB, Available Virtual Mem. = 1.90GB, Page File space = 1.97GB at runtime. Result comparison with a strictly sequential version reveals that the SPPVM executes in less time than ½ x Tseq (thetheoretical value for two processes in parallel). The key benefit of this paper is to enable programmers explore parallel programming features on micro systems and develop their own SPPVM. Keywords: Shell function, Self-diagnostic, Multi processing, Multi-Platform, Parallel Processing. [Global Jnl Mathematical Sci Vol.2(1) 2003: 29-36]

Parallelization of a Transient Method of Lines Navier–Stokes Code

Parallel implementation of a serial code, namely method of lines (MOL) solution for momentum equations (MOLS4ME), previously developed for the solution of transient Navier–Stokes equations for incompressible separated internal flows in regular and complex geometries, is described. *E-mail: ceme@metu.edu.tr The serial code was parallelized by using a domain decomposition strategy with overlapped boundary at the interfaces for information exchange between the sub-domains. A parallel virtual machine (PVM) and dual-processor personal computer (PC) connected over a switch are employed as the message passing software and the parallel computing platform, respectively. Performance of the parallelization strategy was first tested on a transient 1-D laminar pipe flow problem for MOL, explicit and implicit finite difference methods. Assessment of the performance of the parallel code (PARMOLS4ME) was then tested by applying it to two test cases; (i) laminar pipe flow with sudden expansion and (ii) turbulent flow in a gas turbine combustor simulator (GTCS) and comparing the results of serial and parallel codes. Comparisons show that the flow fields predicted by parallel codes agree well with those of serial codes at considerably lower execution times. Effects of load balancing are also investigated and it is seen that the load balancing has a significant effect on the execution time of the parallel code.

Parallel Large Scale High Accuracy Navier-Stokes Computations on Distributed Memory Clusters

We present a highly scalable parallelization of a high-accuracy 3D serial multiblock Navier-Stokes solver. The code solves the full Navier-Stokes equations and is capable of performing large-scale computations for practical configurations in an industrial enviroment. The parallelization strategy is based on the geometrical domain decomposition principle, and on the overlapped communication and computation concept. The important advantage of the strategy is that the suggested type of message-passing ensures a very high scalability of the algorithm from the network point of view, because, on the average, the communication work per processor is not increased if the number of processors is increased. The parallel multiblock-structured Navier-Stokes solver based on the parallel virtual machine (PVM) routines was implemented on 106-processors distributed memory cluster managed by the MOSIX software package. Analysis of the results demonstrated a high level of parallel efficiency (speed up) of the computational algorithm. This allowed the reduction of the execution time for large-scale computations employing 10 million of grid points, from an estimated 46 days on the SGI ORIGIN 2000 computer (in the serial single-user mode) to 5–6 hours on 106-processors cluster. Thus, the parallel multiblock full Navier-Stokes code can be successfully used for large-scale practical aerodynamic simulations of a complete aircraft on millions-points grids on a daily basis, as needed in industry.

PHASE CENTRE OPTIMIZATION IN PROFILED CORRUGATED CIRCULAR HORNS WITH PARALLEL GENETIC ALGORITHMS

Achieving a high stability of the phase centre position in horn antennas with respect to frequency is a very desirable aim in reflector antenna design; a highly stable phase centre reduces efficiency dropping for defocusing at the frequency band extremes. By using an appropriate profile for the horn antenna it is possible to obtain horns both compact and with a stable phase centre. In this paper an automatic design procedure, based on Genetic Algorithms, to obtain such horns is described. The algorithm operates on many horn profile parameters, including corrugations, and is based on an accurate fullwave mode matching/combined field integral equation analysis code. To keep computing time down a full parallel algorithm over a 12 CPU parallel virtual machine is described.

To a Performance Evaluation of Parallel Algorithms in Now

A recent trend in high performance computing (HPC) is to use networks of workstations (NOW) as a cheaper alternative to massively parallel multiprocessors or supercomputers. In such parallel systems (NOW's) individual workstations are connected through widely used communication standard networks and co-operate to solve one large problem. Every workstation is treated similarly as a processing element in a conventional multiprocessor system. To make the whole system appear to the applications as a single parallel computing engine (a virtual parallel system), run-time environments such as PVM (Parallel virtual machine), MPI (Message passing interfaces) are often used to provide an extra layer of abstraction. In this paper, we discuss a new performance evaluation method on the example of multidimensional DFFT (Discrete Fast Fourier Transform) in a NOW's based on Intel's personal computers.

The design and implementation of a runtime system for graph-oriented parallel and distributed programming

Graph has been widely used in modeling, specification, and design of parallel and distributed systems. Many parallel and distributed programs can be expressed as a collection of parallel functional modules whose relationships can be defined by a graph. Often, the basic functions of communication and coordination of the parallel modules are expressed in terms of the underlying graph. Furthermore, parallel/distributed graph algorithms are used to realize various control functions. To facilitate the implementation of these algorithms, it is desirable to have an integrated approach that provides direct support for efficient operations on graphs. We have proposed a graph-oriented programming model, called GOP, which aims at providing high-level abstractions for configuring and programming cooperative parallel processes. GOP enables the programmer to configure the logical structure of a distributed program by using a logical graph and to write the program using communications and synchronization primitives based on the logical structure. In this paper, we describe the design and implementation of a portable run-time system for the GOP framework. The runtime system provides an interface with a library of programming primitives to the low-level facilities required to support graph-oriented communications and synchronization. The implementation is on top of the parallel virtual machine in a local area network of Sun workstations. We focus our discussion on the following four aspects: the software architecture, including the structure of runtime system and interfaces between user programs and the runtime kernel; graph representation; implementation of graph operations; and performance of the run-time in terms of the implementation of graph-oriented communications.

Simulations of the low-dimensional magnetic systems by the quantum transfer-matrix technique

A numerical quantum transfer-matrix approach based on the Suzuki–Trotter formula and the checker-board decomposition is presented in the framework of quantum statistical mechanics. It is applied to the supramolecular cluster Mn 6 (i.e. [Mn(hfac) 2NITPh] 6) and to a number of the macroscopic quasi-one-dimensional magnets. The latter include: (i) the macroscopic Haldane-gap spin S=1 chains and molecular-based magnetic spin S=1 chains (with uniform and alternating interaction couplings); (ii) the spin-Peierls CuGeO 3 and Pb[Cu(SO 4)(OH) 2] compounds subject to frustration; (iii) Yb 4As 3 which is a new semimetallic material being the first example of the antiferromagnetic S=1/2 spin chain with the induced staggered field appearing as a result of the antisymmetric Dzyaloshinsky–Moriya interaction. The compounds in question can be characterized within the spin Heisenberg models and their thermodynamic properties at finite temperatures are calculated using our own codes. Our simulation results are compared with the available experimental results and a quantitative agreement has been established. The large-scale numerical simulations were carried out on the Cray and Silicon Graphics supercomputers, using the parallelized and vectorized codes, exploiting the Parallel Virtual Machine (PVM) and the Message Passing Interface (MPI) system libraries.

Implementation of generalized cross validation based image denoising in parallel virtual machine environment

Paper presents a wavelet based, high performance image denoising scheme implemented in a parallel virtual machine (PVM) environment. Performance of generalized cross validation (GCV) based threshold selection scheme is poorer from CPU time viewpoint when implemented sequentially. In contrast to the traditional parallel approaches, which rely on specialized parallel machines, the present work explores the potential of distributed systems for parallelism. The master/slave model is adopted for control of machines. In this paper the performance of GCV based thresholding algorithm is presented in PVM environment. The proposed algorithm outperforms Donoho and Johnstone's Universal and SureSrink thresholding scheme in most of the test cases. The present investigation validates suitability of GCV thresholding scheme from parallel implementation viewpoint.

On the nuclear response of the water-cooled Pb–17Li test blanket module for ITER-FEAT

Within the European Fusion Technology Programme, the Water-Cooled Lithium Lead (WCLL) DEMO breeding blanket line was selected in 1995 as one of the two EU lines to be developed in the next decades, in particular with the aim of manufacturing a Test Blanket Module (TBM) to be tested in ITER-FEAT. The present paper is focused on the study of the WCLL-TBM nuclear response in ITER-FEAT, being specifically oriented to the investigation of the local effects due to the typical C-shaped tubes of the breeder zone, since they could play a pivotal role in the module-relevant thermo–mechanical design. A 3D heterogeneous model of the WCLL-TBM, realistically simulating its new lay out and taking into account 9% Cr martensitic steel as reference structural material, has been set-up. A particular attention has been paid to the simulation of the characteristic ‘C’ shape of the breeder zone double walled tubes, which have been realistically reproduced. The WCLL-TBM model has been inserted into an existing ITER-FEAT 3D semi-heterogeneous model accounting for a proper D–T neutron source. Analyses have been performed by means of MCNP-4C code running on a cluster of four workstations through the implementation of a parallel virtual machine. For each analysis a large number of histories (>10.000.000) have been simulated, obtaining statistical uncertainties on the results lower than 3%. The main features of the WCLL-TBM nuclear response have been determined focusing the attention on power deposition density, material damage through DPA and He and H production rate, daily tritium production and tritium production rate radial distribution in the module. The obtained results are herewith presented and critically discussed.

Three Mile Island Unit 1 Main Steam Line Break Three-Dimensional Neutronics/Thermal-Hydraulics Analysis: Application of Different Coupled Codes

A comprehensive analysis of the double ended main steam line break (MSLB) accident assumed to occur in the Babcock & Wilcox Three Mile Island Unit 1 (TMI-1) has been carried out at the Dipartimento di Ingegneria Meccanica, Nucleare e della Produzione of the University of Pisa, Italy, in cooperation with the University of Zagreb, Croatia. The overall activity has been completed within the framework of the participation in the Organization for Economic Cooperation and Development-Committee on the Safety of Nuclear Installations-Nuclear Science Committee pressurized water reactor MSLB benchmark.Thermal-hydraulic system codes (various versions of Relap5), three-dimensional (3-D) neutronics codes (Parcs, Quabbox, and Nestle), and one subchannel code (Cobra) have been adopted for the analysis. Results from the following codes (or code versions) are assumed as reference:1. Relap5/mod3.2.2, beta version, coupled with the 3-D neutron kinetics Parcs code parallel virtual machine (PVM) coupling2. Relap5/mod3.2.2, gamma version, coupled with the 3-D neutron kinetics Quabbox code (direct coupling)3. Relap5/3D code coupled with the 3-D neutron kinetics Nestle code.The influence of PVM and of direct coupling is also discussed.Boundary and initial conditions of the system, including those relevant to the fuel status, have been supplied by Pennsylvania State University in cooperation with GPU Nuclear Corporation (the utility, owner of TMI) and the U.S. Nuclear Regulatory Commission. The comparison among the results obtained by adopting the same thermal-hydraulic nodalization and the coupled code version is discussed in this paper.The capability of the control rods to recover the accident has been demonstrated in all the cases as well as the capability of all the codes to predict the time evolution of the assigned transient. However, one stuck control rod caused some “recriticality” or “return to power” whose magnitude is largely affected by boundary and initial conditions.

Special Issue: New parallel algorithm and its speed-up for the explicit difference method

New parallel “ring” algorithm for solution of a spatially one dimensional initial-boundary-value problem (IBVP) for a parabolic equation using an explicit difference method in this article is suggested. The parallel algorithm has been verified by implementation on a workstation-cluster running under parallel virtual machine (PVM). A speed-up function is determined as a ratio of time needed for realization of the algorithm in sequential case to time in parallel. Theoretical estimates of the speed-up function show the significant speed-up of the parallel algorithm in comparison with the serial one.

Periodic steady state solution of electric systems with nonlinear components using parallel processing

This paper introduces a very efficient methodology for obtaining the periodic steady state solution of electric systems with nonlinear components. It combines the benefits of the fast time domain solution using Newton techniques for the acceleration of the convergence of state variables to the limit cycle (LC) with the application of parallel processing based on the parallel virtual machine (PVM) platform.

A parallel implementation of nec for the analysis of large structures

We present a new, parallel version of the numerical electromagnetics code (NEC). The parallelization is based on a bidimensional block-cyclic distribution of matrices on a rectangular processor grid, assuring a theoretically optimal load balance among the processors. The code is portable to any platform supporting message passing parallel environments such as message passing interface and parallel virtual machine, where it could even be executed on heterogeneous clusters of computers running on different operating systems. The developed parallel NEC was successfully implemented on two parallel supercomputers featuring different architectures to test portability. Large structures containing up to 24000 segments, which exceeds currently available computer resources were successfully executed and timing and memory results are presented. The code is applied to analyze the penetration of electromagnetic fields inside a vehicle. The computed results are validated using other numerical methods and experimental data obtained using a simplified model of a vehicle (consisting essentially of the body shell) illuminated by an electromagnetic pulse (EMP) simulator.

Pp-Blast: a "pseudo-parallel" Blast.

We have developed a software called pp-Blast that uses the publicly available Blast package and PVM (parallel virtual machine) to partition a multi-sequence query across a set of nodes with replicated or shared databases. Benchmark tests show that pp-Blast running in a cluster of 14 PCs outperformed conventional Blast running in large servers. In addition, using pp-Blast and the cluster we were able to map all human cDNAs onto the draft of the human genome in less than 6 days. We propose here that the cost/benefit ratio of pp-Blast makes it appropriate for large-scale sequence analysis. The source code and configuration files for pp-Blast are available at http://www.ludwig.org.br/biocomp/tools/pp-blast.

A framework of Web-based Decision Support Systems for portfolio selection with OLAP and PVM

In this paper, we provide an integrated framework for portfolio selection, which is adaptable to the needs of financial organizations and individual investors, and as an organized approach of selecting efficient portfolios for investments. We focus our discussion on the implementation of this framework for a Web-based Decision Support System (DSS) based on our prototype named WPSS—A Web-based Portfolio Selection System for Chinese financial markets. In this system, we adopt technologies such as online analytical processing as an add-on tool for analytical purpose, as well as using Parallel Virtual Machine (PVM) to improve overall performance.

Negative wake in the uniform flow past a cylinder

The upstream/downstream streamline shift and the associated negative wake generation (streamwise velocity overshoot in the wake) in a viscoelastic flow past a cylinder are studied in this paper, for the Oldroyd-B, UCM, PTT, and FENE-CR fluids, using the Discrete Elastic Viscous Split Stress Vorticity (DEVSS-ω) scheme (Dou HS, Phan-Thien N (1999). The flow of an Oldroyd-B fluid past a cylinder in a channel: adaptive viscosity vorticity (DAVSS-ω) formulation. J Non-Newtonian Fluid Mech 87:47–73). The numerical algorithm is a parallelized unstructured Finite Volume Method (FVM), running under a distributed computing environment through the Parallel Virtual Machine (PVM) library. It is demonstrated that both the normal stress and its gradient are responsible for the negative wake generation and streamline shifting. Fluid extensional rheology plays an important role in the generation of the negative wake. The negative wake can occur in flows where the fluid extensional viscosity does not increase rapidly with strain rate. The formation of the negative wake does not depend on whether the streamlines undergo an upstream or a downstream shift. Shear-thinning viscosity weakens the velocity overshoot and while shear-thinning first normal stress coefficient enhances the velocity overshoot. Wall proximity is not necessary for the velocity overshoot; however, it enhances the strength of the negative wake. For the Oldroyd-B fluid, the ratio of the solvent viscosity to the zero-shear viscosity plays an important role in the streamline shift. In addition, mesh dependent behaviour of normal stresses along the centreline at high De in most cylinder/sphere simulations is due to the convection of normal stress from the cylinder to the wake, which results in the maximum of the normal stress being located off the centreline by a short distance at high De.