This dissertation presents a wide-band lumped element equivalent circuit model and a building block-based scalable circuit model for multiple quantum well laser diodes. The wide-band multiple-resonance model expresses two important laser diode characteristics such as input reflection and electrical-to-optical transmission together. Additionally, it demonstrates good agreements with the measurement results of the selected commercial discrete laser diodes. The proposed building block-based modeling approach proves its validity using a numerically derived scalable rate equation. Since success in a circuit design depends largely on the availability of accurate device models, the practical application of the proposed models provides improved accuracy, simple implementation and a short design time.

FPGA devices are an important component in many modern devices. This means that it is important that VLSI designers have a thorough knowledge of how to optimize designs for FPGAs. While the design flows for ASICs and FPGAs are similar, there are many differences as well due to the limitations inherent in FPGA devices. To be able to use an FPGA efficiently it is important to be aware of both the strengths and oweaknesses of FPGAs. If an FPGA design should be ported to an ASIC at a later stage it is also important to take this into account early in the design cycle so that the ASIC port will be efficient. This thesis investigates how to optimize a design for an FPGA through a number of case studies of important SoC components. One of these case studies discusses high speed processors and the tradeoffs that are necessary when constructing very high speed processors in FPGAs. The processor has a maximum clock frequency of 357~MHz in a Xilinx Virtex-4 devices of the fastest speedgrade, which is significantly higher than Xilinx' own processor in the same FPGA. Another case study investigates floating point datapaths and describes how a floating point adder and multiplier can be efficiently implemented in an FPGA. The final case study investigates Network-on-Chip architectures and how these can be optimized for FPGAs. The main focus is on packet switched architectures, but a circuit switched architecture optimized for FPGAs is also investigated. All of these case studies also contain information about potential pitfalls when porting designs optimized for an FPGA to an ASIC. The focus in this case is on systems where initial low volume production will be using FPGAs while still keeping the option open to port the design to an ASIC if the demand is high. This information will also be useful for designers who want to create IP cores that can be efficiently mapped to both FPGAs and ASICs. Finally, a framework is also presented which allows for the creation of custom backend tools for the Xilinx design flow. The framework is already useful for some tasks, but the main reason for including it is to inspire researchers and developers to use this powerful ability in their own design tools.

<div> Due to a combination of algorithmic advances, wide-spread availability of rich data sets, and tremendous growth in compute availability, Deep Neural Networks (DNNs) have seen considerable success in a wide variety of fields, achieving state-of-the art accuracy in a number of perceptual domains, such as text, video and audio processing. Recently, there have been many efforts to extend this success in the perceptual, Euclidean-based domain to non-perceptual tasks, such as task planning or reasoning, as well as to non-Euclidean domains, such as graphs. While several DNN accelerators have been proposed in the past decade, they largely focus on traditional DNN workloads, such as Multi-layer Perceptions (MLPs), Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs). These accelerators are ill-suited to the unique computational needs of the emerging neural networks. In this dissertation, we aim to fix this gap by proposing novel hardware architectures that are specifically tailored to emerging neural workloads.</div><div><br></div><div>First, we consider memory-augmented neural networks (MANNs), a new class of neural networks that exhibits capabilities such as one-shot learning and task planning that are well beyond those of traditional DNNs. MANNs augment a traditional DNN with an external differentiable memory that is used to store dynamic state. This dissertation proposes a novel accelerator that targets the main bottleneck of MANNs: the soft reads and writes to this external memory, each of which requires access to all the memory locations.</div><div><br></div><div>We then focus on Transformer networks, which have become very popular for Natural Language Processing (NLP). A key to the success of these networks is a technique called self-attention, which employs a softmax operation. Softmax is poorly supported in modern, matrix multiply-focused accelerators since it accounts for a very small fraction of traditional DNN workloads. We propose a hardware/software co-design approach to realize softmax efficiently by utilize a suite of approximate computing techniques.</div><div><br></div><div>Next, we address graph neural networks (GNNs). GNNs are achieving state-of-the-art results in a variety of fields such as physics modeling, chemical synthesis, and electronic design automation. These GNNs are a hybrid between graph processing workloads and DNN workloads; they utilize DNN-based feature extractors to form hidden representations for each node in a graph and then combine these representations through some form of a graph traversal. As a result, existing hardware specialized for either graph processing workloads or DNN workloads is insufficient. Instead, we design a novel architecture that balances the needs of these two heterogeneous compute patterns. We also propose a novel feature dimension-blocking dataflow to further increase performance by mitigating the memory bottleneck.</div><div><br></div><div>Finally, we address the growing difficulty in tightly coupling new DNNs and a hardware platform. Given the extremely large DNN-HW design space consisting of DNN selection, hardware operating condition, and DNN-to-HW mapping, it is infeasible to exhaustively search this space by running each sample on a physical hardware device. This has led to the need for highly accurate, machine learning-based performance models which can \emph{predict} the latency/power/energy even faster than direct execution. We first present a taxonomy to characterize the possible approaches to these performance estimators. Based on the insights from this taxonomy, we present a new performance estimator that combines coarse-grained and fine-grained to achieve superior accuracy with a limited number of training samples. Finally, we propose a flexible framework for creating these DNN-HW performance estimators.</div><div><br></div><div>In summary, this dissertation identifies the growing gap between current hardware and new emerging neural networks. We first propose three novel hardware architectures that address this gap for MANNs, Transformers, and GNNs. We then propose a novel hardware-aware DNN estimator and framework to ease addressing this gap for new networks in the future.</div>

This work describes a novel approach of HW/SW codesign on the Xilinx Zynq and similar platforms. It deals with interconnections between the Processing System (ARM Cortex-A9 MPCore) and the Programmable Logic (FPGA) to find an abstract and universal way to develop applications that are partially offloaded into the programmable hardware and that run in the Linux operating system. For that purpose a framework for HW/SW codesign on the Zynq and similar platforms is designed. No such framework is currently available.

This dissertation explores the impact of providing platforms for school learners to positively express themselves through multiple educational activities, in turn, developing a collective identity through performance. The existing school typologies in Mamelodi East, Pretoria, are found to hinder such explorative and all-round inclusive activities, resulting in the exclusion of many types of learners. Through a filter of music-making, the architecture of Tsako Thabo Secondary School in Mamelodi East has been assessed to have inadequate spaces and cause a disconnect with the community. Identified through literature, there is an opportunity to investigate the connection between music-making, participation and architecture as a way of transforming both the social and physical aspects of the school and surrounding community. The proposed intervention networks between existing nodes of music-making. With a focus on the node of Tsako Thabo Secondary School, an expressive architectural language is used to provide platforms for inclusive and adaptable educational and social activities. This language ripples out through the built fabric of the community and creates a sense of identity through the intersection of music and architecture. This dissertation uses participatory research in order to identify real world problems and social agents of change within the community. Furthering participation, codesign methods are used to influence the iterative design process to work beyond the single author limitations and remain contextually and socially relevant throughout. This was largely affected by the COVID-19 pandemic, proving resilience and adaptability for the approach, relevant for current research methodologies. / Mini Dissertation (MArch (Prof))--University of Pretoria, 2020. / This dissertation is a result of the investigation and the UkuDoba research methodology developed in the NRF project, STINT: Stitching the City (2020). This was a collaborative effort between the University of Pretoria and Chalmers University in Gothenberg, Sweden. / Architecture / MArch (Prof) / Unrestricted

Engineers working internationally are increasingly concerned with the social impacts of their work. New frontiers in design show promise in helping practitioners address these concerns. One of these is codesign, a practice of making stakeholders co-decision-makers in the design process. Codesign has the potential to greatly improve the social sustainability of engineered products, but some concerns remain surrounding codesign’s practicability in engineering. I explore three such concerns: that conflicting institutional logics may undermine codesign’s collaborative aspirations, that codesign can perpetuate developmental idealism, and that codesign may insufficiently account for the needs and perspectives of marginalized populations. Through more than a year of ethnographic research, including dozens of interviews and hundreds of hours of observation, I examine the realities of codesign as it is carried out by a team of engineers in the Brazilian Amazon. I find that conflicting logics do undermine codesign at times, but that the engineers are still able to explore new tools and practices for socially sustainable engineering, even in times of codesign failure. I also find that the engineers are better equipped to respond to modernizing stakeholders than they are traditional ones. This may lead to the spread of developmental idealism and the further marginalization of disadvantaged groups.

Yes / Background: Experience-based co-design (EBCD) brings patients and staff together to co-design services. It is normally conducted in one organization which initiates and implements the process. We used the traditional EBCD method with a number of adaptations as part of a larger research study in the British National Health Service.Methods: The primary aim was to assess the feasibility and acceptability of conduct-ing research-initiated EBCD, to enhance intervention development prior to testing. As well as embedding the method in a research study, there were 3 further key adap-tations: (a) working across primary and secondary care sectors, (b) working on multi-ple sites and (c) incorporating theory-informed analysis.Results: We recruited four sites (covering both primary and secondary care) and, on each site, conducted the initial traditional EBCD meetings, with separate staff and patient groups—followed by a single joint patient-staff event, where four priority areas for co-design were agreed. This event was driven by theory-informed analysis, as well as the traditional trigger film of patient experiences. Each site worked on one priority area, and the four co-design groups met over 2-3 months to design prototype tools. A second joint event was held (not usually undertaken in single-site EBCD) where they shared and compared outputs. The research team combined elements of these outputs to create an intervention, now being tested in a cluster randomized controlled trial.Conclusions: EBCD can be successfully adapted for use across an entire patient pathway with multiple organizations and as part of a research process to identify an intervention for subsequent testing in a randomized trial. Our pragmatic approach used the patient experience to identify areas for improvement and co-designed an intervention which directly reflected patient priorities. / National Institute for Health Research programme ‘Improving the safety and continuity of medicines management at care transitions (ISCOMAT)’ RP‐PG‐0514‐2009.

As iniciativas comunitárias em assentamentos urbanos informais e áreas periféricas metropolitanas vêm resistindo por décadas questões de vulnerabilidade social por meio da cultura, arte, educação e outras práticas que resultam em engajamento social e transformações espaciais. Neste sentido, a presente pesquisa escolheu como recorte a atuação de um coletivo de ativistas culturais - ativos nos eixos \"Arte e Cultura\", \"Desenvolvimento Territorial\" e \"Comunicação\" - da organização não governamental Redes da Maré. Trata-se de uma iniciativa comunitária que atua desde 2007 em diversos espaços de uso coletivo do Complexo da Maré, um agrupamento de dezessete favelas localizadas na Zona Norte da cidade do Rio de Janeiro. Este estudo de base experimental qualitativa tem como objetivo apresentar aspectos relevantes levantados em três etapas de colaboração realizadas no Complexo da Maré nos últimos dois anos de pesquisa. A partir da vivência e da troca de experiências e aprendizados da designer pesquisadora em parceria com coordenadores, funcionários e participantes da iniciativa, o trabalho apoiou-se em abordagens do design social participativo, codesign e design anthropology. Com foco nas formas de improvisação e os modos de cooperação presentes em ações que permeiam as atividades da iniciativa, esta pesquisa buscou gerar aprendizados e reflexões para designers e arquitetos brasileiros no intuito de suscitar caminhos para processos de co-criação no espaço urbano por meio de abordagens processuais, experimentais e colaborativas. Além disso, a investigação almeja dar relevância às práticas populares e periféricas nos espaços de uso coletivo da metrópole e, paralelamente, enriquecer o campo de atuação do design no âmbito informal e urbano. / The community initiatives in informal urban settlements and metropolitan suburban areas have been resisting questions of social vulnerability by way of culture, art, education and other practices that result in social engagement and spatial transformation for decades. In that sense this study chose as its subject the action of a collective of cultural activists - active in the axis \"Culture and Art\", \"Territorial Development\" and \"Communication\" - of the non-governmental organization Redes da Maré. This community initiative acts since 2007 in diverse spaces of collective use in Complexo da Maré, a grouping of seventeen favelas located in the north side (Zona Norte) of the city of Rio de Janeiro. This practice-led research study has the objective of introducing relevant aspects that emerged from three phases of collaboration executed in Complexo da Maré in the past two years of research. From the day to day exchange of knowledge and living experiences of the design researcher partnering with coordinators, staff and participants of the initiative, the research was supported by social participatory design, codesign and design anthropology approaches. Focusing on forms of improvisation and modes of cooperation present in actions that permeate in the initiative\'s activities, this research study sought to generate learning and reflections for brazilian designers and architects with the purpose of opening pathways for co-creation processes in the urban space by way of procedural, experimental and collaborative approaches. Furthermore, the investigation aims to give relevance to popular and suburban practices in the metropolitan spaces of collective use, and in parallel enrich the field of action of informal and urban design.

Esta pesquisa analisou o processo de criação de concepts de projeto para estimular os hábitos nutricionais saudáveis na infância, na perspectiva do design para mudança de comportamento. O princípio deste campo do design é transformar a compreensão do comportamento humano em estratégias para a criação de soluções que impactem positivamente a vida dos indivíduos e a sociedade. A fim de atender aos seus objetivos, o estudo adotou a pesquisa-ação como abordagem metodológica, por meio da qual foi realizado um workshop de codesign. Através desta dinâmica foi possível estabelecer a interação colaborativa entre designers e usuárias (mães), no papel de principais educadoras nutricionais dos filhos, e a criação concepts para incentivar o comportamento alimentar saudável nas crianças. O Modelo de Design para o Comportamento Saudável (MDCS – LUDDEN; HEKKERT, 2014), criado para potencializar projetos focados no bem-estar do usuário, foi utilizado como suporte para a análise do processo projetual. No âmbito teórico, os resultados alcançados na pesquisa possibilitaram um avanço nos estudos sobre design para mudança de comportamento, assim como sobre o MDCS. A pesquisa permitiu ainda a aproximação entre princípios do codesign e do design para a mudança de comportamento, diante da lacuna identificada entre essas abordagens. Na perspectiva aplicada, este trabalho promoveu a associação entre conceitos da psicologia comportamental e do design aplicados ao contexto da nutrição infantil. A partir desta conexão foi possível gerar conhecimentos técnicos para o aperfeiçoamento dos processos projetuais focados na alimentação saudável na infância e para a prospecção de novas possibilidades de projeto para prevenção da obesidade. / This research analyzed the creation process of project concepts to stimulate the healthy nutritional habits in childhood, in the perspective of the design for behavior change. The principle of this field of design is to turn the understanding of human behavior into strategies for the project of solutions that impact positively the lives of individuals and the society itself. In order to reach its goals, the study adopted the action research methodology as procedural approach, through which a codesign workshop was held. This dynamic allowed the collaborative interactive between designers and users (mothers), as the main nutritional educators of their children, and the creation of concepts to encourage the healthy eating behavior in kids. The Design for Healthy Behavior Framework (DHBF), created to enhance projects focused on the well-being of the user, was used to support the analysis of the design process. In the theoretical scope, the results achieved in the research made possible an advance in the studies of design for behavior change, as well as on DHBF. The research also allowed the approach between the principles of codesign and design for behavior change, given the gap identified between these two approaches. In the applied perspective, this research promoted the association between behavior psychology and design concepts applied to the context of child nutrition. From this connection, it was possible to generate theoretical knowledge for the improvement of design processes focused in childhood healthy eating and to the prospecting of new possibilities of obesity prevention projects.

Esta pesquisa analisou o processo de criação de concepts de projeto para estimular os hábitos nutricionais saudáveis na infância, na perspectiva do design para mudança de comportamento. O princípio deste campo do design é transformar a compreensão do comportamento humano em estratégias para a criação de soluções que impactem positivamente a vida dos indivíduos e a sociedade. A fim de atender aos seus objetivos, o estudo adotou a pesquisa-ação como abordagem metodológica, por meio da qual foi realizado um workshop de codesign. Através desta dinâmica foi possível estabelecer a interação colaborativa entre designers e usuárias (mães), no papel de principais educadoras nutricionais dos filhos, e a criação concepts para incentivar o comportamento alimentar saudável nas crianças. O Modelo de Design para o Comportamento Saudável (MDCS – LUDDEN; HEKKERT, 2014), criado para potencializar projetos focados no bem-estar do usuário, foi utilizado como suporte para a análise do processo projetual. No âmbito teórico, os resultados alcançados na pesquisa possibilitaram um avanço nos estudos sobre design para mudança de comportamento, assim como sobre o MDCS. A pesquisa permitiu ainda a aproximação entre princípios do codesign e do design para a mudança de comportamento, diante da lacuna identificada entre essas abordagens. Na perspectiva aplicada, este trabalho promoveu a associação entre conceitos da psicologia comportamental e do design aplicados ao contexto da nutrição infantil. A partir desta conexão foi possível gerar conhecimentos técnicos para o aperfeiçoamento dos processos projetuais focados na alimentação saudável na infância e para a prospecção de novas possibilidades de projeto para prevenção da obesidade. / This research analyzed the creation process of project concepts to stimulate the healthy nutritional habits in childhood, in the perspective of the design for behavior change. The principle of this field of design is to turn the understanding of human behavior into strategies for the project of solutions that impact positively the lives of individuals and the society itself. In order to reach its goals, the study adopted the action research methodology as procedural approach, through which a codesign workshop was held. This dynamic allowed the collaborative interactive between designers and users (mothers), as the main nutritional educators of their children, and the creation of concepts to encourage the healthy eating behavior in kids. The Design for Healthy Behavior Framework (DHBF), created to enhance projects focused on the well-being of the user, was used to support the analysis of the design process. In the theoretical scope, the results achieved in the research made possible an advance in the studies of design for behavior change, as well as on DHBF. The research also allowed the approach between the principles of codesign and design for behavior change, given the gap identified between these two approaches. In the applied perspective, this research promoted the association between behavior psychology and design concepts applied to the context of child nutrition. From this connection, it was possible to generate theoretical knowledge for the improvement of design processes focused in childhood healthy eating and to the prospecting of new possibilities of obesity prevention projects.