GPGPU implementation of growing neural gas: application to 3D scene reconstruction

Self-organising neural models have the ability to provide a good representation of the input space. In particular the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time-consuming, especially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This paper proposes a Graphics Processing Unit (GPU) parallel implementation of the GNG with Compute Unified Device Architecture (CUDA). In contrast to existing algorithms, the proposed GPU implementation allows the acceleration of the learning process keeping a good quality of representation. Comparative experiments using iterative, parallel and hybrid implementations are carried out to demonstrate the effectiveness of CUDA implementation. The results show that GNG learning with the proposed implementation achieves a speed-up of 6× compared with the single-threaded CPU implementation. GPU implementation has also been applied to a real application with time constraints: acceleration of 3D scene reconstruction for egomotion, in order to validate the proposal.

Large scale data processing using MapReduce

As massive data sets become increasingly available, people are facing the problem of how to effectively process and understand these data. Traditional sequential computing models are giving way to parallel and distributed computing models, such as MapReduce, both due to the large size of the data sets and their high dimensionality. This dissertation, as in the same direction of other researches that are based on MapReduce, tries to develop effective techniques and applications using MapReduce that can help people solve large-scale problems. Three different problems are tackled in the dissertation. The first one deals with processing terabytes of raster data in a spatial data management system. Aerial imagery files are broken into tiles to enable data parallel computation. The second and third problems deal with dimension reduction techniques that can be used to handle data sets of high dimensionality. Three variants of the nonnegative matrix factorization technique are scaled up to factorize matrices of dimensions in the order of millions in MapReduce based on different matrix multiplication implementations. Two algorithms, which compute CANDECOMP/PARAFAC and Tucker tensor decompositions respectively, are parallelized in MapReduce based on carefully partitioning the data and arranging the computation to maximize data locality and parallelism.

Multi-Tenant Virtual GPUs for Optimising Performance of a Financial Risk Application

Graphics Processing Units (GPUs) are becoming popular accelerators in modern High-Performance Computing (HPC) clusters. Installing GPUs on each node of the cluster is not efficient resulting in high costs and power consumption as well as underutilisation of the accelerator. The research reported in this paper is motivated towards the use of few physical GPUs by providing cluster nodes access to remote GPUs on-demand for a financial risk application. We hypothesise that sharing GPUs between several nodes, referred to as multi-tenancy, reduces the execution time and energy consumed by an application. Two data transfer modes between the CPU and the GPUs, namely concurrent and sequential, are explored. The key result from the experiments is that multi-tenancy with few physical GPUs using sequential data transfers lowers the execution time and the energy consumed, thereby improving the overall performance of the application.

Computer aided parallelisation of unstructured mesh codes

Abstract not available

Load sharing and power quality enhanced operation of a distributed microgrid

This paper describes control methods for proper load sharing between parallel converters connected in a microgrid and supplied by distributed generators (DGs). It is assumed that the microgrid spans a large area and it supplies loads in both in grid connected and islanded modes. A control strategy is proposed to improve power quality and proper load sharing in both islanded and grid connected modes. It is assumed that each of the DGs has a local load connected to it which can be unbalanced and/or nonlinear. The DGs compensate the effects of unbalance and nonlinearity of the local loads. Common loads are also connected to the microgrid, which are supplied by the utility grid under normal conditions. However during islanding, each of the DGs supplies its local load and shares the common load through droop characteristics. Both impedance and motor loads are considered to verify the system response. The efficacy of the controller has been validated through simulation for various operating conditions using PSCAD. It has been found through simulation that the total Harmonic Distortion (THD) of the of the microgrid voltage is about 10% and the negative and zero sequence component are around 20% of the positive sequence component before compensation. After compensation, the THD remain below 0.5%, whereas, negative and zero sequence components of the voltages remain below 0.02% of the positive sequence component.

Parallel linear system solution and its application to railway power network simulation

The Streaming SIMD extension (SSE) is a special feature embedded in the Intel Pentium III and IV classes of microprocessors. It enables the execution of SIMD type operations to exploit data parallelism. This article presents improving computation performance of a railway network simulator by means of SSE. Voltage and current at various points of the supply system to an electrified railway line are crucial for design, daily operation and planning. With computer simulation, their time-variations can be attained by solving a matrix equation, whose size mainly depends upon the number of trains present in the system. A large coefficient matrix, as a result of congested railway line, inevitably leads to heavier computational demand and hence jeopardizes the simulation speed. With the special architectural features of the latest processors on PC platforms, significant speed-up in computations can be achieved.

SSE based parallel solution for power systems network equations

Streaming SIMD Extensions (SSE) is a unique feature embedded in the Pentium III class of microprocessors. By fully exploiting SSE, parallel algorithms can be implemented on a standard personal computer and a theoretical speedup of four can be achieved. In this paper, we demonstrate the implementation of a parallel LU matrix decomposition algorithm for solving power systems network equations with SSE and discuss advantages and disadvantages of this approach.

Distributed optimization in power networks and general multi-agent systems

The dissertation studies the general area of complex networked systems that consist of interconnected and active heterogeneous components and usually operate in uncertain environments and with incomplete information. Problems associated with those systems are typically large-scale and computationally intractable, yet they are also very well-structured and have features that can be exploited by appropriate modeling and computational methods. The goal of this thesis is to develop foundational theories and tools to exploit those structures that can lead to computationally-efficient and distributed solutions, and apply them to improve systems operations and architecture.

Specifically, the thesis focuses on two concrete areas. The first one is to design distributed rules to manage distributed energy resources in the power network. The power network is undergoing a fundamental transformation. The future smart grid, especially on the distribution system, will be a large-scale network of distributed energy resources (DERs), each introducing random and rapid fluctuations in power supply, demand, voltage and frequency. These DERs provide a tremendous opportunity for sustainability, efficiency, and power reliability. However, there are daunting technical challenges in managing these DERs and optimizing their operation. The focus of this dissertation is to develop scalable, distributed, and real-time control and optimization to achieve system-wide efficiency, reliability, and robustness for the future power grid. In particular, we will present how to explore the power network structure to design efficient and distributed market and algorithms for the energy management. We will also show how to connect the algorithms with physical dynamics and existing control mechanisms for real-time control in power networks.

The second focus is to develop distributed optimization rules for general multi-agent engineering systems. A central goal in multiagent systems is to design local control laws for the individual agents to ensure that the emergent global behavior is desirable with respect to the given system level objective. Ideally, a system designer seeks to satisfy this goal while conditioning each agent’s control on the least amount of information possible. Our work focused on achieving this goal using the framework of game theory. In particular, we derived a systematic methodology for designing local agent objective functions that guarantees (i) an equivalence between the resulting game-theoretic equilibria and the system level design objective and (ii) that the resulting game possesses an inherent structure that can be exploited for distributed learning, e.g., potential games. The control design can then be completed by applying any distributed learning algorithm that guarantees convergence to the game-theoretic equilibrium. One main advantage of this game theoretic approach is that it provides a hierarchical decomposition between the decomposition of the systemic objective (game design) and the specific local decision rules (distributed learning algorithms). This decomposition provides the system designer with tremendous flexibility to meet the design objectives and constraints inherent in a broad class of multiagent systems. Furthermore, in many settings the resulting controllers will be inherently robust to a host of uncertainties including asynchronous clock rates, delays in information, and component failures.

Modular distributed manufacturing systems and the implications for integrated control

Driven by the need for more responsive manufacturing processes and as a consequence of increasing complexity in products and production systems, this short paper introduces a number of developments in the area of modular, distributed manufacturing systems. Requirements for the development of such systems are addressed and, in particular, the relevance to current and future integrated control systems is examined. One of the key issues for integrated control systems in the future is the need to provide support for distributed decision-making in addition to existing distributed control capabilities.

Design and Evaluation of the Hamal Parallel Computer

Parallel shared-memory machines with hundreds or thousands of processor-memory nodes have been built; in the future we will see machines with millions or even billions of nodes. Associated with such large systems is a new set of design challenges. Many problems must be addressed by an architecture in order for it to be successful; of these, we focus on three in particular. First, a scalable memory system is required. Second, the network messaging protocol must be fault-tolerant. Third, the overheads of thread creation, thread management and synchronization must be extremely low. This thesis presents the complete system design for Hamal, a shared-memory architecture which addresses these concerns and is directly scalable to one million nodes. Virtual memory and distributed objects are implemented in a manner that requires neither inter-node synchronization nor the storage of globally coherent translations at each node. We develop a lightweight fault-tolerant messaging protocol that guarantees message delivery and idempotence across a discarding network. A number of hardware mechanisms provide efficient support for massive multithreading and fine-grained synchronization. Experiments are conducted in simulation, using a trace-driven network simulator to investigate the messaging protocol and a cycle-accurate simulator to evaluate the Hamal architecture. We determine implementation parameters for the messaging protocol which optimize performance. A discarding network is easier to design and can be clocked at a higher rate, and we find that with this protocol its performance can approach that of a non-discarding network. Our simulations of Hamal demonstrate the effectiveness of its thread management and synchronization primitives. In particular, we find register-based synchronization to be an extremely efficient mechanism which can be used to implement a software barrier with a latency of only 523 cycles on a 512 node machine.

Mermera: Non-Coherent Distributed Shared Memory for Parallel Computing

The proliferation of inexpensive workstations and networks has prompted several researchers to use such distributed systems for parallel computing. Attempts have been made to offer a shared-memory programming model on such distributed memory computers. Most systems provide a shared-memory that is coherent in that all processes that use it agree on the order of all memory events. This dissertation explores the possibility of a significant improvement in the performance of some applications when they use non-coherent memory. First, a new formal model to describe existing non-coherent memories is developed. I use this model to prove that certain problems can be solved using asynchronous iterative algorithms on shared-memory in which the coherence constraints are substantially relaxed. In the course of the development of the model I discovered a new type of non-coherent behavior called Local Consistency. Second, a programming model, Mermera, is proposed. It provides programmers with a choice of hierarchically related non-coherent behaviors along with one coherent behavior. Thus, one can trade-off the ease of programming with coherent memory for improved performance with non-coherent memory. As an example, I present a program to solve a linear system of equations using an asynchronous iterative algorithm. This program uses all the behaviors offered by Mermera. Third, I describe the implementation of Mermera on a BBN Butterfly TC2000 and on a network of workstations. The performance of a version of the equation solving program that uses all the behaviors of Mermera is compared with that of a version that uses coherent behavior only. For a system of 1000 equations the former exhibits at least a 5-fold improvement in convergence time over the latter. The version using coherent behavior only does not benefit from employing more than one workstation to solve the problem while the program using non-coherent behavior continues to achieve improved performance as the number of workstations is increased from 1 to 6. This measurement corroborates our belief that non-coherent shared memory can be a performance boon for some applications.

Parallel Auditory Filtering By Sustained and Transient Channels Separates Coarticulated Vowels and Consonants

A neural model of peripheral auditory processing is described and used to separate features of coarticulated vowels and consonants. After preprocessing of speech via a filterbank, the model splits into two parallel channels, a sustained channel and a transient channel. The sustained channel is sensitive to relatively stable parts of the speech waveform, notably synchronous properties of the vocalic portion of the stimulus it extends the dynamic range of eighth nerve filters using coincidence deteectors that combine operations of raising to a power, rectification, delay, multiplication, time averaging, and preemphasis. The transient channel is sensitive to critical features at the onsets and offsets of speech segments. It is built up from fast excitatory neurons that are modulated by slow inhibitory interneurons. These units are combined over high frequency and low frequency ranges using operations of rectification, normalization, multiplicative gating, and opponent processing. Detectors sensitive to frication and to onset or offset of stop consonants and vowels are described. Model properties are characterized by mathematical analysis and computer simulations. Neural analogs of model cells in the cochlear nucleus and inferior colliculus are noted, as are psychophysical data about perception of CV syllables that may be explained by the sustained transient channel hypothesis. The proposed sustained and transient processing seems to be an auditory analog of the sustained and transient processing that is known to occur in vision.

Distributed Linear Precoding and User Selection in Coordinated Multicell Systems

In this manuscript we tackle the problem of semidistributed user selection with distributed linear precoding for sum rate maximization in multiuser multicell systems. A set of adjacent base stations (BS) form a cluster in order to perform coordinated transmission to cell-edge users, and coordination is carried out through a central processing unit (CU). However, the message exchange between BSs and the CU is limited to scheduling control signaling and no user data or channel state information (CSI) exchange is allowed. In the considered multicell coordinated approach, each BS has its own set of cell-edge users and transmits only to one intended user while interference to non-intended users at other BSs is suppressed by signal steering (precoding). We use two distributed linear precoding schemes, Distributed Zero Forcing (DZF) and Distributed Virtual Signalto-Interference-plus-Noise Ratio (DVSINR). Considering multiple users per cell and the backhaul limitations, the BSs rely on local CSI to solve the user selection problem. First we investigate how the signal-to-noise-ratio (SNR) regime and the number of antennas at the BSs impact the effective channel gain (the magnitude of the channels after precoding) and its relationship with multiuser diversity. Considering that user selection must be based on the type of implemented precoding, we develop metrics of compatibility (estimations of the effective channel gains) that can be computed from local CSI at each BS and reported to the CU for scheduling decisions. Based on such metrics, we design user selection algorithms that can find a set of users that potentially maximizes the sum rate. Numerical results show the effectiveness of the proposed metrics and algorithms for different configurations of users and antennas at the base stations.

Type Systems for Distributed Programs: Components and Sessions

Modern software systems, in particular distributed ones, are everywhere around us and are at the basis of our everyday activities. Hence, guaranteeing their cor- rectness, consistency and safety is of paramount importance. Their complexity makes the verification of such properties a very challenging task. It is natural to expect that these systems are reliable and above all usable. i) In order to be reliable, compositional models of software systems need to account for consistent dynamic reconfiguration, i.e., changing at runtime the communication patterns of a program. ii) In order to be useful, compositional models of software systems need to account for interaction, which can be seen as communication patterns among components which collaborate together to achieve a common task. The aim of the Ph.D. was to develop powerful techniques based on formal methods for the verification of correctness, consistency and safety properties related to dynamic reconfiguration and communication in complex distributed systems. In particular, static analysis techniques based on types and type systems appeared to be an adequate methodology, considering their success in guaranteeing not only basic safety properties, but also more sophisticated ones like, deadlock or livelock freedom in a concurrent setting. The main contributions of this dissertation are twofold. i) On the components side: we design types and a type system for a concurrent object-oriented calculus to statically ensure consistency of dynamic reconfigurations related to modifications of communication patterns in a program during execution time. ii) On the communication side: we study advanced safety properties related to communication in complex distributed systems like deadlock-freedom, livelock- freedom and progress. Most importantly, we exploit an encoding of types and terms of a typical distributed language, session π-calculus, into the standard typed π- calculus, in order to understand their expressive power.

Divided we stand: Parallel distributed stack memory management

We present an overview of the stack-based memory management techniques that we used in our non-deterministic and-parallel Prolog systems: &-Prolog and DASWAM. We believe that the problems associated with non-deterministic and-parallel systems are more general than those encountered in or-parallel and deterministic and-parallel systems, which can be seen as subsets of this more general case. We develop on the previously proposed "marker scheme", lifting some of the restrictions associated with the selection of goals while keeping (virtual) memory consumption down. We also review some of the other problems associated with the stack-based management scheme, such as handling of forward and backward execution, cut, and roll-backs.