Thesis (MEng)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Iterative error correcting codes such as LDPC codes have become prominent in modern forward error correction systems. A particular subclass of LDPC codes known as quasicyclic LDPC codes has been incorporated in numerous high speed wireless communication and video broadcasting standards. These standards feature multiple codes with varying codeword lengths and code rates and require a high throughput. Flexible hardware that is capable of decoding multiple quasi-cyclic LDPC codes is therefore desirable. This thesis investigates binary quasi-cyclic LDPC codes and designs a generic, flexible VHDL decoder. The decoder is further enhanced to automatically select the most likely decoder based on the initial a posterior probability of the parity-check equation syndromes. A software system is developed that generates hardware code for such a decoder based on a small user specification. The system is extended to provide performance simulations for this generated decoder. / AFRIKAANSE OPSOMMING: Iteratiewe foutkorreksiekodes soos LDPC-kodes word wyd gebruik in moderne voorwaartse foutkorreksiestelsels. ’n Subklas van LDPC-kodes, bekend as kwasisikliese LDPC-kodes, word in verskeie hoëspoed-kommunikasie- en video-uitsaaistelselstandaarde gebruik. Hierdie standaarde inkorporeer verskeie kodes van wisselende lengtes en kodetempos, en vereis hoë deurset. Buigsame apparatuur, wat die vermoë het om ’n verskeidenheid kwasisikliese LDPC-kodes te dekodeer, is gevolglik van belang. Hierdie tesis ondersoek binêre kwasisikliese LDPC-kodes, en ontwerp ’n generiese, buigsame VHDL-dekodeerder. Die dekodeerder word verder verbeter om outomaties die mees waarskynlike dekodeerder te selekteer, gebaseer op die aanvanklike a posteriori-waarskynlikheid van die pariteitstoetsvergelykings se sindrome. ’n Programmatuurstelsel word ontwikkel wat die fermware-kode vir so ’n dekodeerder genereer, gebaseer op ’n beknopte gebruikerspesifikasie. Die stelsel word uitgebrei om werksverrigting te simuleer vir die gegenereerde dekodeerder.

Generating code from software models is considered to be a new generation leap within software development methods. The objective of this M.Sc. project is to evaluate how different approaches to modelling and code generation affect embedded systems software development and propose recommendations on how to improve software development. Two product areas at Saab Surveillance EW Systems in Järfälla, Stockholm, are used as study objects. A research overview is made to highlight themes regarding modelling, code generation and software development in general. Based on these, interviews are held with system engineers and software developers at each product area, where they use different modelling and code generation approaches. The two development processes are described thoroughly. Challenges and advantages related to each area’s approach are investigated. Software development within product area A is affected by the product complexity and the larger scale of the development, including projects running over a longer time with more teams involved. Recommendations include enabling code generation by aligning it with other investments on process improvement and limiting the approach to generating some system components. Software developers within product area B can use full code generation, enabled by the limited product complexity. The product area is affected by software standards and external requirements on the process. Recommendations include extending the modelling approach to make it easier to trace functionality from system to software level. Conclusions are that both product areas can apply modelling and code generation to more software development activities to improve their respective development processes.

La compilation est une étape indispensable dans la création d'applications performantes.Cette étape autorise l'utilisation de langages de haut niveau et indépendants de la cible tout en permettant d'obtenir de bonnes performances.Cependant, de nombreux freins empêchent les compilateurs d'optimiser au mieux les applications.Pour les compilateurs statiques, le frein majeur est la faible connaissance du contexte d'exécution, notamment sur l'architecture et les données utilisées.Cette connaissance du contexte se fait progressivement pendant le cycle de vie de l'application.Pour tenter d'utiliser au mieux les connaissances du contexte d'exécution, les compilateurs ont progressivement intégré des techniques de génération de code dynamique.Cependant ces techniques ne se focalisent que sur l'utilisation optimale du matériel et n'utilisent que très peu les données.Dans cette thèse, nous nous intéressons à l'utilisation des données dans le processus d'optimisation d'applications pour GPU Nvidia.Nous proposons une méthode utilisant différents moments pour créer des bibliothèques adaptatives capables de prendre en compte la taille des données.Ces bibliothèques peuvent alors fournir les noyaux de calcul les plus adapté au contexte.Sur l'algorithme de la GEMM, la méthode permet d'obtenir des gains pouvant atteindre 100~\% tout en évitant une explosion de la taille du code.La thèse s'intéresse également aux gains et coûts de la génération de code lors de l'exécution, et ce du point de vue de la vitesse d'exécution, de l'empreinte mémoire et de la consommation énergétique.Nous proposons et étudions 2 approches de génération de code à l'exécution permettant la spécialisation de code avec un faible surcoût.Nous montrons que ces 2 approches permettent d'obtenir des gains en vitesse et en consommation comparables, voire supérieurs, à LLVM mais avec un coût moindre. / Compilation is an essential step to create efficient applications.This step allows the use of high-level and target independent languages while maintaining good performances.However, many obstacle prevent compilers to fully optimize applications.For static compilers, the major obstacle is the poor knowledge of the execution context, particularly knowledge on the architecture and data.This knowledge is progressively known during the application life cycle.Compilers progressively integrated dynamic code generation techniques to be able to use this knowledge.However, those techniques usually focuses on improvement of hardware capabilities usage but don't take data into account.In this thesis, we investigate data usage in applications optimization process on Nvidia GPU.We present a method that uses different moments in the application life cycle to create adaptive libraries able to take into account data size.Those libraries can therefore provide more adapted kernels.With the GEMM algorithm, the method is able to provide gains up to 100~\% while avoiding code size explosion.The thesis also investigate runtime code generation gains and costs from the execution speed, memory footprint and energy consumption point of view.We present and study 2 light-weight runtime code generation approaches that can specialize code.We show that those 2 approaches can obtain comparable, and even superior, gains compared to LLVM but at a lower cost.

No mercado altamente globalizado de hoje, a manutenção dos equipamentos tem se tornado um fator crucial para as empresas dos mais diversos segmentos. Técnicas de manutenção baseadas no nível de degradação dos equipamentos estão sendo preferidas em detrimento das técnicas tradicionais como manutenção corretiva e preventiva, e trazem benefícios como tempos de paradas reduzidos, tarefas de manutenção facilitadas e melhor gerenciamento de ativos. Com o desenvolvimento das técnicas de manutenção inteligente, os sistemas embarcados que comportarão estes algoritmos necessitarão cada vez mais de alta flexibilidade, combinada com alta velocidade de processamento e baixo consumo. Em outras palavras, eles tornam-se cada vez mais complexos, o que tem impacto direto no projeto destes sistemas. Neste contexto, a programação baseada em modelos em conjunto com a capacidade de geração automática de código para uma dada plataforma tem despertado grande interesse. O presente trabalho tem como objetivo realizar a análise dos espaços de projeto e também do desempenho de diferentes implementações para algoritmos de manutenção inteligente quando executados em hardware e software. A partir de implementações disponíveis nos ambientes MATLAB e LabVIEW™ de um sistema de manutenção inteligente chamado Watchdog Agent™, e utilizando ferramentas de geração automática de código, o desempenho dos sistemas de manutenção gerados é comparado usando-se parâmetros como tempo de execução e ocupação de memória ou da área do FPGA. Para os testes são utilizados dados de vibração coletados de uma bancada de testes composta por um atuador eletromecânico para válvulas. / In today’s highly globalized market, equipment maintenance has become a crucial factor for companies from several segments. Maintenance strategies based on equipment’s condition level are being preferred in place of traditional techniques such as corrective and preventive maintenance, and incur in benefits such as reduced downtime, facilitated maintenance tasks and better assets management. With the development of intelligent maintenance techniques, the embedded systems that will be used with such algorithms will need increasingly more flexibility, combined with high processing speed and low power consumption. In other words, they became increasingly more complex, what directly impact in their project. Within this context, model based engineering associated with automatic platform-specific code generation capabilities are of great interest. This work has as objective to perform a design space exploration by analyzing the performance of different implementations for intelligent maintenance algorithms when executed in hardware and software. Based on implementations available in MATLAB™ and LabVIEW™ environments of an intelligent maintenance system called Watchdog Agent, and using automatic code generation tools, the performance of the generated systems are compared using parameters such as execution time and memory or FPGA area occupation. For the validation tests, vibration data collected from a test bench composed by an electric mechanical actuator will be used.

No mercado altamente globalizado de hoje, a manutenção dos equipamentos tem se tornado um fator crucial para as empresas dos mais diversos segmentos. Técnicas de manutenção baseadas no nível de degradação dos equipamentos estão sendo preferidas em detrimento das técnicas tradicionais como manutenção corretiva e preventiva, e trazem benefícios como tempos de paradas reduzidos, tarefas de manutenção facilitadas e melhor gerenciamento de ativos. Com o desenvolvimento das técnicas de manutenção inteligente, os sistemas embarcados que comportarão estes algoritmos necessitarão cada vez mais de alta flexibilidade, combinada com alta velocidade de processamento e baixo consumo. Em outras palavras, eles tornam-se cada vez mais complexos, o que tem impacto direto no projeto destes sistemas. Neste contexto, a programação baseada em modelos em conjunto com a capacidade de geração automática de código para uma dada plataforma tem despertado grande interesse. O presente trabalho tem como objetivo realizar a análise dos espaços de projeto e também do desempenho de diferentes implementações para algoritmos de manutenção inteligente quando executados em hardware e software. A partir de implementações disponíveis nos ambientes MATLAB e LabVIEW™ de um sistema de manutenção inteligente chamado Watchdog Agent™, e utilizando ferramentas de geração automática de código, o desempenho dos sistemas de manutenção gerados é comparado usando-se parâmetros como tempo de execução e ocupação de memória ou da área do FPGA. Para os testes são utilizados dados de vibração coletados de uma bancada de testes composta por um atuador eletromecânico para válvulas. / In today’s highly globalized market, equipment maintenance has become a crucial factor for companies from several segments. Maintenance strategies based on equipment’s condition level are being preferred in place of traditional techniques such as corrective and preventive maintenance, and incur in benefits such as reduced downtime, facilitated maintenance tasks and better assets management. With the development of intelligent maintenance techniques, the embedded systems that will be used with such algorithms will need increasingly more flexibility, combined with high processing speed and low power consumption. In other words, they became increasingly more complex, what directly impact in their project. Within this context, model based engineering associated with automatic platform-specific code generation capabilities are of great interest. This work has as objective to perform a design space exploration by analyzing the performance of different implementations for intelligent maintenance algorithms when executed in hardware and software. Based on implementations available in MATLAB™ and LabVIEW™ environments of an intelligent maintenance system called Watchdog Agent, and using automatic code generation tools, the performance of the generated systems are compared using parameters such as execution time and memory or FPGA area occupation. For the validation tests, vibration data collected from a test bench composed by an electric mechanical actuator will be used.

abstract: Mobile apps have improved human lifestyle in various aspects ranging from instant messaging to tele-health. In the current app development paradigm, apps are being developed individually and agnostic of each other. The goal of this thesis is to allow a new world where multiple apps communicate with each other to achieve synergistic benefits. To enable integration between apps, manual communication between developers is needed, which can be problematic on many levels. In order to promote app integration, a systematic approach towards data sharing between multiple apps is essential. However, current approaches to app integration require large code modifications to reap the benefits of shared data such as requiring developers to provide APIs or use large, invasive middlewares. In this thesis, a data sharing framework was developed providing a non-invasive interface between mobile apps for data sharing and integration. A separate app acts as a registry to allow apps to register database tables to be shared and query this information. Two health monitoring apps were developed to evaluate the sharing framework and different methods of data integration between apps to promote synergistic feedback. The health monitoring apps have shown non-invasive solutions can provide data sharing functionality without large code modifications and manual communication between developers. / Dissertation/Thesis / M.S. Computer Science 2014

No mercado altamente globalizado de hoje, a manutenção dos equipamentos tem se tornado um fator crucial para as empresas dos mais diversos segmentos. Técnicas de manutenção baseadas no nível de degradação dos equipamentos estão sendo preferidas em detrimento das técnicas tradicionais como manutenção corretiva e preventiva, e trazem benefícios como tempos de paradas reduzidos, tarefas de manutenção facilitadas e melhor gerenciamento de ativos. Com o desenvolvimento das técnicas de manutenção inteligente, os sistemas embarcados que comportarão estes algoritmos necessitarão cada vez mais de alta flexibilidade, combinada com alta velocidade de processamento e baixo consumo. Em outras palavras, eles tornam-se cada vez mais complexos, o que tem impacto direto no projeto destes sistemas. Neste contexto, a programação baseada em modelos em conjunto com a capacidade de geração automática de código para uma dada plataforma tem despertado grande interesse. O presente trabalho tem como objetivo realizar a análise dos espaços de projeto e também do desempenho de diferentes implementações para algoritmos de manutenção inteligente quando executados em hardware e software. A partir de implementações disponíveis nos ambientes MATLAB e LabVIEW™ de um sistema de manutenção inteligente chamado Watchdog Agent™, e utilizando ferramentas de geração automática de código, o desempenho dos sistemas de manutenção gerados é comparado usando-se parâmetros como tempo de execução e ocupação de memória ou da área do FPGA. Para os testes são utilizados dados de vibração coletados de uma bancada de testes composta por um atuador eletromecânico para válvulas. / In today’s highly globalized market, equipment maintenance has become a crucial factor for companies from several segments. Maintenance strategies based on equipment’s condition level are being preferred in place of traditional techniques such as corrective and preventive maintenance, and incur in benefits such as reduced downtime, facilitated maintenance tasks and better assets management. With the development of intelligent maintenance techniques, the embedded systems that will be used with such algorithms will need increasingly more flexibility, combined with high processing speed and low power consumption. In other words, they became increasingly more complex, what directly impact in their project. Within this context, model based engineering associated with automatic platform-specific code generation capabilities are of great interest. This work has as objective to perform a design space exploration by analyzing the performance of different implementations for intelligent maintenance algorithms when executed in hardware and software. Based on implementations available in MATLAB™ and LabVIEW™ environments of an intelligent maintenance system called Watchdog Agent, and using automatic code generation tools, the performance of the generated systems are compared using parameters such as execution time and memory or FPGA area occupation. For the validation tests, vibration data collected from a test bench composed by an electric mechanical actuator will be used.

Made available in DSpace on 2014-06-12T15:49:38Z (GMT). No. of bitstreams: 1 license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2011 / Antonius de Castro Tavares, Marcellus; Pierre Louis Robin, Jacques. Behavioral java code generation from imperative object constraint language expressions in plataformindependent UML models. 2011. Dissertação (Mestrado). Programa de Pós-Graduação em Ciência da Computação, Universidade Federal de Pernambuco, Recife, 2011.

Made available in DSpace on 2014-06-12T15:49:47Z (GMT). No. of bitstreams: 1 license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2009 / A grande expansão do mercado de dispositivos digitais tem forçado empresas desenvolvedoras de sistemas embarcados em lidar com diversos desafios para prover sistemas complexos nesse nicho de mercado. Um dos desafios prominentes está relacionado ao consumo de energia, principalmente, devido aos seguintes fatores: (i) mobilidade; (ii) problemas ambientais; e (iii) o custo da energia. Como consequência, consideráveis esforços de pesquisa têm sido dedicados para a criação de técnicas voltadas para aumentar a economia de energia. Na última década, diversas técnicas foram desenvolvidas para reduzir o consumo de energia em sistemas embarcados. Muitos métodos lidam com gerenciamento dinâmico de energia (DPM), como, por exemplo, dynamic voltage scaling (DVS), cooperativamente com sistemas operacionais especializados, a fim de controlar o consumo de energia durante a execução do sistema. Entretanto, apesar da disponibilidade de muitos métodos de redução de consumo de energia, diversas questões estão em aberto, principalmente, no contexto de sistemas de tempo real crítico. Este trabalho propõe um método de síntese de software, o qual leva em consideração relação entre tarefas, overheads, restrições temporais e de energia. O método é composto por diversas atividades, as quais incluem: (i) medição; (ii) especificação; (iii) modelagem formal; (vi) escalonamento; e (v) geração de código. O método também é centrado no formalismo redes de Petri, o qual define uma base para geração precisa de escalas em tempo de projeto, adotando DVS para reduzir o consumo de energia. A partir de uma escala viável, um código customizado é gerado satisfazendo as restrições especificadas, e, dessa forma, garantindo previsibilidade em tempo de execução. Para lidar com a natureza estática das escalas geradas em tempo de projeto, um escalonador simples em tempo de execução é também proposto para melhorar o consumo de energia durante a execução do sistema. Diversos experimentos foram conduzidos, os quais demonstram a viabilidade da abordagem proposta para satisfazer restrições críticas de tempo e energia. Adicionalmente, um conjunto integrado de ferramentas foram desenvolvidas para automatizar algumas atividades do método de síntese de software proposto

The purpose of the thesis was to study the feasibility of using geneticalgorithm (GA) to do the integrated software pipelining (ISP). Different from phasedcode generation, ISP is a technique which integrates instruction selection, instructionscheduling, and register allocation together when doing code generation. ISP is able toprovide a lager solution space than phased way does, which means that ISP haspotential to generate more optimized code than phased code generation. However,integrated compiling costs more than phased compiling. GA is stochastic beam searchalgorithm which can accelerate the solution searching and find an optimized result.An experiment was designed for verifying feasibility of implementing GA for ISP(GASP). The implemented algorithm analyzed data dependency graphs of loop bodies,created genes for the graphs and evolved, generated schedules, calculated andevaluated fitness, and obtained optimized codes. The fitness calculation wasimplemented by calculating the maximum value between the smallest possibleresource initiation interval and the smallest possible recurrence initiation interval. Theexperiment was conducted by generating codes from data dependency graphsprovided in FFMPEG and comparing the performance between GASP and integerlinear programming (ILP). The results showed that out of eleven cases that ILP hadgenerated code, GASP performed close to ILP in seven cases. In all twelve cases thatILP did not have result, GASP did generate optimized code. To conclude, the studyindicated that GA was feasible of being implemented for ISP. The generated codesfrom GASP performed similar with the codes from ILP. And for the dependencygraphs that ILP could not solve in a limited time, GASP could also generate optimizedresults.