Parallelizing irregular and pointer-based computations automatically: perspectives from logic and constraint programming

Irregular computations pose sorne of the most interesting and challenging problems in automatic parallelization. Irregularity appears in certain kinds of numerical problems and is pervasive in symbolic applications. Such computations often use dynamic data structures, which make heavy use of pointers. This complicates all the steps of a parallelizing compiler, from independence detection to task partitioning and placement. Starting in the mid 80s there has been significant progress in the development of parallelizing compilers for logic pro­gramming (and more recently, constraint programming) resulting in quite capable paralle­lizers. The typical applications of these paradigms frequently involve irregular computations, and make heavy use of dynamic data structures with pointers, since logical variables represent in practice a well-behaved form of pointers. This arguably makes the techniques used in these compilers potentially interesting. In this paper, we introduce in a tutoríal way, sorne of the problems faced by parallelizing compilers for logic and constraint programs and provide pointers to sorne of the significant progress made in the area. In particular, this work has resulted in a series of achievements in the areas of inter-procedural pointer aliasing analysis for independence detection, cost models and cost analysis, cactus-stack memory management, techniques for managing speculative and irregular computations through task granularity control and dynamic task allocation such as work-stealing schedulers), etc.

Automatic parallelization of irregular and pointer-based computations: perspectives from logic and constraint programming

Irregular computations pose some of the most interesting and challenging problems in automatic parallelization. Irregularity appears in certain kinds of numerical problems and is pervasive in symbolic applications. Such computations often use dynamic data structures which make heavy use of pointers. This complicates all the steps of a parallelizing compiler, from independence detection to task partitioning and placement. In the past decade there has been significant progress in the development of parallelizing compilers for logic programming and, more recently, constraint programming. The typical applications of these paradigms frequently involve irregular computations, which arguably makes the techniques used in these compilers potentially interesting. In this paper we introduce in a tutorial way some of the problems faced by parallelizing compilers for logic and constraint programs. These include the need for inter-procedural pointer aliasing analysis for independence detection and having to manage speculative and irregular computations through task granularity control and dynamic task allocation. We also provide pointers to some of the progress made in these áreas. In the associated talk we demónstrate representatives of several generations of these parallelizing compilers.

Advanced handover procedure for cellular communication systems

Third Generation cellular communication systems are expected to support mixed cell architecture in which picocells, microcells and macrocells are used to achieve full coverage and increase the spectral capacity. Supporting higher numbers of mobile terminals and the use of smaller cells will result in an increase in the number of handovers, and consequently an increase in the time delays required to perform these handovers. Higher time delays will generate call interruptions and forced terminations, particularly for time sensitive applications like real-time multimedia and data services. Currently in the Global System for Mobile communications (GSM), the handover procedure is initiated and performed by the fixed part of the Public Land Mobile Network (PLMN). The mobile terminal is only capable of detecting candidate base stations suitable for the handover; it is the role of the network to interrogate a candidate base station for a free channel. Handover signalling is exchanged via the fixed network and the time delay required to perform the handover is greatly affected by the levels of teletraffic handled by the network. In this thesis, a new handover strategy is developed to reduce the total time delay for handovers in a microcellular system. The handover signalling is diverted from the fixed network to the air interface to prevent extra delays due to teletraffic congestion, and to allow the mobile terminal to exchange signalling directly with the candidate base station. The new strategy utilises Packet Reservation Multiple Access (PRMA) technique as a mechanism to transfer the control of the handover procedure from the fixed network to the mobile terminal. Simulation results are presented to show a dramatic reduction in the handover delay as compared to those obtained using fixed channel allocation and dynamic channel allocation schemes.

Cost-effective upgrade of WDM all-optical networks using overlay fibres and hop reduction links

Fibre overlay is a cost-effective technique to alleviate wavelength blocking in some links of a wavelength-routed optical network by increasing the number of wavelengths in those links. In this letter, we investigate the effects of overlaying fibre in an all-optical network (AON) based on GÉANT2 topology. The constraint-based routing and wavelength assignment (CB-RWA) algorithm locates where cost-efficient upgrades should be implemented. Through numerical examples, we demonstrate that the network capacity improves by 25 per cent by overlaying fibre on 10 per cent of the links, and by 12 per cent by providing hop reduction links comprising 2 per cent of the links. For the upgraded network, we also show the impact of dynamic traffic allocation on the blocking probability. Copyright © 2010 John Wiley & Sons, Ltd.

Cost-effective upgrade of WDM all-optical networks using overlay fibres and hop reduction links

Fibre overlay is a cost-effective technique to alleviate wavelength blocking in some links of a wavelength-routed optical network by increasing the number of wavelengths in those links. In this letter, we investigate the effects of overlaying fibre in an all-optical network (AON) based on GÉANT2 topology. The constraint-based routing and wavelength assignment (CB-RWA) algorithm locates where cost-efficient upgrades should be implemented. Through numerical examples, we demonstrate that the network capacity improves by 25 per cent by overlaying fibre on 10 per cent of the links, and by 12 per cent by providing hop reduction links comprising 2 per cent of the links. For the upgraded network, we also show the impact of dynamic traffic allocation on the blocking probability. Copyright © 2010 John Wiley & Sons, Ltd.

Comparison of threshold hydrologic response across northern catchments

Funded by Leverhulme Trust. Grant Number: F/00 152/AG

SDN enabled dynamically reconfigurable high capacity optical access architecture for converged services

Dynamically reconfigurable time-division multiplexing (TDM) dense wavelength division multiplexing (DWDM) long-reach passive optical networks (PONs) can support the reduction of nodes and network interfaces by enabling a fully meshed flat optical core. In this paper we demonstrate the flexibility of the TDM-DWDM PON architecture, which can enable the convergence of multiple service types on a single physical layer. Heterogeneous services and modulation formats, i.e. residential 10G PON channels, business 100G dedicated channel and wireless fronthaul, are demonstrated co-existing on the same long reach TDM-DWDM PON system, with up to 100km reach, 512 users and emulated system load of 40 channels, employing amplifier nodes with either erbium doped fiber amplifiers (EDFAs) or semiconductor optical amplifiers (SOAs). For the first time end-to-end software defined networking (SDN) management of the access and core network elements is also implemented and integrated with the PON physical layer in order to demonstrate two service use cases: a fast protection mechanism with end-to-end service restoration in the case of a primary link failure; and dynamic wavelength allocation (DWA) in response to an increased traffic demand.

Parallel Cyclostationarity-Exploiting Algorithm for Energy-Efficient Spectrum Sensing

The evolution of wireless communication systems leads to Dynamic Spectrum Allocation for Cognitive Radio, which requires reliable spectrum sensing techniques. Among the spectrum sensing methods proposed in the literature, those that exploit cyclostationary characteristics of radio signals are particularly suitable for communication environments with low signal-to-noise ratios, or with non-stationary noise. However, such methods have high computational complexity that directly raises the power consumption of devices which often have very stringent low-power requirements. We propose a strategy for cyclostationary spectrum sensing with reduced energy consumption. This strategy is based on the principle that p processors working at slower frequencies consume less power than a single processor for the same execution time. We devise a strict relation between the energy savings and common parallel system metrics. The results of simulations show that our strategy promises very significant savings in actual devices.

Parallel Cyclostationarity-Exploiting Algorithm for Energy-Efficient Spectrum Sensing

The evolution of wireless communication systems leads to Dynamic Spectrum Allocation for Cognitive Radio, which requires reliable spectrum sensing techniques. Among the spectrum sensing methods proposed in the literature, those that exploit cyclostationary characteristics of radio signals are particularly suitable for communication environments with low signal-to-noise ratios, or with non-stationary noise. However, such methods have high computational complexity that directly raises the power consumption of devices which often have very stringent low-power requirements. We propose a strategy for cyclostationary spectrum sensing with reduced energy consumption. This strategy is based on the principle that p processors working at slower frequencies consume less power than a single processor for the same execution time. We devise a strict relation between the energy savings and common parallel system metrics. The results of simulations show that our strategy promises very significant savings in actual devices.

Selection of energy storage system for a regenerative dynamic dynamometer

This paper presents a design technique of a fully regenerative dynamic dynamometer. It incorporates an energy storage system to absorb the energy variation due to dynamometer transients. This allows the minimum power electronics requirement at the grid to supply the losses. The simulation results of the full system over a driving cycle show the amount of energy required to complete a driving cycle, therefore the size of the energy storage system can be determined.

Constrained control design for dynamic positioning of marine vehicles with control allocation

In this paper, we address the control design problem of positioning of over-actuated marine vehicles with control allocation. The proposed design is based on a combined position and velocity loops in a multi-variable anti-windup implementation together with a control allocation mapping. The vehicle modelling is considered with appropriate simplifications related to low-speed manoeuvring hydrodynamics and vehicle symmetry. The control design is considered together with a control allocation mapping. We derive analytical tuning rules based on requirements of closed-loop stability and performance. The anti- windup implementation of the controller is obtained by mapping the actuator-force constraint set into a constraint set for the generalized forces. This approach ensures that actuation capacity is not violated by constraining the generalized control forces; thus, the control allocation is simplified since it can be formulated as an unconstrained problem. The mapping can also be modified on-line based on actuator availability to provide actuator-failure accommodation. We provide a proof of the closed-loop stability and illustrate the performance using simulation scenarios for an open-frame underwater vehicle.

Anti-wind-up designs for dynamic positioning of marine vehicles with control allocation

This paper presents a framework for the design of a joint motion controller and a control allocation strategy for dynamic positioning of marine vehicles. The key aspects of the proposed designs are a systematic approach to deal with actuator saturation and to inform the motion controller about saturation. The proposed system uses a mapping that translates the actuator constraint sets into constraint sets at the motion controller level. Hence, while the motion controller addresses the constraints, the control allocation algorithm can solve an unconstrained optimisation problem. The constrained control design is approached using a multivariable anti-wind-up strategy for strictly proper controllers. This is applicable to the implementation of PI and PID type of motion controllers.

The Learning Component of Dynamic Allocation Indices

For a multiarmed bandit problem with exponential discounting the optimal allocation rule is defined by a dynamic allocation index defined for each arm on its space. The index for an arm is equal to the expected immediate reward from the arm, with an upward adjustment reflecting any uncertainty about the prospects of obtaining rewards from the arm, and the possibilities of resolving those uncertainties by selecting that arm. Thus the learning component of the index is defined to be the difference between the index and the expected immediate reward. For two arms with the same expected immediate reward the learning component should be larger for the arm for which the reward rate is more uncertain. This is shown to be true for arms based on independent samples from a fixed distribution with an unknown parameter in the cases of Bernoulli and normal distributions, and similar results are obtained in other cases.

An algorithm for dynamic optimal bandwidth allocation in communication networks

We study the problem of optimal bandwidth allocation in communication networks. We consider a queueing model with two queues to which traffic from different competing flows arrive. The queue length at the buffers is observed every T instants of time, on the basis of which a decision on the amount of bandwidth to be allocated to each buffer for the next T instants is made. We consider a class of closed-loop feedback policies for the system and use a twotimescale simultaneous perturbation stochastic approximation(SPSA) algorithm to find an optimal policy within the prescribed class. We study the performance of the proposed algorithm on a numerical setting. Our algorithm is found to exhibit good performance.