Using Computational Significance and Resilience in System Software Stacks.:Keynote Talk

This talk explores how the runtime system and operating system can leverage metrics that express the significance and resilience of application components in order to reduce the energy footprint of parallel applications. We will explore in particular how software can tolerate and indeed exploit higher error rates in future processors and memory technologies that may operate outside their safe margins.

High-pressure deformation mechanism in scolecite: A combined computational-experimental study

Pressure-induced structural modifications in scolecite were studied by means of in situ synchrotron X-ray powder diffraction and density functional computations. The experimental cell parameters were refined up to 8.5 GPa. Discontinuities in the slope of the unit-cell parameters vs. pressure dependence were observed; as a consequence, an increase in the slope of the linear pressure-volume dependence is observed at about 6 GPa, suggesting an enhanced compressibility at higher pressures. Weakening and broadening of the diffraction peaks reveals increasing structural disorder with pressure, preventing refinement of the lattice parameters above 8.5 GPa. Diffraction patterns collected during decompression show that the disorder is irreversible. Atomic coordinates within unit cells of different dimensions were determined by means of Car-Parrinello simulations. The discontinuous rise in compressibility at about 6 GPa is reproduced by the computation, allowing us to attribute it to re-organization of the hydrogen bonding network, with the formation of water dimers. Moreover we found that, with increasing pressure, the tetrahedral chains parallel to c rotate along their elongation axis and display an increasing twisting along a direction perpendicular to c. At the same time, we observed the compression of the channels. We discuss the modification of the Ca polyhedra under pressure, and the increase in coordination number (from 4 to 5) of one of the two Al atoms, resulting from the approach of a water molecule. We speculate that this last transformation triggers the irreversible disordering of the system.

Model selection approaches for non-linear system identification: a review

The identification of non-linear systems using only observed finite datasets has become a mature research area over the last two decades. A class of linear-in-the-parameter models with universal approximation capabilities have been intensively studied and widely used due to the availability of many linear-learning algorithms and their inherent convergence conditions. This article presents a systematic overview of basic research on model selection approaches for linear-in-the-parameter models. One of the fundamental problems in non-linear system identification is to find the minimal model with the best model generalisation performance from observational data only. The important concepts in achieving good model generalisation used in various non-linear system-identification algorithms are first reviewed, including Bayesian parameter regularisation and models selective criteria based on the cross validation and experimental design. A significant advance in machine learning has been the development of the support vector machine as a means for identifying kernel models based on the structural risk minimisation principle. The developments on the convex optimisation-based model construction algorithms including the support vector regression algorithms are outlined. Input selection algorithms and on-line system identification algorithms are also included in this review. Finally, some industrial applications of non-linear models are discussed.

Robust fuzzy Gustafson-Kessel clustering for nonlinear system identification

This paper deals with Takagi-Sugeno (TS) fuzzy model identification of nonlinear systems using fuzzy clustering. In particular, an extended fuzzy Gustafson-Kessel (EGK) clustering algorithm, using robust competitive agglomeration (RCA), is developed for automatically constructing a TS fuzzy model from system input-output data. The EGK algorithm can automatically determine the 'optimal' number of clusters from the training data set. It is shown that the EGK approach is relatively insensitive to initialization and is less susceptible to local minima, a benefit derived from its agglomerate property. This issue is often overlooked in the current literature on nonlinear identification using conventional fuzzy clustering. Furthermore, the robust statistical concepts underlying the EGK algorithm help to alleviate the difficulty of cluster identification in the construction of a TS fuzzy model from noisy training data. A new hybrid identification strategy is then formulated, which combines the EGK algorithm with a locally weighted, least-squares method for the estimation of local sub-model parameters. The efficacy of this new approach is demonstrated through function approximation examples and also by application to the identification of an automatic voltage regulation (AVR) loop for a simulated 3 kVA laboratory micro-machine system.

Query cost estimation through remote system contention states analysis over the Internet

Query processing over the Internet involving autonomous data sources is a major task in data integration. It requires the estimated costs of possible queries in order to select the best one that has the minimum cost. In this context, the cost of a query is affected by three factors: network congestion, server contention state, and complexity of the query. In this paper, we study the effects of both the network congestion and server contention state on the cost of a query. We refer to these two factors together as system contention states. We present a new approach to determining the system contention states by clustering the costs of a sample query. For each system contention state, we construct two cost formulas for unary and join queries respectively using the multiple regression process. When a new query is submitted, its system contention state is estimated first using either the time slides method or the statistical method. The cost of the query is then calculated using the corresponding cost formulas. The estimated cost of the query is further adjusted to improve its accuracy. Our experiments show that our methods can produce quite accurate cost estimates of the submitted queries to remote data sources over the Internet.

Reconfigurable system-on-a-chip motion estimation architecture for multi-standard video coding

A new domain-specific, reconfigurable system-on-a-chip (SoC) architecture is proposed for video motion estimation. This has been designed to cover most of the common block-based video coding standards, including MPEG-2, MPEG-4, H.264, WMV-9 and AVS. The architecture exhibits simple control, high throughput and relatively low hardware cost when compared with existing circuits. It can also easily handle flexible search ranges without any increase in silicon area and can be configured prior to the start of the motion estimation process for a specific standard. The computational rates achieved make the circuit suitable for high-end video processing applications, such as HDTV. Silicon design studies indicate that circuits based on this approach incur only a relatively small penalty in terms of power dissipation and silicon area when compared with implementations for specific standards. Indeed, the cost/performance achieved exceeds that of existing but specific solutions and greatly exceeds that of general purpose field programmable gate array (FPGA) designs.

Computational Verification of Two Universal Relations for Simple Ionic Liquids. Kinetic Properties of a Model 2:1 Molten Salt

Two semianalytical relations [Nature, 1996, 381, 137 and Phys. Rev. Lett. 2001, 87, 245901] predicting dynamical coefficients of simple liquids on the basis of structural properties have been tested by extensive molecular dynamics simulations for an idealized 2:1 model molten salt. In agreement with previous simulation studies, our results support the validity of the relation expressing the self-diffusion coefficient as a Function of the radial distribution functions for all thermodynamic conditions such that the system is in the ionic (ie., fully dissociated) liquid state. Deviations are apparent for high-density samples in the amorphous state and in the low-density, low-temperature range, when ions condense into AB(2) molecules. A similar relation predicting the ionic conductivity is only partially validated by our data. The simulation results, covering 210 distinct thermodynamic states, represent an extended database to tune and validate semianalytical theories of dynamical properties and provide a baseline for the interpretation of properties of more complex systems such as the room-temperature ionic liquids.

A Theoretical and Experimental Study of Resonance in a High Performance Engine Intake System: Part 1

The unsteady gas dynamic phenomena in engine intake systems of the type found in racecars have been examined. In particular, the resonant tuning effects, including cylinder-to-cylinder power variations, which can occur as a result of the interaction between an engine and its airbox have been considered. Frequency analysis of the output from a Virtual 4-Stroke 1D engine simulation was used to characterise the forcing function applied by an engine to an airbox. A separate computational frequency sweeping technique, which employed the CFD package FLUENT, was used to determine the natural frequencies of virtual airboxes in isolation from an engine. Using this technique, an airbox with a natural frequency at 75 Hz was designed for a Yamaha R6 4-cylinder motorcycle engine. The existence of an airbox natural frequency at 75 Hz was subsequently confirmed by an experimental frequency sweeping technique carried out on the engine test bed. A coupled 1D/3D analysis which employed the engine simulation package Virtual 4-Stroke and the CFD package FLUENT, was used to model the combined engine and airbox system. The coupled 1D/3D analysis predicted a 75 Hz resonance of the airbox at an engine speed of 9000 rpm. This frequency was the induction frequency for a single cylinder. An airbox was fabricated and tested on the engine. Static pressure was recorded at a grid of points in the airbox as the engine was swept through a speed range of 3000 to 10000 rpm. The measured engine speed corresponding to resonance in the airbox agreed well with the predicted values. There was also good correlation between the amplitude and phase of the pressure traces recorded within the airbox and the 1D/3D predictions.

Computational techniques for a simple theory of conditional preferences

A simple logic of conditional preferences is defined, with a language that allows the compact representation of certain kinds of conditional preference statements, a semantics and a proof theory. CP-nets and TCP-nets can be mapped into this logic, and the semantics and proof theory generalise those of CP-nets and TCP-nets. The system can also express preferences of a lexicographic kind. The paper derives various sufficient conditions for a set of conditional preferences to be consistent, along with algorithmic techniques for checking such conditions and hence confirming consistency. These techniques can also be used for totally ordering outcomes in a way that is consistent with the set of preferences, and they are further developed to give an approach to the problem of constrained optimisation for conditional preferences.

The Application of Computational Chemistry and Chemometrics to Developing a Method for Online Monitoring of Polymer Degradation in the Manufacture of Bioresorbable Medical Implants

Bioresorbable polymers such as PLA have an important role to play in the development of temporary implantable medical devices with significant benefits over traditional therapies. However, development of new devices is hindered by high manufacturing costs associated with difficulties in processing the material. A major problem is the lack of insight on material degradation during processing. In this work, a method of quantifying degradation of PLA using IR spectroscopy coupled with computational chemistry and chemometric modeling is examined. It is shown that the method can predict the quantity of degradation products in solid-state samples with reasonably good accuracy, indicating the potential to adapt the method to developing an on-line sensor for monitoring PLA degradation in real-time during processing.