[pt] Nas instalações elétricas de baixa tensão de uma planta industrial está concentrado grande parte do custo necessário para sua construção. Nesse contexto, os cabos elétricos são o item mais relevante e poucas iniciativas visando sua redução têm sido notadas. Apesar dos limites impostos pelas normativas vigentes e pelos próprios usuários dessas plantas, existem medidas que podem levar a redução do custo dos cabos elétricos, entre elas, a alocação inteligente dos quadros de distribuição na baixa tensão. O objetivo desse trabalho é definir uma metodologia para o posicionamento ótimo desses quadros de distribuição, minimizando o custo dos cabos elétricos dentro de uma área onde as cargas a serem alimentadas estão posicionadas de modo fixo. São definidas algumas restrições ao posicionamento dos quadros e também é preparada uma interface gráfica que facilita a interpretação dos resultados. Faz-se então uma comparação dos resultados obtidos com dados reais de uma instalação industrial cujo projeto básico fora realizado da forma tradicional, sem uso de qualquer técnica inteligente para a alocação dos quadros de distribuição. Dessa comparação nota-se que o potencial de redução pode alcançar mais de 40 porcento do custo dos cabos elétricos previsto pelo projeto básico. A redução na quantidade de cabos elétricos trás vários efeitos colaterais positivos, entre eles a redução das perdas por efeito Joule e a redução nas emissões de CO2 cujos impactos são mensurados também. Por fim, são lançadas ideias para a evolução da metodologia de modo que sua aplicação seja mais abrangente e simples, preparando-a para o uso em qualquer situação, como uma nova ferramenta dos projetos elétricos. / [en] In electrical installations of low voltage of an industrial plant is concentrated much of the cost required for its construction. In this context, the electrical wires are the most important item and few initiatives aimed at their reduction has been noted. Despite the limitations imposed by current regulations and by the users of these plants, there are steps that can lead to reducing the cost of electric cables, among them the intelligent allocation of switchboards at low voltage. The aim of this study is to define a methodology for the optimal positioning of these distribution boards, minimizing the cost of electrical cables within an area where the loads to be fed are positioned permanently. A graphical interface that facilitates the interpretation of results are set some restrictions on placement of tables and is also prepared. Then it makes a comparison of the results with real data of an industrial installation whose basic design was done the traditional way, without using any smart technique for allocation of the switchboard. This comparison we note that the reduction potential can reach more than 40 percent of the cost of electric cables provided by the basic design. The reduction in the amount of electrical cables behind several positive side effects, including reducing losses by Joule effect and the reduction in CO2 emissions whose impacts are measured as well. Finally, ideas are thrown to the evolution of the methodology so that their application is more comprehensive and simple to prepare it for use in any situation, as a new tool of electrical projects.

Los cambios climáticos originados por el Fenómeno El Niño del 2017 han causado el colapso de una gran cantidad de puentes en Perú. La importancia de estas estructuras radica en las fuertes sumas de dinero que son empleadas para la construcción y mantenimiento, es por ello que se debe tomar importancia en los procesos constructivas para hacer un eficiente uso de la gran cantidad de dinero a emplear para materializar estos proyectos. En presente investigación se mostrarán la diferencia en costo y tiempo mediante una simulación con la metodología de Bridge Information Modeling (BrIM). Dicha metodologia integra información en un modelo visual el cual servirá como herramienta para planificación en la construcción de la superestructura de un puente. Mediante este enfoque se estudiarán la construcción de vigas de concreto presforzado y acero; y losa mediante vaciado in-situ y vaciado con prelosas. Asimismo, se diferenciarán los costos de mantenimiento que requiere cada alternativa, determinando así un sistema estructural optimiza el uso de los recursos económicos y el plazo para la construcción de la superestructura. Como resultado se obtiene el costo integral y tiempo constructivo de la superestructura ideal para el caso de estudio. / The climatic changes caused by the El Niño Phenomenon of 2017 have caused the collapse of a large number of bridges in Peru. Large sums of money are used for construction and maintenance, which is why importance should be taken in the construction processes to make efficient use of the large amount of money to be used to materialize these projects. This research will show the difference in cost and time through a simulation with the Bridge Information Modeling (BrIM) methodology. This methodology integrates information in a visual model which will serve as a tool for planning the construction of the superstructure of a bridge. This approach will study the construction of prestressed concrete and steel beams; and slabs by in-situ casting and precast slabs. Likewise, the maintenance costs required by each alternative will be differentiated, thus determining a structural system that optimizes the use of economic resources and the time frame for the construction of the superstructure. As a result, the integral cost and construction time of the ideal superstructure for the case study is obtained. / Tesis

In the modern era of IT, cloud computing is becoming the new standard. Companies have gone from owning their own data centers to procuring virtualized computational resources as a service. This technology opens up for elasticity and cost savings. Computational resources have gone from being a capital expenditure to an operational expenditure. Vendors, such as Google, Amazon, and Microsoft, offer these services globally with different provisioning alternatives. In this thesis, we focus on providing a cost optimization algorithm for Spotify on the Google Cloud Platform. To achieve this we  construct an algorithm that breaks up the problem in four different parts. Firstly, we generate trajectories of monthly active users. Secondly, we split these trajectories up in regions and redistribute monthly active users to better describe the actual Google Cloud Platform footprint. Thirdly we calculate usage per monthly active users quotas from a representative week of usage and use these to translate the redistributed monthly active users trajectories to usage. Lastly, we apply an optimization algorithm to these trajectories and obtain an objective value. These results are then evaluated using statistical methods to determine the reliability. The final model solves the problem to optimality and provides statistically reliable results. As a consequence, we can give recommendations to Spotify on how to minimize their cloud cost, while considering the uncertainty in demand.

In face of climate change, Europe and other global actors are in the process of transitioning to carbon-neutral economies, aiming to phase out of fossil fuels and power industries with renewable energies. Hydrogen is going to play a crucial role in the transition, replacing fossil fuels in hard-to-decarbonize industries and acting as energy carrier and energy storage for renewable electricity. However, the hydrogen production method with the lowest carbon intensity, water electrolysis in combination with renewable electricity, is often not cost competitive to other production methods. Even though policies and initiatives are providing subsidies to scale up low-carbon hydrogen production, companies hesitate to invest due to the complexity of hydrogen production systems and the uncertainties of cost competitiveness. This research aims to develop a tool for optimizing the capacity of a water electrolysis system to produce low-carbon hydrogen and to lay the groundwork for optimizing the operation of electrolysis hydrogen production plants. The objective is to find the optimal plant capacity to achieve the lowest cost of hydrogen production for a defined hydrogen demand and energy supply. The scope is limited to the electrolysis system as optimizing asset which is modeled with technology-specific costs and characteristics, gained from manufacturer interviews and internal company data. This includes the often neglected characteristics of load-dependent efficiency and degradation effects. Further, the tool is enabled to buy and sell electricity on the spot market according to predicted prices in order to minimize the electricity costs. The developed tool is tested, benchmarked and applied to two different industry-based test scenarios in Germany and Portugal. The test scenario in Germany describes a mid-scale hydrogen production case for a transport application with a demand increase over 10 years (80 to 1,800 tons per year) and regional renewable energy supply via power purchase agreements. The lowest costs of hydrogen production for this scenario can be reached with an alkaline electrolysis system of a capacity of 16 MWel considering only renewable energy sources, achieving a LCOH of 4.75 €/kg of green hydrogen. The second test scenario describes a large-scale production case in Portugal for application in the refinery industry. The yearly hydrogen demand increases from 5,000 tons up to 17,100 tons within three years and is assumed to stay constant for the residual years. The electricity for the electrolysis process is secured through large solar PV and offshore wind power purchase agreements. Utilizing the alkaline electrolysis technology with a capacity of 128 MWel, a LCOH of 3.31 €/kg of green hydrogen can be achieved at the output point of the plant. The study concludes that the optimal solution and the achievable hydrogen production costs are highly dependent on the hydrogen demand (quantity and profile), the energy supply (quantity, profile, costs), and the chosen technology (efficiency, degradation, costs) and need to be evaluated under the case-specific prerequisites. The thesis further highlights the significant impact of the electrolysis system efficiency and capital expenditures on the capacity decision and achievable hydrogen production costs. / Mot bakgrund av klimatförändringarna håller Europa och andra globala aktörer på att ställa om till koldioxidneutrala ekonomier, med målet att fasa ut fossila bränslen och driva industrier med förnybara energikällor. Vätgas kommer att spela en avgörande roll i omställningen genom att ersätta fossila bränslen i industrier som är svåra att koldioxidneutralisera och fungera som energibärare och energilagring för förnybar el. Den metod för vätgasproduktion som har lägst koldioxidintensitet, vattenelektrolys i kombination med förnybar el, är dock ofta inte kostnadsmässigt konkurrenskraftig i förhållande till andra produktionsmetoder. Även om politik och initiativ tillhandahåller subventioner för att skala upp koldioxidsnål vätgasproduktion, tvekar företagen på grund av komplexiteten i vätgasproduktionssystemen och osäkerheten kring konkurrenskraften. Denna forskning syftar till att utveckla ett verktyg för att optimera kapaciteten hos ett vattenelektrolyssystem för att producera grön vätgas och att lägga grunden för att optimera driften av elektrolysanläggningar för vätgasproduktion. Målet är att hitta den optimala anläggningskapaciteten för att uppnå den lägsta kostnaden för vätgasproduktion för en definierad vätgasefterfrågan och definierad energiförsörjning. Omfattningen är begränsad till elektrolyssystemet som en optimerande tillgång som modelleras med teknikspecifika kostnader och egenskaper, hämtade från tillverkar-intervjuer och från företags interna marknadsdata. Detta inkluderar de ofta försummade egenskaperna hos lastberoende effektivitet och degraderingseffekter. Vidare kan verktyget köpa och sälja el på spotmarknaden enligt förutspådda priser för att minimera elkostnaderna. Det utvecklade verktyget testas, jämförs och tillämpas på två olika industribaserade testscenarier i Tyskland och Portugal. Testscenariot i Tyskland beskriver en medelstor vätgasproduktion för en transporttillämpning där efterfrågan ökar över 10 år (80 till 1 800 ton per år) och regional förnybar energiförsörjning via energiköpsavtal (power purchase agreements). De lägsta kostnaderna för vätgasproduktion för detta scenario kan uppnås med ett alkaliskt elektrolyssystem med en kapacitet på 16 MWel som endast använder förnyelsebara energikällor och uppnår en LCOH på 4,75 €/kg grön vätgas. Det andra testscenariot beskriver en storskalig vätgasproduktion i Portugal för tillämpning inom raffinaderi-industrin. Det årliga vätgasbehovet ökas från 5 000 ton till 17 100 ton inom tre år och antogs förbli konstant under de återstående åren. El för elektrolysprocessen säkras genom stora energiköpsavtal (power purchase agreements) för solceller och havsbaserad vindkraft. Genom att använda alkalisk elektrolysteknik med en kapacitet på 128 MWel kan en LCOH på 3,31 €/kg grön vätgas uppnås vid anläggningens utgångspunkt. Studien visar att den optimala lösningen och de uppnåbara vätgasproduktionskostnaderna är starkt beroende av vätgasbehovet (mängd och profil), energiförsörjningen (mängd, profil, kostnader) och den valda tekniken (effektivitet, nedbrytning, kostnader) och måste utvärderas utifrån de fallspecifika förutsättningarna. Avhandlingen belyser vidare den betydande inverkan som elektrolyssystemets effektivitet och kapitalutgifter har på kapacitetsbeslutet och de uppnåeliga kostnaderna för vätgasproduktion.

Partial dynamic reconfiguration of FPGAs has attracted high attention from both academia and industry in recent years. With this technique, the functionality of the programmable devices can be adapted at runtime to changing requirements. The approach allows designers to use FPGAs more efficiently: E. g. FPGA resources can be time-shared between different functions and the functions itself can be adapted to changing workloads at runtime. Thus partial dynamic reconfiguration enables a unique combination of software-like flexibility and hardware-like performance. Still there exists no common understanding on how to assess the overhead introduced by partial dynamic reconfiguration. This dissertation presents a new cost model for both the runtime and the memory overhead that results from partial dynamic reconfiguration. It is shown how the model can be incorporated into all stages of the design optimization for reconfigurable hardware. In particular digital circuits can be mapped onto FPGAs such that only small fractions of the hardware must be reconfigured at runtime, which saves time, memory, and energy. The design optimization is most efficient if it is applied during high level synthesis. This book describes how the cost model has been integrated into a new high level synthesis tool. The tool allows the designer to trade-off FPGA resource use versus reconfiguration overhead. It is shown that partial reconfiguration causes only small overhead if the design is optimized with regard to reconfiguration cost. A wide range of experimental results is provided that demonstrates the benefits of the applied method. / Partielle dynamische Rekonfiguration von FPGAs hat in den letzten Jahren große Aufmerksamkeit von Wissenschaft und Industrie auf sich gezogen. Die Technik erlaubt es, die Funktionalität von progammierbaren Bausteinen zur Laufzeit an veränderte Anforderungen anzupassen. Dynamische Rekonfiguration erlaubt es Entwicklern, FPGAs effizienter einzusetzen: z.B. können Ressourcen für verschiedene Funktionen wiederverwendet werden und die Funktionen selbst können zur Laufzeit an veränderte Verarbeitungsschritte angepasst werden. Insgesamt erlaubt partielle dynamische Rekonfiguration eine einzigartige Kombination von software-artiger Flexibilität und hardware-artiger Leistungsfähigkeit. Bis heute gibt es keine Übereinkunft darüber, wie der zusätzliche Aufwand, der durch partielle dynamische Rekonfiguration verursacht wird, zu bewerten ist. Diese Dissertation führt ein neues Kostenmodell für Laufzeit und Speicherbedarf ein, welche durch partielle dynamische Rekonfiguration verursacht wird. Es wird aufgezeigt, wie das Modell in alle Ebenen der Entwurfsoptimierung für rekonfigurierbare Hardware einbezogen werden kann. Insbesondere wird gezeigt, wie digitale Schaltungen derart auf FPGAs abgebildet werden können, sodass nur wenig Ressourcen der Hardware zur Laufzeit rekonfiguriert werden müssen. Dadurch kann Zeit, Speicher und Energie eingespart werden. Die Entwurfsoptimierung ist am effektivsten, wenn sie auf der Ebene der High-Level-Synthese angewendet wird. Diese Arbeit beschreibt, wie das Kostenmodell in ein neuartiges Werkzeug für die High-Level-Synthese integriert wurde. Das Werkzeug erlaubt es, beim Entwurf die Nutzung von FPGA-Ressourcen gegen den Rekonfigurationsaufwand abzuwägen. Es wird gezeigt, dass partielle Rekonfiguration nur wenig Kosten verursacht, wenn der Entwurf bezüglich Rekonfigurationskosten optimiert wird. Eine Anzahl von Beispielen und experimentellen Ergebnissen belegt die Vorteile der angewendeten Methodik.

I Sverige står bostadssektorn för mer än en tredjedel av landets energianvändning. Byggnader måste minska sin energianvändning för att således kunna uppfylla framtida lagkrav om maximal tillåten energianvändning, men också för att minska påverkan till global uppvärmning. Ytterligare en problematik som råder, däribland i Sverige, är bostadsbrist. Kommunala bostadsbolag står inför utmaningen att kunna bygga bostäder snabbt, billigt och miljövänligt för att minska bostadsbristen i landet. Ett sätt att studera två av tre hållbarhetsaspekter vid val av framtida bostadsbyggande är att utföra en livscykelanalys (LCA) och livscykelkostnadsanalys (LCC) kring de tilltänkta husen. LCA:er indikerar vilken miljöpåverkan en produkt förorsakar under dess livslängd. LCC:er avser att studera vilka kostnader produkter ger upphov till under en given analysperiod. Det svenska kommunala bostadsbolaget Stångåstaden AB står inför utmaningen kring bostadsbrist och vill bygga hållbara bostäder. Bostadsbolaget har önskat en jämförande LCA och LCC av två verkliga flerbostadshus som de genom ramavtal kan upphandla, detta är utgångspunkten för denna studie. Den ena byggnaden har stomme av betong, den andra har stomme av trä. Husen är tänkta att placeras i utkanten av Linköping, Sverige. Studien har valt att analysera miljöpåverkan från husens olika livscykelfaser samt kostnader över analysperioden 50 år. Utöver detta studeras även vilka energieffektiviseringsåtgärder (EEÅ) till byggnaderna som är optimala att genomföra för att öka den termiska prestandan hos huskonstruktionerna. Från litteraturen finns det relativt få studier som kombinerar både LCA och LCC för vanligt förekommande hustyper i städer. I dess standardfall påvisade resultatet från LCA:n att huset med betongkonstruktion hade något lägre påverkan i sex av sju studerade miljöpåverkanskategorier, jämfört med flerbostadshuset i trä. Resultatet skilde sig lite åt då annan typ av indata användes. Vad gäller kostnader under husens livslängd var huset i trä ungefär 20 % dyrare jämfört med huset med betongkonstruktion. Trots annan typ av indata var träkonstruktionen dyrare än betongkonstruktionen. Med en kalkylränta på 7,5 % var det inte lönsamt att genomföra EEÅ för husen, med halverad kalkylränta blev det dock lönsamt att tilläggsisolera krypgrunden i huset med trästomme. Fler studier behöver utföras för att generalla slutsatser ska kunna dras kring vilket konstruktionsmaterial som är mest hållbart. Denna studie baseras på två specifika fall. Samma resultat kan eventuellt inte förväntas för andra byggnader med stomme i betong och trä. / The residential sector accounts for more than a third of the energy use in Sweden. To reduce the energy use of buildings is a necessity in order to meet future regulationof maximum allowable energy, but also important to reduce the impact on global warming. Another complexity arising in Sweden is the shortage of accommodation. Municipal housing corporations face the challenge of constructing residences fast, cheap and with concern of environmental effects in order to reduce the shortage of accommodation. One way of assessing two of the three aspects of sustainability when looking at future construction of residential buildings is to carry out a Life Cycle Assessment (LCA) and a Life Cycle Cost Assessment (LCCA). An LCA can indicate what kind of environmental impact a product causes over its lifetime and the LCC allows for assessing what types of costs are associated with the product. For the municipal housing corporation Stångåstaden AB the shortage of accommodation is a reality and their mindset is sustainable construction of residences. This study was conducted upon request from Stångåstaden who wanted a comparative LCA and LCCA between two prefabricated multi-residential buildings that are available to them through a framework agreement. The first building has a concrete foundation and the second one is made of wood. The houses are planned to be placed at the outskirts of Linköping, Sweden. The focus of this study has been to comparatively assess the environmental impact from the different life cycle phases and the economic costs of the two buildings during a time period of 50 years. Moreover, the thesis also analyze the optimal retrofit strategy for the buildings in order to find the optimal (lowest) life cycle cost. Furthermore, the current literature has conveyed relatively few studies that combine both LCA and LCC methodology for house types that are common in most towns. The result from the LCA indicated that the house with concrete construction had a little less impact in six of the seven studied environmental impact categories compared to the house made of wood. The result differed slightly when the input data were changed. Regarding the LCCA the house made of wood was roughly 20 % more expensive than its concrete counterpart. Changing the input data revealed no difference in the result. With an interest rate of 7,5 % no retrofits were profitable for either building, however reducing the interest rate to half its original value made it cost optimal to increase the floor insulation for the house made of wood. More studies should be conducted to be able to draw general conclusions regarding which construction material that is the most sustainable. This thesis is based on two specific and real cases. The same result could possibly not be expected from other studies comparing buildings with concrete and wood construction.

Aujourd’hui les réseaux électriques sont exploités dans un marché dérégulé. Les gestionnaires des réseaux électriques sont tenus d’assurer un certain nombre de critères de fiabilité et de continuité du service, tout en minimisant le coût total consacré aux efforts effectués pour maintenir la fiabilité des installations. Il s’agit de trouver une stratégie, qui répond à plusieurs exigences, comme :le coût, les performances, la législation, les exigences du régulateur, etc. Cependant, le processus de prise de décision est subjectif, car chaque participant ramène sa contribution sur base de sa propre expérience. Bien que ce processus permette de trouver la « meilleure » stratégie, cette dernière n’est pas forcément la stratégie « optimale ». Ce compromis technico-économique a sensibilisé les gestionnaires des réseaux électriques à la nécessité d’un recours à des outils d’aide à la décision, qui doivent se baser sur des nouvelles approches quantitatives et une modélisation plus proches de la réalité physique.<p>Cette thèse rentre dans le cadre d’un projet de recherche lancé par ELIA, et dénommé COMPRIMa (Cost-Optimization Models for the Planning of the Renewal, Inspection, and Maintenance of Belgian power system facilities). Ce projet vise à développer une méthodologie qui permet de modéliser une partie du réseau électrique de transport (par les réseaux de Petri stochastiques) et de simuler son comportement dynamique sur un horizon donné (simulation de Monte Carlo). L’évaluation des indices de fiabilité permet de comparer les différents scénarios qui tentent d’améliorer les performances de l’installation. L’approche proposée est basée sur la stratégie RCM (Reliability-Centered Maintenance).<p>La méthodologie développée dans cette thèse permet une modélisation plus réaliste du réseau qui tient compte, entre autres, des aspects suivants :<p>- La corrélation quantitative entre le processus de maintenance et le processus de vieillissement des composants (par un modèle d’âge virtuel) ;<p>- Les dépendances liées à l’aspect multi-composant du système, qui tient compte des modes de défaillance spécifiques des systèmes de protection ;<p>- L’aspect économique lié à la stratégie de maintenance (inspection, entretien, réparation, remplacement), aux coupures (programmées et forcées) et aux événements à risque (refus disjoncteur, perte d’un client, perte d’un jeu de barres, perte d’une sous-station, etc.). / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

Partial dynamic reconfiguration of FPGAs has attracted high attention from both academia and industry in recent years. With this technique, the functionality of the programmable devices can be adapted at runtime to changing requirements. The approach allows designers to use FPGAs more efficiently: E. g. FPGA resources can be time-shared between different functions and the functions itself can be adapted to changing workloads at runtime. Thus partial dynamic reconfiguration enables a unique combination of software-like flexibility and hardware-like performance. Still there exists no common understanding on how to assess the overhead introduced by partial dynamic reconfiguration. This dissertation presents a new cost model for both the runtime and the memory overhead that results from partial dynamic reconfiguration. It is shown how the model can be incorporated into all stages of the design optimization for reconfigurable hardware. In particular digital circuits can be mapped onto FPGAs such that only small fractions of the hardware must be reconfigured at runtime, which saves time, memory, and energy. The design optimization is most efficient if it is applied during high level synthesis. This book describes how the cost model has been integrated into a new high level synthesis tool. The tool allows the designer to trade-off FPGA resource use versus reconfiguration overhead. It is shown that partial reconfiguration causes only small overhead if the design is optimized with regard to reconfiguration cost. A wide range of experimental results is provided that demonstrates the benefits of the applied method.:1 Introduction 1 1.1 Reconfigurable Computing . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Reconfigurable System on a Chip (RSOC) . . . . . . . . . . . . 4 1.1.2 Anatomy of an Application . . . . . . . . . . . . . . . . . . . . . . 6 1.1.3 RSOC Design Characteristics and Trade-offs . . . . . . . . . . . 7 1.2 Classification of Reconfigurable Architectures . . . . . . . . . . . . . . . 10 1.2.1 Partial Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.2 Runtime Reconfiguration (RTR) . . . . . . . . . . . . . . . . . . . 10 1.2.3 Multi-Context Configuration . . . . . . . . . . . . . . . . . . . . . 11 1.2.4 Fine-Grain Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2.5 Coarse-Grain Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.3 Reconfigurable Computing Specific Design Issues . . . . . . . . . . . . 12 1.4 Overview of this Dissertation . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 Reconfigurable Computing Systems – Background 17 2.1 Examples for RSOCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2 Partially Reconfigurable FPGAs: Xilinx Virtex Device Family . . . . . . 20 2.2.1 Virtex-II/Virtex-II Pro Logic Architecture . . . . . . . . . . . . . 20 2.2.2 Reconfiguration Architecture and Reconfiguration Control . . 21 2.3 Methods for Design Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.3.1 Behavioural Design Entry . . . . . . . . . . . . . . . . . . . . . . . 25 2.3.2 Design Entry at Register-Transfer Level (RTL) . . . . . . . . . . 25 2.3.3 Xilinx Early Access Partial Reconfiguration Design Flow . . . . 26 2.4 Task Management in Reconfigurable Computing . . . . . . . . . . . . . 27 2.4.1 Online and Offline Task Management . . . . . . . . . . . . . . . 28 2.4.2 Task Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.4.3 Task Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.4.4 Reconfiguration Runtime Overhead . . . . . . . . . . . . . . . . 31 2.5 Configuration Data Compression . . . . . . . . . . . . . . . . . . . . . . . 32 2.6 Evaluation of Reconfigurable Systems . . . . . . . . . . . . . . . . . . . . 35 2.6.1 Energy Efficiency Models . . . . . . . . . . . . . . . . . . . . . . . 35 2.6.2 Area Efficiency Models . . . . . . . . . . . . . . . . . . . . . . . . 37 2.6.3 Runtime Efficiency Models . . . . . . . . . . . . . . . . . . . . . . 37 2.7 Similarity Based Reduction of Reconfiguration Overhead . . . . . . . . 38 2.7.1 Configuration Data Generation Methods . . . . . . . . . . . . . 39 2.7.2 Device Mapping Methods . . . . . . . . . . . . . . . . . . . . . . . 40 2.7.3 Circuit Design Methods . . . . . . . . . . . . . . . . . . . . . . . . 41 2.7.4 Model for Partial Configuration . . . . . . . . . . . . . . . . . . . 44 2.8 Contributions of this Work . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3 Runtime Reconfiguration Cost and Optimization Methods 47 3.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.2 Reconfiguration State Graph . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.2.1 Reconfiguration Time Overhead . . . . . . . . . . . . . . . . . . 52 3.2.2 Dynamic Configuration Data Overhead . . . . . . . . . . . . . . 52 3.3 Configuration Cost at Bitstream Level . . . . . . . . . . . . . . . . . . . . 54 3.4 Configuration Cost at Structural Level . . . . . . . . . . . . . . . . . . . 56 3.4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 3.4.2 Virtual Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.4.3 Reconfiguration Costs in the VA Context . . . . . . . . . . . . . 65 3.5 Allocation Functions with Minimal Reconfiguration Costs . . . . . . . 67 3.5.1 Allocation of Node Pairs . . . . . . . . . . . . . . . . . . . . . . . 68 3.5.2 Direct Allocation of Nodes . . . . . . . . . . . . . . . . . . . . . . 76 3.5.3 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4 Implementation Tools for Reconfigurable Computing 95 4.1 Mapping of Netlists to FPGA Resources . . . . . . . . . . . . . . . . . . . 96 4.1.1 Mapping to Device Resources . . . . . . . . . . . . . . . . . . . . 96 4.1.2 Connectivity Transformations . . . . . . . . . . . . . . . . . . . . 99 4.1.3 Mapping Variants and Reconfiguration Costs . . . . . . . . . . . 100 4.1.4 Mapping of Circuit Macros . . . . . . . . . . . . . . . . . . . . . . 101 4.1.5 Global Interconnect . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.1.6 Netlist Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 4.2 Mapping Aware Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . 103 4.2.1 Generalized Node Mapping . . . . . . . . . . . . . . . . . . . . . 104 4.2.2 Successive Node Allocation . . . . . . . . . . . . . . . . . . . . . 105 4.2.3 Node Allocation with Ant Colony Optimization . . . . . . . . . 107 4.2.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 4.3 Netlist Mapping with Minimized Reconfiguration Cost . . . . . . . . . 110 4.3.1 Mapping Database . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 4.3.2 Mapping and Packing of Elements into Logic Blocks . . . . . . 112 4.3.3 Logic Element Selection . . . . . . . . . . . . . . . . . . . . . . . 114 4.3.4 Logic Element Selection for Min. Routing Reconfiguration . . 115 4.3.5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 5 High-Level Synthesis for Reconfigurable Computing 125 5.1 Introduction to HLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 5.1.1 HLS Tool Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 5.1.2 Realization of the Hardware Tasks . . . . . . . . . . . . . . . . . 128 5.2 New Concepts for Task-based Reconfiguration . . . . . . . . . . . . . . 131 5.2.1 Multiple Hardware Tasks in one Reconfigurable Module . . . . 132 5.2.2 Multi-Level Reconfiguration . . . . . . . . . . . . . . . . . . . . . 133 5.2.3 Resource Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 5.3 Datapath Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 5.3.1 Task Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 5.3.2 Resource Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.3.3 Resource Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.3.4 Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 5.3.5 Constraints for Scheduling and Resource Binding . . . . . . . . 151 5.4 Reconfiguration Optimized Datapath Implementation . . . . . . . . . . 153 5.4.1 Effects of Scheduling and Binding on Reconfiguration Costs . 153 5.4.2 Strategies for Resource Type Binding . . . . . . . . . . . . . . . 154 5.4.3 Strategies for Resource Instance Binding . . . . . . . . . . . . . 157 5.5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 5.5.1 Summary of Binding Methods and Tool Setup . . . . . . . . . . 163 5.5.2 Cost Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 5.5.3 Implementation Scenarios . . . . . . . . . . . . . . . . . . . . . . 166 5.5.4 Benchmark Characteristics . . . . . . . . . . . . . . . . . . . . . . 168 5.5.5 Benchmark Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 5.5.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 6 Summary and Outlook 185 Bibliography 189 A Simulated Annealing 201 / Partielle dynamische Rekonfiguration von FPGAs hat in den letzten Jahren große Aufmerksamkeit von Wissenschaft und Industrie auf sich gezogen. Die Technik erlaubt es, die Funktionalität von progammierbaren Bausteinen zur Laufzeit an veränderte Anforderungen anzupassen. Dynamische Rekonfiguration erlaubt es Entwicklern, FPGAs effizienter einzusetzen: z.B. können Ressourcen für verschiedene Funktionen wiederverwendet werden und die Funktionen selbst können zur Laufzeit an veränderte Verarbeitungsschritte angepasst werden. Insgesamt erlaubt partielle dynamische Rekonfiguration eine einzigartige Kombination von software-artiger Flexibilität und hardware-artiger Leistungsfähigkeit. Bis heute gibt es keine Übereinkunft darüber, wie der zusätzliche Aufwand, der durch partielle dynamische Rekonfiguration verursacht wird, zu bewerten ist. Diese Dissertation führt ein neues Kostenmodell für Laufzeit und Speicherbedarf ein, welche durch partielle dynamische Rekonfiguration verursacht wird. Es wird aufgezeigt, wie das Modell in alle Ebenen der Entwurfsoptimierung für rekonfigurierbare Hardware einbezogen werden kann. Insbesondere wird gezeigt, wie digitale Schaltungen derart auf FPGAs abgebildet werden können, sodass nur wenig Ressourcen der Hardware zur Laufzeit rekonfiguriert werden müssen. Dadurch kann Zeit, Speicher und Energie eingespart werden. Die Entwurfsoptimierung ist am effektivsten, wenn sie auf der Ebene der High-Level-Synthese angewendet wird. Diese Arbeit beschreibt, wie das Kostenmodell in ein neuartiges Werkzeug für die High-Level-Synthese integriert wurde. Das Werkzeug erlaubt es, beim Entwurf die Nutzung von FPGA-Ressourcen gegen den Rekonfigurationsaufwand abzuwägen. Es wird gezeigt, dass partielle Rekonfiguration nur wenig Kosten verursacht, wenn der Entwurf bezüglich Rekonfigurationskosten optimiert wird. Eine Anzahl von Beispielen und experimentellen Ergebnissen belegt die Vorteile der angewendeten Methodik.:1 Introduction 1 1.1 Reconfigurable Computing . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Reconfigurable System on a Chip (RSOC) . . . . . . . . . . . . 4 1.1.2 Anatomy of an Application . . . . . . . . . . . . . . . . . . . . . . 6 1.1.3 RSOC Design Characteristics and Trade-offs . . . . . . . . . . . 7 1.2 Classification of Reconfigurable Architectures . . . . . . . . . . . . . . . 10 1.2.1 Partial Reconfiguration . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.2 Runtime Reconfiguration (RTR) . . . . . . . . . . . . . . . . . . . 10 1.2.3 Multi-Context Configuration . . . . . . . . . . . . . . . . . . . . . 11 1.2.4 Fine-Grain Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2.5 Coarse-Grain Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.3 Reconfigurable Computing Specific Design Issues . . . . . . . . . . . . 12 1.4 Overview of this Dissertation . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 Reconfigurable Computing Systems – Background 17 2.1 Examples for RSOCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2 Partially Reconfigurable FPGAs: Xilinx Virtex Device Family . . . . . . 20 2.2.1 Virtex-II/Virtex-II Pro Logic Architecture . . . . . . . . . . . . . 20 2.2.2 Reconfiguration Architecture and Reconfiguration Control . . 21 2.3 Methods for Design Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.3.1 Behavioural Design Entry . . . . . . . . . . . . . . . . . . . . . . . 25 2.3.2 Design Entry at Register-Transfer Level (RTL) . . . . . . . . . . 25 2.3.3 Xilinx Early Access Partial Reconfiguration Design Flow . . . . 26 2.4 Task Management in Reconfigurable Computing . . . . . . . . . . . . . 27 2.4.1 Online and Offline Task Management . . . . . . . . . . . . . . . 28 2.4.2 Task Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.4.3 Task Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.4.4 Reconfiguration Runtime Overhead . . . . . . . . . . . . . . . . 31 2.5 Configuration Data Compression . . . . . . . . . . . . . . . . . . . . . . . 32 2.6 Evaluation of Reconfigurable Systems . . . . . . . . . . . . . . . . . . . . 35 2.6.1 Energy Efficiency Models . . . . . . . . . . . . . . . . . . . . . . . 35 2.6.2 Area Efficiency Models . . . . . . . . . . . . . . . . . . . . . . . . 37 2.6.3 Runtime Efficiency Models . . . . . . . . . . . . . . . . . . . . . . 37 2.7 Similarity Based Reduction of Reconfiguration Overhead . . . . . . . . 38 2.7.1 Configuration Data Generation Methods . . . . . . . . . . . . . 39 2.7.2 Device Mapping Methods . . . . . . . . . . . . . . . . . . . . . . . 40 2.7.3 Circuit Design Methods . . . . . . . . . . . . . . . . . . . . . . . . 41 2.7.4 Model for Partial Configuration . . . . . . . . . . . . . . . . . . . 44 2.8 Contributions of this Work . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3 Runtime Reconfiguration Cost and Optimization Methods 47 3.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.2 Reconfiguration State Graph . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.2.1 Reconfiguration Time Overhead . . . . . . . . . . . . . . . . . . 52 3.2.2 Dynamic Configuration Data Overhead . . . . . . . . . . . . . . 52 3.3 Configuration Cost at Bitstream Level . . . . . . . . . . . . . . . . . . . . 54 3.4 Configuration Cost at Structural Level . . . . . . . . . . . . . . . . . . . 56 3.4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 3.4.2 Virtual Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.4.3 Reconfiguration Costs in the VA Context . . . . . . . . . . . . . 65 3.5 Allocation Functions with Minimal Reconfiguration Costs . . . . . . . 67 3.5.1 Allocation of Node Pairs . . . . . . . . . . . . . . . . . . . . . . . 68 3.5.2 Direct Allocation of Nodes . . . . . . . . . . . . . . . . . . . . . . 76 3.5.3 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4 Implementation Tools for Reconfigurable Computing 95 4.1 Mapping of Netlists to FPGA Resources . . . . . . . . . . . . . . . . . . . 96 4.1.1 Mapping to Device Resources . . . . . . . . . . . . . . . . . . . . 96 4.1.2 Connectivity Transformations . . . . . . . . . . . . . . . . . . . . 99 4.1.3 Mapping Variants and Reconfiguration Costs . . . . . . . . . . . 100 4.1.4 Mapping of Circuit Macros . . . . . . . . . . . . . . . . . . . . . . 101 4.1.5 Global Interconnect . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.1.6 Netlist Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 4.2 Mapping Aware Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . 103 4.2.1 Generalized Node Mapping . . . . . . . . . . . . . . . . . . . . . 104 4.2.2 Successive Node Allocation . . . . . . . . . . . . . . . . . . . . . 105 4.2.3 Node Allocation with Ant Colony Optimization . . . . . . . . . 107 4.2.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 4.3 Netlist Mapping with Minimized Reconfiguration Cost . . . . . . . . . 110 4.3.1 Mapping Database . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 4.3.2 Mapping and Packing of Elements into Logic Blocks . . . . . . 112 4.3.3 Logic Element Selection . . . . . . . . . . . . . . . . . . . . . . . 114 4.3.4 Logic Element Selection for Min. Routing Reconfiguration . . 115 4.3.5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 5 High-Level Synthesis for Reconfigurable Computing 125 5.1 Introduction to HLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 5.1.1 HLS Tool Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 5.1.2 Realization of the Hardware Tasks . . . . . . . . . . . . . . . . . 128 5.2 New Concepts for Task-based Reconfiguration . . . . . . . . . . . . . . 131 5.2.1 Multiple Hardware Tasks in one Reconfigurable Module . . . . 132 5.2.2 Multi-Level Reconfiguration . . . . . . . . . . . . . . . . . . . . . 133 5.2.3 Resource Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 5.3 Datapath Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 5.3.1 Task Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 5.3.2 Resource Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.3.3 Resource Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.3.4 Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 5.3.5 Constraints for Scheduling and Resource Binding . . . . . . . . 151 5.4 Reconfiguration Optimized Datapath Implementation . . . . . . . . . . 153 5.4.1 Effects of Scheduling and Binding on Reconfiguration Costs . 153 5.4.2 Strategies for Resource Type Binding . . . . . . . . . . . . . . . 154 5.4.3 Strategies for Resource Instance Binding . . . . . . . . . . . . . 157 5.5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 5.5.1 Summary of Binding Methods and Tool Setup . . . . . . . . . . 163 5.5.2 Cost Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 5.5.3 Implementation Scenarios . . . . . . . . . . . . . . . . . . . . . . 166 5.5.4 Benchmark Characteristics . . . . . . . . . . . . . . . . . . . . . . 168 5.5.5 Benchmark Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 5.5.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 6 Summary and Outlook 185 Bibliography 189 A Simulated Annealing 201