Concurrent Architecture Research Articles

An object‐oriented tool for simulating distributed real‐time control systems

AbstractThis paper presents an object‐oriented software tool, called RTSIM, aimed at simulating real‐time embedded controllers. The tool consists of a collection of C++ libraries permitting a separate specification of the functional behaviour of the controller and of the hardware/software architecture to be used for its deployment. In particular, it is possible to provide an accurate modelling of the concurrent architecture of the control tasks and of the run‐time support offered by the operating system for the real‐time scheduling of the shared resources (CPU, memory buffers and network links). In this way, it is possible to compare different scheduling solutions by evaluating their simulated performance directly in the domain of the control application. Moreover, the tool can be utilized to tune up design parameters such as the activation frequencies of the tasks. The application of the tool is shown in a meaningful case study. Copyright © 2002 John Wiley & Sons, Ltd.

A tool for simulation and fast prototyping of embedded control systems

This paper presents a set of C++ libraries, called RTSIM, aimed at realizing a joint simulation of a continuous plant and of a real-time embedded controller. The libraries permit a separate specification of the functional behaviour of the controller and of the software platform to be used for its deployment. In particular, it is possible to provide an accurate modeling of the concurrent architecture of the control tasks and of the run-time support offered by the operating system for the real-time scheduling of the shared resources (CPU, memory buffers, network links). In this way, its is possible to compare different scheduling solutions by evaluating their simulated performance in the domain of the control application. Moreover, the tool can be utilized to tune up such design parameters as the activation frequencies of the tasks. The application of the tool is shown on a meaningful robotic case-study.

Supporting server-level fault tolerance in concurrent-push-based parallel video servers

Parallel video servers have been proposed for building large-scale video-on-demand (VoD) systems from multiple low-cost servers. However, when adding more servers to scale up the capacity, system-level reliability will decrease as failure of any one of the servers will cripple the entire system. To tackle this reliability problem, this paper proposes and analyzes architectures to support server-level fault tolerance in parallel video servers. Based on the concurrent push architecture proposed earlier, this paper tackles three problems pertaining to fault tolerance, namely redundancy management, redundant data transmission protocol, and real-time fault masking. First, redundant data based on erasure codes are introduced to video data stored in the servers, which are then delivered to the clients to support fault tolerance. Despite the success of distributed redundancy striping schemes such as RAID-5 in disk array implementations, we discover that similar schemes extended to the server context do not scale well. Instead, we propose a redundant server scheme that is both scalable, and with lower total server buffer requirement. Second, two protocols are proposed to manage the transmission of redundant data to the clients, namely forward erasure correction which always transmits redundant data, and on-demand correction which transmits redundant data only after a server failure is detected. Third, to enable ongoing video sessions to maintain nonstop video playback during failure, we propose using fault masking at the client to recompute lost video data in real-time. In particular we derive the amount of client buffer required so that nonstop, continuous video playback can be maintained despite server failures.

Computational granularity and parallel models to scale up reactive scattering calculations

The computational granularity is a key factor in determining the suitability of a parallel model for an efficient implementation of complex codes on concurrent processor architectures. The paper discusses these aspects with reference to various theoretical approaches to reactive scattering and to the scale up of related calculations.

Concurrency State Models and Java Programs

Concurrency State Models and Java Programs

Applying object-oriented database technologies in concurrent design processes

A concurrent design can be used to improve productivity and quality by means of shortening product development time and sharing design information. In order to share information among designers, an object-oriented database management system, which has shown promising capabilities in engineering data processing, is used as a foundation for building a concurrent design architecture in this research. The architecture allows multiple designers to work on the design concurrently, store multiple versions of the design, integrate design information, and evolve the design structure. A procedure of plastic filling injection moulding is used to demonstrate the process of concurrent design, and is compared with the process of C-MOLD, a simulation analysis software package for plastic injection moulding, to show the superiority of the proposed concurrent design architecture.

Simulating free-space optical computing architectures

Major issues in optoelectronic system design include timing, synchronization, and control. Designing free-space optical computing architectures is difficult because of the high degree of system complexity, parallelism, and concurrency in conjunction with the high cost and lack of availability of devices. Current simulation tools lack the expressiveness to model the system structure and behavior of parallel and concurrent architectures, thus making them inefficient and ineffective. We show that Petri nets, compared with other system-modeling methodologies, are more efficient and effective at expressing the functional, behavioral, and structural properties of parallel and concurrent architectures. We show how an extended version of the standard Petri net, a timed-colored Petri net, is used to model and simulate free-space optoelectronic computing architectures. We also present methods for analysis of system timing, synchronization, and control behavior.

Expert-Coop: An Environment for Cognitive Multi-Agent Systems

Abstract A modular object-oriented environment, called Expert-Coop, designed to help the implementation of open heterogeneous cognitive Multi-Agent Systems is presented in this paper. The environment consists of a library of classes that supports the creation of agents according to a predefined concurrent agent architecture. The environment is written in Common Lisp/CLOS and its modularity, guaranteed by the object-oriented design methodology, makes it easy to modify and augment it with new features. The environment is being used to develop an application concerning the restoration of an electricity transmission network, covering Brazil's south region, after a blackout has affected part of the network.

Reflection and time-dependent computing: Experiences with the R2 architecture

In this paper we present an application of computational reflection in the programming oftime-dependent systems. A time-dependent system performs its tasks according to timing specifications specified within the system or imposed from outside the system. Reflective techniques can be applied to programming time-dependent systems because (1) some application programs require the introduction of a new language construct for specifying timing requirements and (2) different applications may require domain-specific scheduling algorithms. To allow a programmer to add or modify language constructs or scheduling algorithms, however, a clear reflective architecture and program interfaces must be provided. This paper proposes a concurrent object-based reflective architecture (R 2 architecture) for time-dependent computing. This architecture is based on anindividual reflection scheme and introduces new meta-level objects (real-time meta objects) that are responsible for time-dependent capabilities. An alarm-clock object and a scheduler object are introduced, and message protocols between them and real-time meta objects are defined. We implemented this architecture on ABCL/R2 and created the Sampled Sound Player program as an application. With this application we provided three different scheduler objects and measured the impact of different scheduling algorithms on sound playback. The measured results show that a scheduler with more complex computations at the meta level exhibited less scheduling overhead, thus was capable of better sound playback. The other example, Time-dependent Graceful Degradation Scheme, demonstrates the programming of functionality degradation triggered by failure to satisfy timing specifications.

A mode-changeable 2-D DCT/IDCT processor for digital VCR

In a digital VCR, the DCT/IDCT is performed by both the 8/spl times/8 mode and the 2/spl times/4/spl times/8 mode to improve the coding efficiency. A new 2-D DCT/IDCT processor which requires minimal hardware overhead for 8/spl times/8/2/spl times/4/spl times/8 mode change for a digital VCR is presented. The proposed DCT/TDCT processor uses a concurrent architecture and executes both the DCT and IDCT with 8/spl times/8 and 2/spl times/4/spl times/8 mode selection. This chip is implemented on the basis of the row-column decomposition scheme. The proposed architecture minimizes the hardware overhead for the 2/spl times/4/spl times/8 mode by sharing the same data path with the 8/spl times/8 mode as much as possible. The proposed architecture also reduces the hardware and the chip size by exploiting the table look-up method instead of the extra multiplication circuits in the weighting coefficients handling. The implemented DCT/IDCT processor satisfies the accuracy specification of digital VCR.

A concurrent architecture for serializable production systems

This paper presents a new production system architecture that takes advantage of modern associative memory devices to allow parallel production firing, concurrent matching, and overlap among matching, selection, and firing of productions. We prove that the results produced by the architecture are correct according to the serializability criterion. A comprehensive event driven simulator is used to evaluate the scaling properties of the new architecture and to compare it with a parallel architecture that does global synchronization before every production firing. We also present measures for the improvement in speed due to the use of associative memories and an estimate for the amount of associative memory needed. Architectural evaluation is facilitated by a new benchmark program that allows for changes in the number of productions, the size of the database, the variance between the sizes of local data clusters, and the ratio between local and global data. Our results indicate that substantial improvements in speed can be achieved with a very modest increase in hardware cost.

Molecular dynamics simulations of covalent amorphous insulators on parallel computers

Algorithms are designed to implement molecular dynamics (MD) simulations on emerging concurrent architectures. A highly efficient multiresolution algorithm is designed to carry out large-scale MD simulations for systems with long-range Coulomb and three-body covalent interactions. Large-scale MD simulations of amorphous silica are carried out on systems containing up to 41 472 particles. The intermediate-range order represented by the first sharp diffraction peak (FSDP) in the neutron static structure factor shows a significant dependence on the system size. Correlations in the range 4–11 Å are found to play a vital role in determining the shape of the FSDP correctly. The calculated structure factor for the largest system is in excellent agreement with neutron diffraction experiments, including the height of the FSDP. Molecular dynamics simulations of porous silica, in the density range 2.2-0.1 g/cm 3, are carried out on a system of 41 472 particles. The internal surface area, pore surface-to-volume ratio, pore-size distribution, and other structural correlations are determined as a function of the density. Various dissimilar porous structures with different fractal dimensions are obtained by controlling the preparation schedule and temperature. Pore interface growth and the roughness of internally fractured surfaces in silica glasses are investigated by MD simulations of 1.12 million particles. During uniform dilation, the pores coalesce and grow in size. When the mass density is reduced to 1.4 g/cm 3, the pores grow catastrophically to cause fracture. The roughness exponent for fracture surfaces, α = 0.87 ± 0.02, supports experimental claims about the universality of α. The nature of the phonon density-of-states due to low-energy floppy modes in crystalline and glassy states of the high-temperature ceramic Si 3N 4 is investigated. Floppy modes appear continuously in the glass as the connectivity of the system is reduced. In the crystal, they appear suddenly at 30% volume expansion. The density-of-states due to the floppy modes varies linearly with energy, and the specific heat is significantly enhanced by the floppy modes.

Parallel implementation of the fuzzy ARTMAP neural network paradigm on a hypercube

Abstract: The recent surge of interest in connectionist models arose through the availability of high speed parallel supercomputers and the advent of new learning algorithms. The computations performed on concurrent architectures are less costly than similar ones performed on sequential machines. In this paper, the design and implementation of a parallel version of fuzzy ARTMAP (Carpenter et al. 1992), which encompasses both neural and fuzzy logic, is discussed. Fuzzy ARTMAP is a supervised learning algorithm utilising two fuzzy ART modules and an associated mapping network. A simplified version of fuzzy ARTMAP (SFAM) was designed by incorporating a simplification of the match tracking concept on unsupervised fuzzy ART paradigms. The proposed simplified version consists of only one fuzzy ART module and an associated mapping network. A parallel fuzzy ARTMAP (PFAM) algorithm is then designed and implemented on a hypercube simulator (iPSC). The algorithm is parallelised for any architecture and, with the exception of issues related to communications, the implementation remains the same on any type of parallel machine. PFAM enjoys the advantage of reduced training time that makes the algorithm a successful candidate for applications that require both online testing and training. Such applications can range from underwater sonar detection and chemical plant processing control to nuclear reactor process control, flexible manufacturing and systems analysis.

Efficient technique for partitioning and programming linear algebra algorithms on concurrent VLSI architectures

An efficient technique for partitioning and programming linear algebra algorithms on concurrent architectures is described and applied to 2-D wavefront arrays. The mapping of the computational elements (processes) to processors is based on the concept of folding. The mapping pattern on the 2-D full-size mesh of processes is composition of symmetric tiles of size 2 square root (N)*2 square root (N), N being the number of processors. The algorithm can be partitioned according to a globally sequential, locally parallel scheme. The code optimisation is performed by programming a few different types of tile, according to the algorithm. When the size of the problem is much larger than the size of the mesh of processors, a linear speed-up is achieved independently of the number of processors. Experimental results are presented for matrix multiplication, LU decomposition and the solution of triangular system equations on 2-D meshes of transputers programmed in Occam.

Concurrent packaging architecture design

Packaging constitutes one of the primary limits on the performance and partitioning of high density electronic systems. A concurrent design methodology for the design of physical packaging hierarchies is presented. Architecture, electrical performance, and energy management aspects of the system are included. The CAD system AUDIT implements this design methodology. The concurrent design capability has been illustrated using model systems derived from high speed Digital Equipment 3000/500 (Alpha) and IBM RS/6000 workstations. It is found that the choice of the packaging architecture as well as the impact of packaging on system performance is determined by the partitioning of the system. >

BECAUSE European workshop Sophia-Antipolis (France) 14–16 October 1992. General presentation of the BECAUSE Benchmark Set: BBS

Within the framework of the ESPRIT Project named BECAUSE (5417), entitled “Benchmarking of Concurrent Architectures for their Use in Scientific Engineering”, a set of industrial benchmarks was specified. The BECAUSE Benchmark Set (BBS) is scientifically application oriented. The various Test Programs have been extracted and specified from real industrial software such as electromagnetics software, semi-conductor simulation and computational fluid dynamics. The aim was to provide a general tool to industry in order to evaluate new generations of parallel computers and supercomputers in terms of memory capacity and computational power. Before parallelising a complete application, a scientific developer must answer the questions: ‘What is the performance of a given machine on my application and is it worth spending time to parallelise it?’ The BBS has been designed to provide simple benchmarks, as close as possible to critical parts of real scientific applications. Thus, by implementing various tests selected within the BBS, the user can get information useful to answer these questions. The serial version of the entire set of BBS is public and available on a mail server. In this paper, the objectives of the BECAUSE Project are recalled in order to replace the BBS in the context of this ESPRIT Project. Then, the applications from which test programs have been extracted are presented and all tests are described in detail. The next part is dedicated to the practical organisation of the BBS and is concerned with such aspects as documentation, input data and how to get the BBS, so that the BBS can be used inside and outside of the Project to assess parallel machines. As a conclusion, further developments are presented.

Topological properties, communication, and computation on WK‐recursive networks

AbstractRecently, with the support of the CAPRI (Concurrent Architecture and PRogramming environment for highly Integrated system) project, Vecchia and Sanges proposed a new general class of recursively scalable networks, termed WK‐recursive networks, and developed routing and broadcasting algorithms on them. They have also implemented the WK‐recursive networks using VLSI technology. This paper studies WK‐recursive networks by first investigating their topological properties such as diameter, connectivity, and Hamiltonicity. We then develop new and more efficient routing and broadcasting algorithms. Our routing algorithm can guarantee the shortest paths. Our broadcasting algorithm is much simpler and requires fewer extra bits to be transmitted. The broadcasting tree of our broadcasting algorithm is of minimal height (equal to the diameter), and each node receives the message exactly once. Moreover, we show the execution of descend/ ascend algorithms on the WK‐recursive networks using the bitonic sort as an illustrative example. © 1994 by John Wiley & Sons, Inc.

A Methodology for Programming Scalable Architectures

In scalable concurrent architectures, the performance of a parallel algorithm depends on the resource management policies used. Such policies determine, for example, how data is partitioned and distributed and how processes are scheduled. In particular, the performance of a parallel algorithm obtained by using a particular policy can be affected by increasing the size of the architecture or the input. In order to support scalability, we are developing a methodology for modular specification of partition and distribution strategies (PDSs). As a consequence, a PDS may be changed without modifying the code specifying the logic of a parallel algorithm. We illustrate our methodology for parallel algorithms that use dynamic data structures.

An object-oriented structure for transient dynamics on concurrent computers

This paper describes the deep knowledge encoded into sets of objects so that their aggregated behaviour simulates the dynamics of some physical systems. The details of the mathematical models used to date and their application to a number of standard numerical experiments are presented. A multi-time step scheme is given for cases where different time steps have to be used in different parts of a discretized model. Poincaré sections of the chaotic response of a multi-degree of freedom system indicates that complex large scale simulation can be carried out in this way. The natural parallelism of the method ensures good performance on concurrent architectures. Implementations on snared memory, distributed memory machines based on transputers, and the fine grained connection machine are described.

Computer simulation of materials using parallel architectures

Algorithms are designed to implement molecular dynamics (MD) and quantum molecular dynamics (QMD) simulations on emerging concurrent architectures. For systems with finite-range interactions, a domain-decomposition algorithm is used to implement the multiple-time-step (MTS) approach to MD simulations on distributed-memory multiple instruction multiple data (MIMD) machines. Parallel algorithms are also designed for MD simulations of bulk Coulombic systems. The performances of these algorithms are tested on the Intel iPSC/860 system. The computational complexities of these algorithms are O( N) and parallel efficiencies close to 0.9. The core computational kernel of the QMD approach consists of solutions of parabolic partial differential equations (PDE) such as the time-dependent Schrödinger equation or time-dependent Kohn-Sham equation. This problem is coupled with another computationally intensive problem, i.e., solution of elliptic PDEs (the Poisson equation) for the long-range electron-electron interaction. We have designed parallel algorithms for both problems on SIMD (single instruction multiple data) machines. In the past two years, we have used the parallel computer architectures in our Concurrent Computing Laboratory for Materials Simulations (CCLMS) to carry out MD and QMD simulations on network glasses, ceramic composites, nanophase materials, solid C 60 and graphitic tubules, and quantum transport in nanoscale devices. Structural transformation, intermediate-range order and dynamic behavior of SiO 2 glass at high pressures are investigated with molecular dynamics. At high densities, the height of the first sharp diffraction peak is considerably diminished, its position changes from 1.6 to 2.2 Å −1, and a new peak appears at 2.85 Å -1. At twice the normal density, SiO bond length increases, SiO coordination changes from 4 to 6 and OSiO bond-angle changes from 109° to 90°. This is a tetrahedral to octahedral transformation, which was reported recently by Meade, Hem1ey, and Mao. Molecular dynamics simulations of porous silica, in the density range 2.2-0.1 g/cm 3, are carried out on a 41 472 particle system using a MIMD computer. The internal surface area, pore surface-to-volume ratio, pore-size distribution, fractal dimension, correlation length and mean-particle size are determined as a function of the density. Structural transition between a condensed amorphous phase and a low-density porous phase is characterized by these quantities. Various dissimilar porous structures with different fractal dimensions are obtained by controlling the preparation schedule and temperature.