Bin Cheng, Xingang Wang, Ying Li et al. · 2009
The performance analysis of parallel system is critical for tuning, optimizing and algorithm modifying of parallel program. A formal modeling method based on TCPN is proposed in this paper, which makes it possible to organize a complicated parallel system as a set of hierarchical modules and includes a time concept to represent the time taken to execute events and further analyze the performance of the system. In view of the fact that the performance of parallel systems is not only determined by the performance of the hardware itself, but also by the structure of the parallel program and the assignment of program parts to resources, TCPN describes the parallel program and the resources respectively and brings less effect to modify the program structure because of running environment changes.
Cited by 0
F. Ino, K. Hagihara · 2005
This paper describes the design and implementation of Gordini, a performance analysis tool that is capable of automatically locating places in the source code where a communication optimization technique can be applied for performance debugging of message passing parallel programs. Our automatic search approach is based on data dependence analysis on trace files. It currently supports three techniques: communication-computation overlap, message aggregation, and collective communication. In case studies, Gordini assists us in improving the performance of sophisticated programs coded by experts, the two first prize winners of software contests. Therefore, we believe that our automatic approach is useful for application developers to locate communication bottlenecks in programs.
Cited by 0
Dong-Won Kim, Insik Yoon · 2002
This paper investigates the execution behaviors of parallel sorting algorithms on an experimental multiprocessor (KuPP) and compares with predicted performance under LogP and BSP (Bulk Synchronous Parallel) models. Since the communication overhead is considered a primary candidate for improvement, a few schemes are devised and experimented on KuPP to reduce the time spent in communication, thus to enhance the overall performance. The authors believe the ideas can be adopted in other high-performance parallel computers.
Cited by 1
久保田 和人, 板倉 憲一, 佐藤 三久 et al. · Terrestrial Environment Research Center (University of Tsukuba) · 1999
journal article
Cited by 0
Allen Malony, Bernd Mohr, Pete Beckman et al. · 2002
pC++ is a language extension to C++ designed to allow programmers to compose distributed data structures with parallel execution semantics. These data structures are organized as "concurrent aggregate" collection classes which can be aligned and distributed over the memory hierarchy of a parallel machine in a manner consistent with the High Performance Fortran Forum (HPF) directives for Fortran 90. pC++ allows the user to write portable and efficient code which will run on a wide range of scalable parallel computers. We discuss the performance analysis of the pC++ programming system. We describe the performance tools developed and include scalability measurements for four benchmark programs: a "nearest neighbor" grid computation, a fast Poisson solver, and the "Embar" and "Sparse" codes from the NAS suite. In addition to speedup numbers, we present a detailed analysis highlighting performance issues at the language, runtime system, and target system levels.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>
Cited by 24
Tomoo Inoue, Takaharu Fujii, Hideo Fujiwara · 2002
The performance of parallel processing for test generation depends on the method of communication among processors. This paper presents two types of parallel processing which differ in communication methods, and analyzes their performance. We formulate the number of test vectors obtained in parallel processing, and analyze the costs of test generation, fault simulation and interprocessor communication and the speedup ratio. Further, we consider a method which minimizes the cost of interprocessor communication.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>
Cited by 2
Joachim Beer · Lecture notes in computer science · 1989
Cited by 20
Eric R. Marsh, R. Ryan Vallance · 2006
Cited by 0
Geimer, M. · JuSER (Forschungszentrum Jülich) · 2006
Cited by 0
Souto, Roberto Pinto, Campos Velho, Haroldo Fraga de, Stephany, Stephan et al. · Biblioteca Digital da Memória Científica do INPE (National Institute for Space Research) · 2003
Three algorithms for solving the radiative transfer equation (RTE) were studied: Hydrolight, PEESNA and LTSN. These algorithms correspond, respectively, to invariant imbedding, analytical discrete-ordinates and LTSN methods. As a first step, the performance of each algorithm was evaluated running in the same sequential machine. The related codes were used in a Hydrological Optics typical coastal water test case, in order to calculate the surface-emergent radiation intensities (radiances) given the incident radiances and inherent optical properties such as the absorption and the scattering coefficients. Timing and profiling of the three codes was performed in order to evaluate processing times and to identify performance bottlenecks. Next, each algorithm was studied concerning the feasibility of its parallelization using the MPI message passing communication library and execution in a distributed memory machine, a multicomputer based on IA-32 architecture. The three codes perform spatial discretization of the domain and Fourier decomposition of the radiances obtaining independent azimuthal modes. Therefore, an independent RTE can be written for each azimuthal mode and can be assigned to a different processor, in a parallel implementation. The speed-up that can be achieved increases with the fraction of time spent in the azimuthal mode, but total execution time is also an important issue. Results are discussed and further strategies are proposed.
Cited by 1Open access
Gumaste, U., Lanteri, S., Farhat, C. et al. · NASA Technical Reports Server (NASA)
In order to predict the dynamic response of a flexible structure in a fluid flow, the equations of motion of the structure and the fluid must be solved simultaneously. In this paper, we present several partitioned procedures for time-integrating this focus coupled problem and discuss their merits in terms of accuracy, stability, heterogeneous computing, I/O transfers, subcycling, and parallel processing. All theoretical results are derived for a one-dimensional piston model problem with a compressible flow, because the complete three-dimensional aeroelastic problem is difficult to analyze mathematically. However, the insight gained from the analysis of the coupled piston problem and the conclusions drawn from its numerical investigation are confirmed with the numerical simulation of the two-dimensional transient aeroelastic response of a flexible panel in a transonic nonlinear Euler flow regime.
Cited by 0
Jürgen Knopp · Lecture notes in computer science · 1992
Cited by 3Open access
Jordi Guitart, Jordi Torres, Eduard Ayguade et al. · 2001
In this paper we describe an instrumentation environment for the performance analysis and visualization of parallel applications written in JOMP, an OpenMP-like interface for Java. The environment includes two complementary approaches. The first one has been designed to provide a detailed analysis of the parallel behavior at the JOMP programming model level. At this level, the user is faced with parallel, work-sharing and synchronization constructs, which are the core of JOMP. The second mechanism has been designed to support an in-depth analysis of the threaded execution inside the Java Virtual Machine (JVM). At this level of analysis, the user is faced with the supporting threads layer, monitors and conditional variables. The paper discusses the implementation of both mechanisms and evaluates the overhead incurred by them.
Cited by 0
Ronaghi, M., Gumaste, U., Felippa, C. A. et al. · NASA Technical Reports Server (NASA)
Applications are described of high-performance parallel, computation for the analysis of complete jet engines, considering its multi-discipline coupled problem. The coupled problem involves interaction of structures with gas dynamics, heat conduction and heat transfer in aircraft engines. The methodology issues addressed include: consistent discrete formulation of coupled problems with emphasis on coupling phenomena; effect of partitioning strategies, augmentation and temporal solution procedures; sensitivity of response to problem parameters; and methods for interfacing multiscale discretizations in different single fields. The computer implementation issues addressed include: parallel treatment of coupled systems; domain decomposition and mesh partitioning strategies; data representation in object-oriented form and mapping to hardware driven representation, and tradeoff studies between partitioning schemes and fully coupled treatment.
Cited by 0Open access
Gerndt, M., Mohr, B., Miller, T. W. R. · JuSER (Forschungszentrum Jülich) · 2000
Cited by 1
Hunsicker,F. L. · Defense Technical Information Center (DTIC) · 1964
A solid-state sonar transmitter using silicon-controlled rectifiers in a parallel inverter circuit has been developed. The transmitter has an output power of three kilowatts, with an efficiency of 85 to 90 percent at low audio frequencies. An ac equivalent circuit for the parallel inverter and load combination has been analyzed to obtain equations for inverter input resistance, commutation ratio, and peak voltage. The equivalent circuit assumes that the inverter components are lossless and that the input current is constant. The transmitter was used to drive a 36-element array of variable-reluctance transducers which have a resonant frequency of 960 cycles per second. The distorted driving voltage from the inverter circuit did not adversely affect the acoustic performance of the projector array. Comparison of the measured and calculated performance of the inverter circuit indicates that the ac model gives sufficiently accurate results to allow prediction of inverter performance with a known load. (Author)
Cited by 0
Chen, P.-S., Farhat, C., Lesoinne, M. et al. · NASA Technical Reports Server (NASA)
Applications are described of high-performance computing methods to the numerical simulation of complete jet engines. The methodology focuses on the partitioned analysis of the interaction of the gas flow with a flexible structure and with the fluid mesh motion driven by structural displacements. The latter is treated by a ALE technique that models the fluid mesh motion as that of a fictitious mechanical network laid along the edges of near-field elements. New partitioned analysis procedures to treat this coupled three-component problem were developed. These procedures involved delayed corrections and subcycling, and have been successfully tested on several massively parallel computers, including the iPSC-860, Paragon XP/S and the IBM SP2. The NASA-sponsored ENG10 program was used for the global steady state analysis of the whole engine. This program uses a regular FV-multiblock-grid discretization in conjunction with circumferential averaging to include effects of blade forces, loss, combustor heat addition, blockage, bleeds and convective mixing. A load-balancing preprocessor for parallel versions of ENG10 was developed as well as the capability for the first full 3D aeroelastic simulation of a multirow engine stage. This capability was tested on the IBM SP2 parallel supercomputer at NASA Ames.
Cited by 0Open access
Wolf, F. · JuSER (Forschungszentrum Jülich) · 2011
Cited by 0
M. Varalakshmi, Daphne Lopez · Research Journal of Applied Sciences Engineering and Technology · 2015
This study aims at making a comparative study of various parallel programming models for a compute intensive application pertaining to Atmospheric modeling. Atmospheric modeling deals with predicting the behavior of atmosphere through mathematical equations governing the atmospheric fluid flows. The mathematical equations are nonlinear partial differential equations which are difficult to solve analytically. Thus fundamental governing equations of atmospheric motion are discretized into algebraic forms that are solved using numerical methods to obtain flow-field values at discrete points in time and/or space. Solving these equations often requires huge computational resource, which is normally available with high-speed supercomputers. Shallow Water equations provide a useful framework for the analysis of dynamics of large-scale atmospheric flow and for the analysis of various numerical methods that might be applied to the solution of these equations. In this study, Finite volume approach has been used for discretizing these equations that leads to a number of algebraic equations equal to the number of time instants at which the flow field values are to be evaluated. It is apparent that the application is embarrassingly parallel and its parallelization will suppress communication overhead. A High Performance Compute cluster has been employed for solving the equations involved in atmospheric modeling. Use of OpenMP and MPI APIs has paved the way to study the behavior of shared memory programming model and the message passing programming model in the context of such a highly compute intensive application. It is observed that no additional benefit can be enjoyed by creating too many software threads than the available hardware threads, as the execution resources should be shared among them.
Cited by 0Open access
Charles J Archer · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 2023
An apparatus, program product and method optimize the operation of a massively parallel computer system by, in part, receiving actual performance data concerning an application executed by the plurality of interconnected nodes, and analyzing the actual performance data to identify an actual performance pattern. A desired performance pattern may be determined for the application, and an algorithm may be selected from among a plurality of algorithms stored within a memory, the algorithm being configured to achieve the desired performance pattern based on the actual performance data.
Cited by 0Open access