Message Passing Interface Standard Research Articles

Parallel edge-based solution of viscoplastic flows with the SUPG/PSPG formulation

A parallel edge-based solution of three dimensional viscoplastic flows governed by the steady Navier–Stokes equations is presented. The governing partial differential equations are discretized using the SUPG/PSPG stabilized finite element method on unstructured grids. The highly nonlinear algebraic system arising from the convective and material effects is solved by an inexact Newton-Krylov method. The locally linear Newton equations are solved by GMRES with nodal block diagonal preconditioner. Matrix-vector products within GMRES are computed edge-by-edge (EDE), diminishing flop counts and memory requirements. A comparison between EDE and element-by-element data structures is presented. The parallel computations were based in a message passing interface standard. Performance tests were carried out in representative three dimensional problems, the sudden expansion for power-law fluids and the flow of Bingham fluids in a lid-driven cavity. Results have shown that edge based schemes requires less CPU time and memory than element-based solutions.

An Approach to Parallel M × N Communication

High performance scientific applications are frequently multiphysics codes composed from single-physics programs, or have a functional decomposition based on physics as well as a domain decomposition for parallelism. In either case, dividing the application into independent components that can be developed and tested separately is useful. This requires a fast and efficient mechanism to share the large parallel data structures that are used in scientific applications. The “ M× Nproblem” is the transfer of data between two scientific parallel programs with different numbers of processes on each side. We present a solution to the M× Nproblem based on an I/O message passing interface (MPI). In this solution, the client application reads and writes data using the standard MPI I/O, but instead of going to hard drives the data are streamed in parallel between module programs working together on an integrated simulation. This approach builds on existing technology, emphasizes the easy migration of current applications, and simplifies the unit testing of the components while maintaining parallel high performance throughout the application. It also does not require a component to know how many processes another communicating component has.

A parallel FFT accelerated transient field-circuit simulator

A novel fast electromagnetic field-circuit simulator that permits the full-wave modeling of transients in nonlinear microwave circuits is proposed. This time-domain simulator is composed of two components: 1) a full-wave solver that models interactions of electromagnetic fields with conducting surfaces and finite dielectric volumes by solving time-domain surface and volume electric field integral equations, respectively, and 2) a circuit solver that models field interactions with lumped circuits, which are potentially active and nonlinear, by solving Kirchoff's equations through modified nodal analysis. These field and circuit analysis components are consistently interfaced and the resulting coupled set of nonlinear equations is evolved in time by a multidimensional Newton-Raphson scheme. The solution procedure is accelerated by allocating field- and circuit-related computations across the processors of a distributed-memory cluster, which communicate using the message-passing interface standard. Furthermore, the electromagnetic field solver, whose demand for computational resources far outpaces that of the circuit solver, is accelerated by a fast Fourier transform (FFT)-based algorithm, viz. the time-domain adaptive integral method. The resulting parallel FFT accelerated transient field-circuit simulator is applied to the analysis of various active and nonlinear microwave circuits, including power-combining arrays.

PARALIGN: rapid and sensitive sequence similarity searches powered by parallel computing technology

PARALIGN is a rapid and sensitive similarity search tool for the identification of distantly related sequences in both nucleotide and amino acid sequence databases. Two algorithms are implemented, accelerated Smith–Waterman and ParAlign. The ParAlign algorithm is similar to Smith–Waterman in sensitivity, while as quick as BLAST for protein searches. A form of parallel computing technology known as multimedia technology that is available in modern processors, but rarely used by other bioinformatics software, has been exploited to achieve the high speed. The software is also designed to run efficiently on computer clusters using the message-passing interface standard. A public search service powered by a large computer cluster has been set-up and is freely available at , where the major public databases can be searched. The software can also be downloaded free of charge for academic use.

Fault Tolerance in Message Passing Interface Programs

In this paper we examine the topic of writing fault-tolerant Message Passing Interface (MPI) applications. We discuss the meaning of fault tolerance in general and what the MPI Standard has to say about it. We survey several approaches to this problem, namely checkpointing, restructuring a class of standard MPI programs, modifying MPI semantics, and extending the MPI specification. We conclude that, within certain constraints, MPI can provide a useful context for writing application programs that exhibit significant degrees of fault tolerance.

PEGASUS 5: An Automated Preprocessor for Overset-Grid Computational Fluid Dynamics

An all new, automated version of the PEGASUS software has been developed and tested. PEGASUS provides the hole-cutting and connectivity information between overlapping grids and is used as the final part of the grid-generation process for overset-grid computational fluid dynamics approaches. The new PEGASUS code (Version 5) has many new features: automated hole cutting, a projection scheme for fixing small discretization errors in overset surfaces, more efficient interpolation search methods using an alternating digital tree and a stencil-jumping scheme, hole-size optimization based on adding additional layers of fringe points, and an automatic restart capability. The new code has also been parallelized using the message-passing interface standard. The parallelization performance provides efficient speedup of the execution time by an order of magnitude, and up to a factor of 30 for very large problems. The results of two example cases are presented: a three-element high-lift airfoil and a complete Boeing 777-200 aircraft in a high-lift landing configuration

Towards Efficient Execution of MPI Applications on the Grid: Porting and Optimization Issues

The message passing interface (MPI) is a standard used by many parallel scientific applications. It offers the advantage of a smoother migration path for porting applications from high performance computing systems to the Grid. In this paper Grid-enabled tools and libraries for developing MPI applications are presented. The first is MARMOT, a tool that checks the adherence of an application to the MPI standard. The second is PACX-MPI, an implementation of the MPI standard optimized for Grid environments. Besides the efficient development of the program, an optimal execution is of paramount importance for most scientific applications. We therefore discuss not only performance on the level of the MPI library, but also several application specific optimizations, e.g., for a sparse, parallel equation solver and an RNA folding code, like latency hiding, prefetching, caching and topology-aware algorithms.

Portable and Scalable Algorithm for Irregular All-to-All Communication

In irregular all-to-all communication, messages are exchanged between every pair of processors. The message sizes vary from processor to processor and are known only at run time. This is a fundamental communication primitive in parallelizing irregularly structured scientific computations. Our algorithm reduces the total number of message start-ups. It also reduces node contention by smoothing out the lengths of the messages communicated. As compared to the earlier approaches, our algorithm provides deterministic performance and also reduces the buffer space at the nodes during message passing. The performance of the algorithm is characterised using a simple communication model of high-performance computing (HPC) platforms. We show the implementation on T3D and SP2 using C and the message passing interface standard. These can be easily ported to other HPC platforms. The results show the effectiveness of the proposed technique as well as the interplay among the machine size, the variance in message length, and the network interface.

Portable and Scalable Algorithm for Irregular All-to-All Communication

In irregular all-to-all communication, messages are exchanged between every pair of processors. The message sizes vary from processor to processor and are known only at run time. This is a fundamental communication primitive in parallelizing irregularly structured scientific computations. Our algorithm reduces the total number of message start-ups. It also reduces node contention by smoothing out the lengths of the messages communicated. As compared to the earlier approaches, our algorithm provides deterministic performance and also reduces the buffer space at the nodes during message passing. The performance of the algorithm is characterised using a simple communication model of high-performance computing (HPC) platforms. We show the implementation on T3D and SP2 using C and the message passing interface standard. These can be easily ported to other HPC platforms. The results show the effectiveness of the proposed technique as well as the interplay among the machine size, the variance in message length, and the network interface.

ENIGMA: maximum-entropy method program package for huge systems

ENIGMA(Electron and Nuclear Image Generator by Max-ent Analysis) is a program package to evaluate three-dimensional electron and nuclear density from X-ray and neutron diffraction data by using the maximum-entropy method (MEM). Compared with the previous program packageMEED,ENIGMAsaves computing time and frees memory space at the same time by employing parallel data processing. The fast Fourier transformation (FFT) technique is also implemented. As a consequence of these improvements, the MEM analysis byENIGMAbecomes applicable to huge systems, such as proteins and polymers, when the phased structure factors are provided. The package is transferable to a wide variety of parallel computers, because it is written in Fortran 90 and a standard message-passing interface (MPI).

Development and Validation of a Massively Parallel Flow Solver for Turbomachinery Flows

The development and validation of the unsteady, three-dimensional, multiblock, parallel turbomachinery  ow solver TFLO is presented. The unsteady Reynolds-averaged Navier–Stokes equations are solved using a cellcentered discretizationon arbitrarymultiblockmeshes. The solutionprocedure is basedon efŽ cient explicit Runge– Kuttamethodswith several convergenceacceleration techniques such asmultigrid,implicitresidual smoothing,and local time stepping.The solver is parallelized usingdomaindecomposition,a single programmultipledata strategy, and the message passing interface standard. Details of the communication scheme and load balancing algorithms are discussed. A general and efŽ cient procedure for parallel interblade row interfacing is developed. The dual-time stepping technique is used to advance unsteady computations in time. The focus is on improving the parallel efŽ ciency and scalability of the  ow solver, as well as on its initial validationof steady-state calculations in multiblade row environment.The result of this careful implementation is a solverwith demonstrated scalability up to 1024processors. For validationand veriŽ cation purposes, results fromTFLO are comparedwith both existing experimental data and computational results from other computational  uid dynamics codes used in aircraft engine industry.

PARALLEL NEUROSYS: A system for the simulation of very large networks of biologically accurate neurons on parallel computers

We present a software package for the simulation of very large neuronal networks on parallel computers. The package can be run on any system with an implementation of the Message Passing Interface standard. We also present some example results for a simple neuronal model in networks of up to a quarter of a million neurons. The full software package as well as usage and installation guidelines can be found in [http://cs.usfca.edu/neurosys].

Object serialization for marshaling data in a Java interface to MPI

Several Java bindings to Message Passing Interface (MPI) software have been developed recently. Message buffers have usually been restricted to arrays with elements of primitive type. We discuss adoption of the Java object serialization model for marshaling general communication data in MPI-like APIs. This approach is compared with a Java transcription of the standard MPI derived datatype mechanism. We describe an implementation of the mpiJava interface to MPI that incorporates automatic object serialization. Benchmark results confirm that current JDK implementations of serialization are not fast enough for high performance messaging applications. Means of solving this problem are discussed, and benchmarks for greatly improved schemes are presented. Copyright © 2000 John Wiley & Sons, Ltd.

Flutter Computation of Turbomachinery Cascades Using a Parallel Unsteady Navier-Stokes Code

A quasi-three-dimensional multigrid Navier-Stokes solver on single- and multiple-passage domains is presented for solving unsteady flows around oscillating turbine and compressor blades. The conventional direct store method is used for applying the phase-shifted periodic boundary condition over a single blade passage. A parallel version of the solver using the message passing interface standard is developed for multiple-passage computations. In parallel multiple-passage computations, the phase-shifted periodic boundary condition is converted to a simple in-phase periodic condition. Euler and Navier-Stokes solutions are obtained for unsteady flows through an oscillating turbine cascade blade row with both the sequential and the parallel codes. It is found that the parallel code offers almost linear speedup with a slope close to 1 on multiple CPUs. In addition, significant improvement is achieved in convergence of the computation to a periodic unsteady state in the parallel multiple-passage computations due to the use of the in-phase periodic boundary conditions compared with that in the single-passage computations with phase-shifted periodic boundary conditions via the direct store method. The parallel Navier-Stokes code is also used to calculate the flow through an oscillating compressor cascade. Results are compared with experimental data and computations by other authors.

Thread-oriented Message-Passing interface

In this paper the limitations of the standard Message-Passing Interface (MPI) are analyzed, then the technique of multithreading is introduced and a stand-alone LWP (light weight process) extension to MPI is implemented. Experimental results show that system overhead is considerably diminished.

Massively parallel molecular dynamics simulations with EAM potentials

Molecular dynamics of cascades in pure iron and iron–copper alloys using embedded atom method type of interatomic potentials are presented. Reliable simulations of radiation damage at the atomic scale with high energy Primary Knocked Atoms (PKA) need systems with large numbers of particles and very long computational time. To perform the simulation in a reasonable amount of time high-performance computer systems such as massively parallel machines need to be used. This paper presents the parallelisation strategy applied to a serial classical Molecular Dynamics code: DYMOKA. The original sequential Fortran code CDCMD from the University of Connecticut was first improved algorithmically by applying a link cell method for the neighbour list construction of the Verlet list, resulting in a fully linear algorithm. The parallelisation strategy adopted is a multidimensional domain decomposition of the simulation box using a link cell method and a Verlet list method for each subdomain independently. The program paradigm is based on explicit message passing, and the standard Message-Passing Interface (MPI) was chosen in order to achieve portability. First measurements have demonstrated that the simulation of a system of 2.000.000 (750.000) atoms on 128 (32) processors costs 2.5 (10.) μs per atom per step. The current implementation has proven good scalability up to 32 processors on a NEC Cenju-3 machine. To study the effects of irradiation on copper segregation, simulations with up to 1.000.000 atoms in iron and iron–copper were performed with PKA energies up to 20 keV.

Early Applications in the Message-Passing Interface (Mpi)

We describe a number of early efforts to make use of the Message-Passing Interface (MPI) standard in appli cations, based on an informal survey conducted in May-June, 1994. Rather than a definitive statement of all MPI developmental work, this paper addresses the initial successes, progress, and impressions that appli cation developers have had with MPI, according to the responses received. We summarize the important as pects of each survey response, and draw conclusions about the spread of MPI into applications. An understanding of message passing and access to the MPI standard are prerequisites for appreciating this paper. Some background material is provided to ease this requirement.