Deployment on GPUs of an application in computational atomic physics

This paper describes the deployment on GPUs of PROP, a program of the 2DRMP suite which models electron collisions with H-like atoms and ions. Because performance on GPUs is better in single precision than in double precision, the numerical stability of the PROP program in single precision has been studied. The numerical quality of PROP results computed in single precision and their impact on the next program of the 2DRMP suite has been analyzed. Successive versions of the PROP program on GPUs have been developed in order to improve its performance. Particular attention has been paid to the optimization of data transfers and of linear algebra operations. Performance obtained on several architectures (including NVIDIA Fermi) are presented.

Scale of production and choice of technique in the engineering industries in developing countries

This article examines the relationship between scale of production, optimal choice of technique and costs for three engineering industries: nuts and bolts, iron founding and machine tools. In all three industries costs of production fell as the scale of output increased. This was associated with switches of technique and the spread of fixed costs over a larger number of units. The capital-output ratio fell and labour productivity increased with increases in scale while, in most cases, the capital-labour ratio increased. The implications of these findings are briefly discussed.

Investigation of precision grinding process for production of silicon diaphragms

The application of precision grinding for the formation of a silicon diaphragm is investigated. The test structures involved 2-6 mm diam diaphragms with thicknesses in the range of 25-150 //m. When grinding is performed without supporting the diaphragm, bending occurs due to nonuniform removal of the silicon material over the diaphragm region. The magnitude of bending depends on the µNal thickness of the diaphragm. The results demonstrate that the use of a porous silicon support can significantly reduce the amount of bending, by a factor of up to 300 in the case of 50 m thick diaphragms. The use of silicon on insulator (SOI) technology can also suppress or eliminate bending although this may be a less economical process. Stress measurements in the diaphragms were performed using x-ray and Raman spectroscopies. The results show stress of the order of 1 X107-! X108 Pa in unsupported and supported by porous silicon diaphragms while SOI technology provides stress-free diaphragms. Results obtained from finite element method analysis to determine deterioration in the performance of a 6 mm diaphragm due to bending are presented. These results show a 10% reduction in performance for a 75 µm thick diaphragm with bending amplitude of 30 fim, but negligible reduction if the bending is reduced to

Learning soft computing control strategies in a modular neural network architecture

Modelling and control of nonlinear dynamical systems is a challenging problem since the dynamics of such systems change over their parameter space. Conventional methodologies for designing nonlinear control laws, such as gain scheduling, are effective because the designer partitions the overall complex control into a number of simpler sub-tasks. This paper describes a new genetic algorithm based method for the design of a modular neural network (MNN) control architecture that learns such partitions of an overall complex control task. Here a chromosome represents both the structure and parameters of an individual neural network in the MNN controller and a hierarchical fuzzy approach is used to select the chromosomes required to accomplish a given control task. This new strategy is applied to the end-point tracking of a single-link flexible manipulator modelled from experimental data. Results show that the MNN controller is simple to design and produces superior performance compared to a single neural network (SNN) controller which is theoretically capable of achieving the desired trajectory. (C) 2003 Elsevier Ltd. All rights reserved.

High-solid-content hydroxyapatite slurry for the production of bone substitute scaffolds

Key to various bone substitute scaffold production techniques is the development of free-flowing ceramic slurry with optimum theological properties. The aim is to achieve a colloidal suspension with as high a solid content as possible while maintaining a low viscosity which easily penetrates the pores of relevant sacrificial templates. The following investigation describes the optimization of a hydroxyapatite slip and demonstrates its potential application in scaffold production. Using predominantly spherical particles of hydroxyapatite of between 0.82 mu m and 16.2 mu m, coupled with a 2 wt % addition of the anionic polyelectrolyte, ammonium polyacrylate, an 80 wt % (55.9 vol %) hydroxyapatite solid loaded slip with a viscosity of approximately 126 mPa s has been developed. Its ability to infiltrate and replicate porous preforms has been shown using polyurethane foam. The enhanced particle packing achieved has allowed for the production of scaffolds with highly dense and uniform grain structures. The results represent a significant improvement in current slurry production techniques and can be utilized to develop high-density ceramic bone substitute scaffolds.