Despite the progress made in developing more advanced compilers for embedded systems, programming of embedded high-performance computing systems based on Digital Signal Processors (DSPs) is still a highly skilled manual task. This is true for single-processor systems, and even more for embedded systems based on multiple DSPs. Compilers often fail to optimise existing DSP codes written in C due to the employed programming style. Parallelisation is hampered by the complex multiple address space memory architecture, which can be found in most commercial multi-DSP configurations. This thesis develops an integrated optimisation and parallelisation strategy that can deal with low-level C codes and produces optimised parallel code for a homogeneous multi-DSP architecture with distributed physical memory and multiple logical address spaces. In a first step, low-level programming idioms are identified and recovered. This enables the application of high-level code and data transformations well-known in the field of scientific computing. Iterative feedback-driven search for “good” transformation sequences is being investigated. A novel approach to parallelisation based on a unified data and loop transformation framework is presented and evaluated. Performance optimisation is achieved through exploitation of data locality on the one hand, and utilisation of DSP-specific architectural features such as Direct Memory Access (DMA) transfers on the other hand. The proposed methodology is evaluated against two benchmark suites (DSPstone & UTDSP) and four different high-performance DSPs, one of which is part of a commercial four processor multi-DSP board also used for evaluation. Experiments confirm the effectiveness of the program recovery techniques as enablers of high-level transformations and automatic parallelisation. Source-to-source transformations of DSP codes yield an average speedup of 2.21 across four different DSP architectures. The parallelisation scheme is – in conjunction with a set of locality optimisations – able to produce linear and even super-linear speedups on a number of relevant DSP kernels and applications.

published_or_final_version / Chemistry / Doctoral / Doctor of Philosophy

published_or_final_version / Computer Science and Information Systems / Doctoral / Doctor of Philosophy

Finite element analysis is versatile and used widely in a range of engineering andscientific disciplines. As time passes, the problems that engineers and designers areexpected to solve are becoming more computationally demanding. Often theproblems involve the interplay of two or more processes which are physically andtherefore mathematically coupled. Although parallel computers have been availablefor about twenty years to satisfy this demand, finite element analysis is still largelyexecuted on serial machines. Parallelisation appears to be difficult, even for thespecialist. Parallel machines, programming languages, libraries and tools are used toparallelise old serial programs with mixed success. In some cases the serialalgorithm is not naturally suitable for parallel computing. Some argue that rewritingthe programs from scratch, using an entirely different solution strategy is a betterapproach. Taking this point of view, using MPI for portability, a mesh free elementby element method for simple data distribution and the appropriate iterative solvers,a general parallel strategy for finite element analysis is developed and assessed.

This study was conducted to set up a systematic approach utilizing advantages of both wet lab and bioinformatic methodologies to study biodegradation abilities and microbial bacterial-functional relationship within bioremediation communities. Firstly, 11pharmaceuticals and personal care products (PPCPs)were selected as target chemicals for establishing an effective determination process in analyzing trace-level concentrations in the environment, and understanding the removal routes during pollutants removal process in wastewater treatment process using activated sludge. Ultra performance liquid chromatography-tandem mass spectrometry was utilized to develop a rapid, sensitive and reliable method without solid phase extraction pre-concentration for trace analysis of 11 PPCPs in influent and effluent from municipal wastewater treatment plants. Shorten the detection time and significant reduction of detection cost were achieved due to the omitting usage of solid phase extraction (SPE)process and avoiding the consumption of hydrophiliclipophilic balancced (HLB)cartridge. Research on removal routes of ten selected PPCPs in activated sludge found activated sludge hardly removed carbamazepine. Biodegradation was the sole route to remove acyclovir, metronidazole, benzylparaben, ethylparaben, methylparaben and propylparaben. Both adsorption and biodegradation were involved in the removal of ranitidine and benzophenone-3, while fluoxetine could be totally removed by adsorption in activated sludge. Secondly, as the target microbial community, activated sludge community was used to set up the global bioinformatic analysis process. Both metagenomic and metatranscriptomic approaches were processed to characterize microbial structure and gene expression of activated sludge community. The taxonomic profile showed thatactivated sludge was dominated by Proteobacteria, Actinobacteria, Bacteroidetes, Firmicutes and Verrucomicrobiaphyla. Gene expression annotation of nitrogen removal revealed that denitrification-related genes sequences dominated in both DNA and cDNA datasets while nitrifying genes were also expressed in relative high levels. Specially, ammonia monooxygenase and hydroxylamine oxidase demonstrated the high cDNA/DNA ratios, indicating strong nitrification activity. Ammonia-oxidizing bacteria present mainly belonged to Nitrosomonas and Nitrosospira species. A fast method to construct local sub-databases has been established for the quick similarity search and annotation of huge metagenomic datasets. The conducted tests showed sub-database annotation pipeline achieved a speedup of ~150-385 times, and got exactly the same annotation results with those of the direct NCBI-nr database BLAST-MEGAN method. This approach provides a new time-efficient and convenient annotation similarity search strategy for laboratories without access to high performance computing facilities. Thirdly, bisphenol A(BPA), which has a partially known biodegradation pathway and relevant bioremediating genes, was chosen as a model to establish a pipeline for systematical understanding the pathways and gene/bacteria relationships in an enriched microbial community. 11 new metabolites were detected during BPA degradation. Thereby, a novel pathway of degrading BPA metabolite was proposed. Sphingomonas strains were dominant taxa in initial degradation of BPA, while the other taxa were competing BPA metabolites during degradation. Metagenomic binning results showed a cytochrome P450 monooxygenase system, which was previously reported BPA mediator, was sharing by two Sphingomonas strains, showing the undergoing mechanism of competition of the two strains. The observations suggested bacterial specialization may occur in that community that each taxon was selected to degrade certain metabolite in a community economical way. / published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Energy efficiency is of the utmost importance in modern high-performance embedded processor design. As the number of transistors on a chip continues to increase each year, and processor logic becomes ever more complex, the dynamic switching power cost of running such processors increases. The continual progression in fabrication processes brings a reduction in the feature size of the transistor structures on chips with each new technology generation. This reduction in size increases the significance of leakage power (a constant drain that is proportional to the number of transistors). Particularly in embedded devices, the proportion of an electronic product’s power budget accounted for by the CPU is significant (often as much as 50%). Dynamic branch prediction is a hardware mechanism used to forecast the direction, and target address, of branch instructions. This is essential to high performance pipelined and superscalar processors, where the direction and target of branches is not computed until several stages into the pipeline. Accurate branch prediction also acts to increase energy efficiency by reducing the amount of time spent executing mis-speculated instructions. ‘Stalling’ is no longer a sensible option when the significance of static power dissipation is considered. Dynamic branch prediction logic typically accounts for over 10% of a processor’s global power dissipation, making it an obvious target for energy optimisation. Previous approaches at increasing the energy efficiency of dynamic branch prediction logic has focused on either fully dynamic or fully static techniques. Dynamic techniques include the introduction of a new cache-like structure that can decide whether branch prediction logic should be accessed for a given branch, and static techniques tend to focus on scheduling around branch instructions so that a prediction is not needed (or the branch is removed completely). This dissertation explores a method of combining static techniques and profiling information with simple hardware support in order to reduce the number of accesses made to a branch predictor. The local delay region is used on unconditional absolute branches to avoid prediction, and, for most other branches, Adaptive Branch Bias Measurement (through profiling) is used to assign a static prediction that is as accurate as a dynamic prediction for that branch. This information is represented as two hint-bits in branch instructions, and then interpreted by simple hardware logic that bypasses both the lookup and update phases for appropriate branches. The global processor power saving that can be achieved by this Combined Algorithm is around 6% on the experimental architectures shown. These architectures are based upon real contemporary embedded architecture specifications. The introduction of the Combined Algorithm also significantly reduces the execution time of programs on Multiple Instruction Issue processors. This is attributed to the increase achieved in global prediction accuracy.

Approved for public release; distribution is unlimited / S tasks. Discrete Event Simulation (DES) often involves repeated, independent runs of the same models with different input parameters. A system which is able to run many replications quickly is more useful than one in which a single monolithic application runs quickly. A loosely coupled parallel system is indicated. Inexpensive commodity hardware, high speed local area networking, and open source software have created the potential to create just such loosely coupled parallel systems. These systems are constructed from Linux-based computers and are called Beowulf clusters. This thesis presents an analysis of clusters in high-performance computing and establishes a testbed implementation at the MOVES Institute. It describes the steps necessary to create a cluster, factors to consider in selecting hardware and software, and describes the process of creating applications that can run on the cluster. Monitoring the running cluster and system administration are also addressed. / Major, Hellenic Army

A mini-dissertation in partial compliance with the requirements for a Master's Degree in Technology: Homoeopathy, Durban Institute of Technology, 2002. / The purpose of this study was to compare the amount of caffeine extracted from triturated samples and ethanolic samples of Coffea tosta using high pressure liquid chromatography (HPLC) as a method of analysis. The study wanted to expand on homoeopathic pharmaceutical knowledge, specifically looking at the two methods of remedy preparation of plant materials. From the same batch of ground roasted coffee beans, using the decimal scale of dilution, the mother tincture (bill) and the first triturated (bill) samples were prepared. The subsequent 2xH and 3xH triturated and ethanolic potencies were then made in accordance with homoeopathic methodology. Each group contained three different dilution levels (bill, 2xH and 3xH), 18 samples per group giving a total of36 samples that were analysed using HPLC. Three samples were analysed from the three dilution levels in each Group, in total there were 18 samples from the triturated group and 18 from the ethanolic group. . The samples were analysed quantitatively using the highly accurate and advanced method of high pressure liquid chromatography. This method gives accurate readings of the caffeine concentrations of a sample compared to a caffeine standard. This allowed for quantification of the caffeine concentration of each sample. The percentage caffeine was calculated from each sample. The aim of the study was to evaluate the difference in each method of preparation by measuring the amount of caffeine extracted from the samples. The results obtained from the inter-Group Mann-Whitney and ANOVA tests showed that there was a significant difference between the ethanolic dilutions and triturated dilutions with regards to the 1xH and 2xH dilutions. In the 1xH dilution the ethanolic method retained / M

Doctor of Philosophy / Department of Industrial & Manufacturing Systems Engineering / Zhijian Pei / High-performance materials, such as silicon, aerospace stainless steels, titanium alloys, and carbon fiber reinforced plastic (CFRP) composites, have a variety of engineering applications. However, they usually have poor machinability and are classified as hard-to-machine materials. Drilling is one of the important machining processes for these materials. Industries are always under tremendous pressure to meet the ever-increasing demand for lower cost and better quality of the products made from these high-performance materials. Rotary ultrasonic machining (RUM) is a non-traditional machining process that combines the material removal mechanisms of diamond grinding and ultrasonic machining. It is a relatively low-cost, environment-benign process that easily fits in the infrastructure of the traditional machining environment. Other advantages of this process include high hole accuracy, superior surface finish, high material removal rate, low tool pressure, and low tool wear rate. The goal of this research is to provide new knowledge of machining these high performance materials with RUM for further improvement in the machined hole quality and decrease in the machining cost. A thorough research in this dissertation has been conducted by experimental, numerical, and theoretical investigations on output variables, including cutting force, torque, surface roughness, tool wear, cutting temperature, material removal rate, edge chipping (for silicon), power consumption (for CFRP), delamination (for CFRP), and feasible regions (for dry machining of CFRP). In this dissertation, an introduction of workpiece materials and RUM are discussed first. After that, two literature reviews on silicon drilling and dry drilling are presented. Then, design of experiment and finite element analysis on edge chipping in RUM of silicon, experimental investigations and finite element analysis on RUM of aerospace stainless steels, an ultrasonic vibration amplitude measurement method and a cutting temperature measurement method for RUM using titanium alloys as workpiece, experimental and theoretical investigations on RUM of CFRP composites, and experimental studies on CFRP/Ti stacks are presented, respectively. Finally, conclusions and contributions on RUM drilling are discussed.

O uso de concreto de alto desempenho constitui tendência irreversível, devido às vantagens que apresenta em relação aos concretos de resistência usual, principalmente na execução de pilares de edifícios, pois as áreas das seções transversais podem ser reduzidas. Este trabalho trata da obtenção e do estudo teórico e experimental de pilares de concreto de alto desempenho, analisando o comportamento de modelos, solicitados à compressão centrada e flexão normal composta. Foram analisados pilares de seções transversais quadradas de 20cm x 20cm e retangulares de 15cm x 30cm e 12cm x 30cm, dimensões que representam pilares usuais em estruturas de edifícios. Na compressão centrada ficou caracterizado que o estado limite último foi atingido por ruptura da seção mais solicitada e comprovado que a ruptura ocorreu quando o núcleo, definido pelo perímetro considerado pelos eixos dos estribos, se rompeu. Nos ensaios de flexão normal composta os resultados mostraram que as hipóteses de distribuição de tensões, utilizadas para os concretos de resistências usuais, não devem ser consideradas para concreto de alta resistência. As forças normais experimentais e as teóricas, calculadas por um modelo proposto pelo Autor, ficaram iguais enquanto que, para os momentos fletores, os valores experimentais ficaram muito acima dos teóricos. / The use of high performance concrete is an irreversible tendency in structural design due to the advantages it presents, compared to the concrete of usual strength, these advantages can be seen mainly in the execution of buildings columns as the cross sections can be reduced. This thesis is about the theoretical and experimental study to obtain high performance concrete columns. The behaviour of models subjected to axial load and bending plus axial load was analysed. Columns with the square (20cm x 20cm) and rectangular (15cm x 30cm and 12cm x 30cm) cross sections were studied, because they are the usual buildings columns dimensions. As far as axial load is concerned, the ultimate state limit was reached when the mostly demanded section broke. It was proven that the rupture occurred when the core, defined by the perimeter considered by the axial stirrups, broke. The results of bending plus axial load tests have shown that the distribution of stress hypotheses used for concrete of usual strength must not be considered for high strength concrete. The experimental axial forces and the internal ones calculated by a given model were the same, whereas the experimental values for the internal moments were much higher.