Systolic VLSI compiler (SVC) for high performance vector quantisation chips

An overview is given of a systolic VLSI compiler (SVC) tool currently under development for the automated design of high performance digital signal processing (DSP) chips. Attention is focused on the design of systolic vector quantization chips for use in both speech and image coding systems. The software in question consists of a cell library, silicon assemblers, simulators, test pattern generators, and a specially designed graphics shell interface which makes it expandable and user friendly. It allows very high performance digital coding systems to be rapidly designed in VLSI.

On the application of genetic algorithms for optimising composites against impact loading

A genetic algorithm (GA) was adopted to optimise the response of a composite laminate subject to impact. Two different impact scenarios are presented: low-velocity impact of a slender laminated strip and high-velocity impact of a rectangular plate by a spherical impactor. In these cases, the GA's objective was to, respectively, minimise the peak deflection and minimise penetration by varying the ply angles.

The GA was coupled to a commercial finite-element (FE) package LS DYNA to perform the impact analyses. A comparison with a commercial optimisation package, LS OPT, was also made. The results showed that the GA was a robust, capable optimisation tool that produced near optimal designs, and performed well with respect to LS OPT for the more complex high-velocity impact scenario tested.

Optimising sorting and washing of home-grown maize to reduce fumonisin contamination under laboratory-controlled conditions

Subsistence farming communities with low socio-economic status reliant on a mono cereal maize diet are exposed to fumonisin levels that exceed the provisional maximum tolerable daily intake of 2 mu g kg(-1) body weight day(-1) recommended by the Joint FAO/WHO Expert Committee on Food Additives. In the rural Centane magisterial district, Eastern Cape Province, South Africa, it is customary during food preparation to sort visibly infected maize kernels from good maize kernels and to wash the good kernels prior to cooking. However, this customary practice seems not to sufficiently reduce the fumonisin levels. This is the first study to optimise the reduction of fumonisin mycotoxins in home-grown maize based on customary methods of a rural population, under laboratory-controlled conditions. Maize obtained from subsistence farmers was analysed for the major naturally occurring fumonisins (FB1, FB2 and FB3) by fluorescence HPLC. Large variations were observed in the unsorted and the experimental maize batches attributable to the non-homogeneous distribution of fumonisin contamination in maize kernels. Optimised hand-sorting of maize kernels by removing the visibly infected/damaged kernels (fumonisins, 53.7 +/- 15.0 mg kg(-1), 2.5% by weight) reduced the mean fumonisins from 2.32 +/- 1.16 mg kg(-1) to 0.68 +/- 0.42 mg kg(-1). Hand washing of the sorted good maize kernels for a period of 10 min at 25 degrees C resulted in optimal reduction with no additional improvement for wash periods up to 15 h. The laboratory optimised sorting reduced the fumonisins by 71 +/- 18% and an additional 13 +/- 12% with the 10 min wash. Based on these results and on local practices and practicalities the protocol that would be recommended to subsistence farmers consists of the removal of the infected/damaged kernels from the maize followed by a 10 min ambient temperature water wash. (C) 2010 Elsevier Ltd. All rights reserved.

Optimising the locations of thermally sensitive equipment in an aircraft crown compartment

A Design of Experiments (DoE) analysis was undertaken to generate a list of configurations for CFD numerical simulation of an aircraft crown compartment. Fitted regression models were built to predict the convective heat transfer coefficients of thermally sensitive dissipating elements located inside this compartment. These are namely the SEPDC and the Route G. Currently they are positioned close to the fuselage and it is of interest to optimise the heat transfer for reliability and performance purposes. Their locations and the external fuselage surface temperature were selected as input variables for the DoE. The models fit the CFD data with values ranging from 0.878 to 0.978, and predict that the optimum locations in terms of heat transfer are when the elements are positioned as close to the crown floor as possible ( and ?min. limits), where they come in direct contact with the air flow from the cabin ventilation system, and when they are positioned close to the centreline ( and ?CL). The methodology employed allows aircraft thermal designers to optimise equipment placement in confined areas of an aircraft during the design phase. The determined models should be incorporated into global aircraft numerical models to improve accuracy and reduce model size and computational time. © 2012 Elsevier Masson SAS. All rights reserved.

An application of bi-directional evolutionary structural optimisation for optimising energy absorbing structures using a material damage model

The Bi-directional Evolutionary Structural Optimisation (BESO) method is a numerical topology optimisation method developed for use in finite element analysis. This paper presents a particular application of the BESO method to optimise the energy absorbing capability of metallic structures. The optimisation objective is to evolve a structural geometry of minimum mass while ensuring that the kinetic energy of an impacting projectile is reduced to a level which prevents perforation. Individual elements in a finite element mesh are deleted when a prescribed damage criterion is exceeded. An energy absorbing structure subjected to projectile impact will fail once the level of damage results in a critical perforation size. It is therefore necessary to constrain an optimisation algorithm from producing such candidate solutions. An algorithm to detect perforation was implemented within a BESO framework which incorporated a ductile material damage model.

Compiler and Runtime Support for Hybrid Static/Dynamic Scheduling

One of the outstanding issues in parallel computing is the selection of task granularity. This work proposes a solution to the task granularity problem by lowering the overhead of the task scheduler and as such supporting very fine-grain tasks. Using a combination of static (compile-time) scheduling and dynamic (run-time) scheduling, we aim to make scheduling decisions as fast as with static scheduling while retaining the dynamic load- balancing properties of fully dynamic scheduling. We present an example application and discuss the requirements on the compiler and runtime system to realize hybrid static/dynamic scheduling.

Optimising drug solubilisation in amorphous polymer dispersions: rational selection of hot-melt extrusion processing parameters

The aim of this article was to construct a T–ϕ phase diagram for a model drug (FD) and amorphous polymer (Eudragit® EPO) and to use this information to understand the impact of how temperature–composition coordinates influenced the final properties of the extrudate. Defining process boundaries and understanding drug solubility in polymeric carriers is of utmost importance and will help in the successful manufacture of new delivery platforms for BCS class II drugs. Physically mixed felodipine (FD)–Eudragit® EPO (EPO) binary mixtures with pre-determined weight fractions were analysed using DSC to measure the endset of melting and glass transition temperature. Extrudates of 10 wt% FD–EPO were processed using temperatures (110°C, 126°C, 140°C and 150°C) selected from the temperature–composition (T–ϕ) phase diagrams and processing screw speed of 20, 100 and 200rpm. Extrudates were characterised using powder X-ray diffraction (PXRD), optical, polarised light and Raman microscopy. To ensure formation of a binary amorphous drug dispersion (ADD) at a specific composition, HME processing temperatures should at least be equal to, or exceed, the corresponding temperature value on the liquid–solid curve in a F–H T–ϕ phase diagram. If extruded between the spinodal and liquid–solid curve, the lack of thermodynamic forces to attain complete drug amorphisation may be compensated for through the use of an increased screw speed. Constructing F–H T–ϕ phase diagrams are valuable not only in the understanding drug–polymer miscibility behaviour but also in rationalising the selection of important processing parameters for HME to ensure miscibility of drug and polymer.

Optimising postbuckling composite panels for damage resistance

The design of current composite primary aerostructures, such as fuselage or wing stiffened panels, tends to be conservative due to the susceptibility of the relatively weak skin-stiffener interface. This weakness is due to through-thickness stresses which are exacerbated by deformations due to buckling. This paper presents a finite-elementbased optimization strategy, utilizing a global-local modelling approach, for postbuckling stiffened panels which takes into account damage mechanisms which may lead to delamination and subsequent failure of the panel due to stiffener debonding. A genetic algorithm was linked to a finite element package to automate the iterative procedure and maximize the damage resistance of the panel in postbuckling. For a given loading condition, the procedure optimized the panel’s skin layup leading to a design displaying superior damage resistance compared to non-optimized designs

Optimising element choice for nanoparticle radiosensitisers

There is considerable interest in the use of heavy atom nanoparticles as theranostic contrast agents due to their high radiation cross-section compared to soft tissue. However, published studies have primarily focused on applications of gold nanoparticles. This study applies Monte Carlo radiation transport modelling using Geant4 to evaluate the macro- and micro-scale radiation dose enhancement following X-ray irradiation with both imaging and therapeutic energies on nanoparticles consisting of stable elements heavier than silicon. An approach based on the Local Effect Model was also used to assess potential biological impacts. While macroscopic dose enhancement is well predicted by simple absorption cross-sections, nanoscale dose deposition has a much more complex dependency on atomic number, with local maxima around germanium (Z = 32) and gadolinium (Z = 64), driven by variations in secondary Auger electron spectra, which translate into significant variations in biological effectiveness. These differences may provide a valuable tool for predicting and elucidating fundamental mechanisms of these agents as they move towards clinical application.

FPGA Soft-core Processors, Compiler and Hardware Optimizations validated using HOG

There is demand for an easily programmable, high performance image processing platform based on FPGAs. In previous work, a novel, high performance processor - IPPro was developed and a Histogram of Orientated Gradients (HOG) algorithm study undertaken on a Xilinx Zynq platform. Here, we identify and explore a number of mapping strategies to improve processing efficiency for soft-cores and a number of options for creation of a division coprocessor. This is demonstrated for the revised high definition HOG implementation on a Zynq platform, resulting in a performance of 328 fps which represents a 146% speed improvement over the original realization and a tenfold reduction in energy.

Optimising the management of polycystic kidney disease

Polycystic kidney disease (PKD) is the most common inherited renal disorder that results in chronic kidney disease. Clinical features include visible haematuria, loin pain, UTI and hypertension. The typical clinical course is a progressive increase in the number and size of renal cysts associated with gradual loss of kidney function (falling eGFR).

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.