Coarse-grained (CG) molecular dynamics (MD) simulations have seen a wide range of applications from biomolecules, polymers to graphene and metals. In CG MD simulations, atomistic groups are represented by beads, which reduces the degrees of freedom in the systems and allows larger timesteps. Thus, large time and length scales could be achieved in CG MD simulations with inexpensive computational cost. The representative example of large time- and length-scale phenomena is the conformation transitions of single polymer chains as well as polymer chains in their architectures, self-assembly of biomaterials, etc. Polymers exist in many aspects of our life, for example, plastic packages, automobile parts, and even medical devices. However, the large chemical and structural diversity of polymers poses a challenge to the existing CG MD models due to their limited accuracy and transferabilities. In this regard, this dissertation has developed CG models of polymers on the basis of accurate and transferable hydrocarbon models, which are important components of the polymer backbone. CG hydrocarbon models were created with 2:1 and 3:1 mapping schemes and their force-field (FF) parameters were optimized by using particle swarm optimization (PSO). The newly developed CG hydrocarbon models could reproduce their experimental properties including density, enthalpy of vaporization, surface tension and self-diffusion coefficients very well. The cross interaction parameters between CG hydrocarbon and water models were also optimized by the PSO to repeat the experimental properties of Gibbs free energies and interfacial tensions. With the hydrocarbon models as the backbone, poly(acrylic acid) (PAA) and polystyrene (PS) models were constructed. Their side chains were represented by one COOH (carboxylic acid) and three BZ beads, respectively. Before testing the PAA and PS models, their monomer models, propionic acid and ethylbenzene, were created and validated, to confirm that the cross interactions between hydrocarbon and COOH beads, and between hydrocarbon and BZ beads could be accurately predicted by the Lorentz-Berthelot (LB) combining rules. Then the experimental properties, density of polymers at 300 K and glass transition temperatures, and the conformations of their all-atom models in solvent mixtures of water and dimethylformamide (DMF) were reproduced by the CG models. The CG PAA and PS models were further used to build the bottlebrush copolymers of PAA-PS and to predict the structures of PAA-PA in different compositions of binary solvents water/DMF. Although CG models are useful in understanding the phenomena at large time- or length- scales, atomistic information is lost. Backmapping is usually involved in reconstructing atomistic models from their CG models. Here, four machine learning (ML) algorithms, artificial neural networks (ANN), k-nearest neighbor (kNN), gaussian process regression (GPR), and random forest (RF) were developed to improve the accuracy of the backmapped all-atom structures. These optimized four ML models showed R2 scores of more than 0.99 when testing the backmapping against four representative molecules: furan, benzene, naphthalene, graphene. / Doctor of Philosophy / Polymers have a wide range of applications from packaging, foams, coating to pipes, tanks and even medical devices and biosensors. To improve the properties of these materials it is important to understand their structure and features responsible for controlling their properties at the molecular-level. Molecular dynamic (MD) simulations are a powerful tool to study their structures and properties at microscopic level. However, studying the molecular-level conformations of polymers and their architectures usually requires large time- or length-scales, which is challenging for the all-atom MD simulations because of the high computational cost. Coarse-grained (CG) MD simulations can be used to study these soft-materials as they represent atomistic groups with beads, enabling the reduction of the system sizes drastically, and allowing the use of large timesteps in MD simulations. In MD simulations, force-fields (FF) that describe the intramolecular and intermolecular interactions determine the performance of simulations. Here, we firstly optimized the FF parameters for hydrocarbons. With the optimized CG hydrocarbon models, two representative CG polymer models, poly(acrylic acid) (PAA) and polystyrene (PS) were built by using hydrocarbons as the backbones of polymers. Furthermore, the PAA and PS chains were grafted on a linear hydrocarbon backbone to form a bottlebrush copolymer. Although CG MD models are useful in studying the complex process of polymers, the atomic detailed information is lost. To reconstruct accurate atomistic structures, backmapping by using machine learning (ML) algorithms was performed. The performance of the ML models was better than that of the existing backmapping packages built in Visual Molecular Dynamics (VMD).

Polyetherimides are an important class of engineering thermoplastics used in a broad range of industries and applications because of their high heat resistance and stability, high strength and moduli, excellent electrical properties over a wide range of temperatures and frequencies, good processability, good adhesive properties, and chemical stability. All-atom molecular dynamics (MD) simulation is a useful tool to study polymers, but the accessible length and time scales are limited. In this thesis, we explore several computational methods that go beyond all-atom MD simulations to investigate polyetherimides. First, we have developed a transferable coarse-grained MD model of polyetherimides that captures their mechanical and thermal expansion properties. Our results show that in order to make the model transferable, it is critical to include an entropic correction term in the coarse-grained force field and require the coarse-grained model to capture the thermal expansion property of polyetherimides. Secondly, we have constructed a predictive model of the glass transition temperature (Tg) for polyimides by using machine-learning algorithms to analyze existing data on Tg reported in the literature. The predictive model is validated by comparing its predictions to experimental data not used in the training process of the model. We further demonstrate that the diffusion coefficients of small gas molecules can be quickly computed with all-atom MD simulations and used to determine Tg. Finally, we have developed a Monte Carlo (MC) program to model the polymerization process of branched polyetherimides and to compute their molecular weight distribution for a wide range of systems, including fully reacted, partially reacted, stoichiometric, and nonstoichiometric ones. The MC results are compared to the predictions of the Flory-Stockmayer theory of branched polymers and an excellent agreement is found below the gel point of the system under consideration. Above the gel point, the Flory- Stockmayer theory starts to fail but the MC method can still be used to quickly determine the molecular weight distribution of branched polyetherimides under very general conditions. / Doctor of Philosophy / Polyetherimides are an important category of engineering plastics with wide applications in many fields because of their superior mechanical, thermal, chemical, and electrical properties. All-atom molecular dynamics simulations serve as a useful tool to study the properties of polyetherimides in silico. However, such simulations are computationally expensive and therefore limited to small system sizes and short time scales. To overcome these issues, we employed various computational techniques in this thesis to model polyetherimides. First, we have developed a coarse-grained model of polyetherimides where atoms are grouped into beads. We show that molecular dynamics simulations on the basis of the coarse-grained model can be used to provide a reasonable description of the mechanical and thermal expansion properties of polyetherimides. Secondly, we have constructed a predictive model of the glass transition temperature, which is the temperature at which a material enters a glassy state when cooled rapidly, of polyimides using machine-learning algorithms. This model is capable of estimating the glass transition temperature of polyimides within an accuracy of ± 15 K even for those not synthesized yet. We further show that the diffusion coefficients of gas molecules, in addition to the polymer density, can be computed accurately with all-atom molecular dynamics simulations and used to determine the glass transition temperature of polyimides. Finally, we have developed a Monte Carlo scheme to efficiently model the polymerization and compute the chain-length distribution of branched polyetherimides under very general conditions. The results from Monte Carlo simulations are compared to the predictions of the Flory-Stockmayer theory of branched polymers. The range of applicability of the theory is revealed. Overall, we have demonstrated several computational techniques that can be used to efficiently model polyetherimides, potentially other polymers as well, beyond the widely-used all-atom molecular dynamics simulations.

L'analyse visuelle de scènes débute par l'extraction en parallèle de l'information à différentes fréquences spatiales, en suivant un mode de traitement par défaut « coarse-to-fine ». L'analyse rapide de l'information grossière (« coarse ») en basses fréquences spatiales fournirait un aperçu global de la scène, qui serait ensuite affiné par l'analyse des détails de la scène (« fine ») en hautes fréquences spatiales. L'objectif de cette thèse a été de préciser les bases cérébrales du traitement des fréquences spatiales lors de la catégorisation de scènes. A travers deux études comportementales, nous avons tout d'abord montré qu'une analyse « coarse-to-fine » est plus avantageuse pour la catégorisation rapide de scènes, et ce, indépendamment de la valeur de contraste de luminance associée aux différentes fréquences spatiales (Expériences 1 et 2). Des études en IRMf nous ont par la suite permis de mettre en évidence l'implication d'un large réseau cérébral lors de l'analyse « coarse-to-fine » des scènes, incluant les aires visuelles primaires et occipito-temporales, mais également le cortex frontal inférieur (Expérience 3). Une analyse de la connectivité a révélé que lors de cette analyse, le cortex frontal inférieur exercerait une influence « top-down » sur le cortex visuel primaire et les gyri fusiforme et parahippocampique au sein du cortex occipito-temporal. Ces résultats soulignent le rôle du cortex visuel primaire comme région intégrative, codant à la fois les afférences rétino-thalamiques et les influences « top-down » de régions supérieures. Nous avons également observé que le gyrus frontal inférieur et le gyrus fusiforme participaient activement à l'intégration de l'information sémantique contenue dans les basses et hautes fréquences spatiales d'une scène (Expérience 4). Enfin, nous avons spécifiquement étudié le traitement des fréquences spatiales au sein de régions occipito-temporales sélectives aux scènes : la « parahippocampal place area » (PPA), le cortex retrosplenial et l'« occipital place area ». Nous avons montré que ces trois régions participent de façon distincte au traitement des fréquences spatiales dans les scènes (Expérience 5) et qu'une stratégie d'analyse « coarse-to-fine » serait privilégiée par la PPA (Expérience 6). Les résultats de ces travaux nous permettent de conforter et de préciser les modèles actuels de la catégorisation visuelle de scènes basés sur un traitement fréquentiel de l'information visuelle. / Visual analysis begins with the parallel extraction of different attributes at different spatial frequencies following a predominantly coarse-to-fine default processing sequence. Rapid processing of low spatial frequency information would permit a coarse parsing of the visual input, prior to the detailed analysis of fine information in high spatial frequencies. Our aim was to further address the neural bases of spatial frequency processing during scene categorization. We first demonstrated in two behavioral studies that a coarse-to-fine processing is indeed an advantageous strategy for rapid scene categorization, and is independent of the luminance contrast values associated with the different spatial frequencies (Experiments 1 and 2). In two fMRI studies, we showed first the involvement of a large cerebral network during coarse-to-fine processing of scenes, including early visual and occipito-temporal areas, but also the inferior frontal cortex. Effective connectivity analysis revealed that the inferior frontal gyrus exerts top-down influence on the early visual cortex as well as on the parahippocampal and fusiform gyri in the occipito-temporal cortex (Experiment 3). These results highlight the role of the primary visual cortex in integrating top-down influences from frontal areas to retino-thalamic incoming signals. We also evidenced that the inferior frontal and fusiform gyri actively participate to the integration of the semantic information contained in low and high spatial frequency (Experiment 4). Finally, we specifically investigated the spatial frequency processing of scenes within scene-selective areas of the occipito-temporal cortex: the parahippocampal place area (PPA), the retrosplenial cortex, and the occipital place area. We demonstrated that these regions participate differently in the spatial frequency processing of scenes (Experiment 5) and that a coarse-to-fine processing is favored within the PPA (Experiment 6). Overall, results allow us to refine current model of visual scene categorization based on a spatial frequency analysis.

Le thème central de cette thèse est l'étude de plongements d'espaces métriques dans des espaces de Banach. La première étude concerne les plongements grossièrement Lipschitz entre les espaces de James Jp pour p≻1 et p fini. On obtient que, pour p,q différents, Jq ne se plonge pas grossièrement Lipschitz dans Jp. Nous avons également obtenu, dans le cas où q≺p, une majoration de l'exposant de compression de Jq dans Jp par q/p. La question naturelle qui se pose ensuite est de savoir si le résultat obtenu pour les espaces de James est vrai aussi en ce qui concerne leurs duaux. Nous obtenons que, pour p,q différents, Jp* ne se plonge pas grossièrement lipschitz dans Jq*. Suite à ce travail, on établit des résultats plus généraux sur la non-plongeabilité des espaces de Banach q-AUS dans les espaces de Banach p-AMUC pour p≺q. On en déduit aussi, à l'aide d'un théorème de renormage, un résultat sur les indices de Szlenk. Par ailleurs, on obtient un résultat sur la plongeabilité quasi-Lipschitz dont la définition diffère légèrement de la plongeabilité presque Lipschitz : pour deux espaces de Banach X et Y, si, pour C≻1, X est C-finiment crûment représentable dans tout sous-espace vectoriel de codimension finie de Y, alors tout sous-espace propre M de X se plonge quasi-Lipschitz dans Y. Pour conclure, on obtient le corollaire suivant : soient X et Y deux espaces de Banach tels que X est localement minimal et Y est finiment crûment représentable dans X. Alors, pour M sous-espace propre de Y, M se plonge quasi-Lipschitz dans X. / The central theme of this thesis is the study of embeddings of metric spaces into Banach spaces.The first study focuses on the coarse Lipschitz embeddings between James Spaces Jp for p≻1 and p finite. We obtain that, for p,q different, Jq does not coarse Lipschitz embed into Jp. We also obtain, in the case where q≺p, that the compression exponent of Jq in Jp is lower or equal to q/p. Another natural question is to know whether we have similar results for the dual spaces of James spaces. We obtain that, for p,q different, Jp* does not coarse Lipschitz embed into Jq*. Further to this work, we establish a more general result about the coarse Lipschitz embeddability of a Banach space which has a q-AUS norm into a Banach space which has a p-AMUC norm for p≺q. With the help of a renorming theorem, we deduce also a result about the Szlenk index. Moreover, after defining the quasi-Lipschitz embeddability, which is slightly different to the almost Lipschitz embeddability, we obtain the following result: For two Banach spaces X, if X is crudely finitely representable with constant C (where C≻1) in any subspace of Y of finite codimension, then every proper subset M of X quasi-Lipschitz embeds into Y. To conclude, we obtain the following corollary: Let X be a locally minimal Banach space, and Y be a Banach space which is crudely finitely representable in X. Then, for M a proper subspace of Y, M quasi-Lipschitz embeds into X.

Numerical simulations of large complex systems such as biomolecules often suffer from the full description of the system having too many dimensions for direct numerical calculations and Monte Carlo methods having trouble overcoming energy barriers. It is therefore desirable to formulate a description in lower dimension which captures the system’s macroscopic behaviour. Recently, Lindahl et al [1] proposed a metric, g(λ), on the extended space Λ based on the dynamics of the system to optimize Monte Carlo sampling within extended ensemble formalism. In this thesis, we formulate a low-dimensional effective coarse-grained dynamic on Λ as a diffusion process and ask if it is possible to use this metric to calculate thelocal effective diffusion matrix as D(λ) = g−1(λ). By testing various scenarios we conclude that computing D(λ) in this manner indeed gives a correct effective dynamic in most cases, where the scale of coarse-graining can be tuned. However, an incorrect dynamic is received for example when the scale of coarse-graining is comparable to the size of oscillations in the energy landscape. / Numeriska simuleringar av stora komplexa system såsom biomolekyler lider ofta av att den fulla beskrivningen av systemet har för många dimensioner för direkta numeriska beräkningar samt att Monte Carlo-metoder har svårt att komma över energibarriärer. Det är därför önskvärt att formulera en beskrivning i lägre dimension som fångar systemets makroskopiska beteende. Nyligen föreslog Lindahl et al [1] en metrik g(λ) på det utvidgade rummet Λ baserad på dynamiken av systemet för att optimera Monte Carlo-sampling inom formalismen av utvidgade ensembler. I den här uppsatsen formulerar vi en lågdimensionell effektiv grov dynamik på Λ som en diffusionsprocess och frågar om det är möjligt att använda den här metriken för att beräkna den lokala effektiva diffusionsmatrisen som D(λ) = g(λ)−1. Genom testning av flera scenarier drar vi slutsatsen att beräkna D(λ) på det här sättet ger en korrekt effektiv dynamik i de flesta fall, där skalan på förgrovningen kan ställas in. Däremot fås en inkorrekt dynamik till exempel när skalan på förgrovningen ärjämförbar med storleken på oscillationer i energilandskapet.

Este trabalho é uma contribuição ao estudo da geometria de larga escala de espaços de Banach e de grupos topológicos. Embora esses dois campos sejam tradicionalmente estudados de forma independente, em 2017, Christian Rosendal mostrou que eles podem ser encarados como faces distintas de algo maior: a geometria grosseira de grupos topológicos. Uma ferramenta essencial para o desenvolvimento dessa nova abordagem é a noção de estrutura coarse, introduzida por John Roe em 2003, a qual pode ser vista como a contraparte de larga escala do conceito de estrutura uniforme. Por essa razão, os capítulos iniciais da dissertação destinam-se a apresentar uma introdução elementar à teoria dos espaços uniformes e dos espaços coarse, destacando os conceitos-chave para a compreensão dos demais capítulos e conferindo particular atenção ao estudo de uniformidades e estruturas coarse associadas a grupos topológicos, dentre as quais são enfatizadas as estruturas uniforme à esquerda e coarse à esquerda de um grupo topológico. No capítulo 5, são discutidos resultados recentes de Christian Rosendal acerca da existência de mergulhos uniformes e mergulhos grosseiros entre espaços de Banach. Dois dos mais importantes afirmam que, se existir uma função f uniformemente contínua e não colapsada entre os espaços de Banach (X, ||·||_X) e (E, ||·||_E), então, para todo p em [1, + infty[, existirá um mergulho uniforme de (X, ||·||_X) em (l_p(E), ||·||_p) o qual é, também, um mergulho grosseiro, e que, se f for, também, limitada, existirá um mergulho grosseiro uniformemente contínuo de (X, ||·||_X) em (ExE, ||·||_(ExE)). Já no capítulo 6, estuda-se a classe das estruturas coarse invariantes à esquerda sobre grupos. Inicialmente, mostra-se como uma estrutura coarse invariante à esquerda em um grupo (G, · ) pode ser descrita em função de um certo ideal sobre G, e vice-versa. Em seguida, utiliza-se esse resultado para caracterizar a estrutura coarse à esquerda E_L de um grupo topológico (G, · , T) em termos da coleção dos conjuntos grosseiramente limitados em (G, E_L) e, com isso, provar que a estrutura coarse à esquerda associada ao grupo aditivo de um espaço normado coincide com a estrutura coarse limitada induzida pela norma. / This work is a contribution to the study of large-scale geometry of Banach spaces and topological groups. Although these two fields are traditionally studied independently, in 2017, Christian Rosendal showed they can be regarded as different aspects of a more general theory: the coarse geometry of topological groups. An essential tool for the development of this new approach is the notion of coarse structure, introduced by John Roe in 2003, which can be seen as the large-scale counterpart of the concept of uniform structure. For this reason, the initial chapters of this work intend to present an elementary introduction to both uniform and coarse spaces theory, highlighting the key concepts for the understanding of the other chapters and paying particular attention to the study of uniform and coarse structures associated with topological groups, and, mainly, to the left-uniform and the left-coarse structures of a topological group. In Chapter 5, we discuss Rosendal\'s recent results on the existence of uniform and coarse embeddings between Banach spaces. Two of the most important state that, if there is an uncollapsed uniformly continuous function f between the Banach spaces (X, ||·||_X) and (E, ||·||_E), then, for all p in [1, + infty[, (X, ||·||_X) admits a simultaneously uniform and coarse embedding into (l_p(E), ||·||_p), and that, if, in addition, we assume that f maps into a bounded set, then (X, ||·||_X) also admits a uniformly continuous coarse embedding into (ExE, ||·||_(ExE)). On the other hand, in chapter 6, we focus our attention on the class of left-invariant coarse structures on groups. In the first section, we show how a left-invariant coarse structure on a group (G, · ) can be described in terms of a certain ideal on G, and vice versa. After that, we use this result to characterize the left-coarse structure E_L of a topological group (G, · , T) in terms of the collection of the coarsely bounded sets of (G, E_L) and, with this, we prove that the left-coarse structure associated with the additive group of a normed space is simply the bounded coarse structure induced by its norm.

A Coarse-Grained Reconfigurable Architecture (CGRA) is a processing platform which constitutes an interconnection of coarse-grained computation units (viz. Function Units (FUs), Arithmetic Logic Units (ALUs)). These units communicate directly, viz. send-receive like primitives, as opposed to the shared memory based communication used in multi-core processors. CGRAs are a well-researched topic and the design space of a CGRA is quite large. The design space can be represented as a 7-tuple (C, N, T, P, O, M, H) where each of the terms have the following meaning: C -choice of computation unit, N -choice of interconnection network, T -Choice of number of context frame (single or multiple), P -presence of partial reconfiguration, O choice of orchestration mechanism, M -design of memory hierarchy and H host-CGRA coupling. In this thesis, we develop an architectural framework for a Macro-Dataflow based CGRA where we make the following choice for each of these parameters: C -ALU, N -Network-on-Chip (NoC), T -Multiple contexts, P -support for partial reconfiguration, O -Macro Dataflow based orchestration, M -data memory banks placed at the periphery of the reconfigurable fabric (reconfigurable fabric is the name given to the interconnection of computation units), H -loose coupling between host processor and CGRA, enabling our CGRA to execute an application independent of the host-processor’s intervention. The motivations for developing such a CGRA are: To execute applications efficiently through reduction in reconfiguration time (i.e. the time needed to transfer instructions and data to the reconfigurable fabric) and reduction in execution time through better exploitation of all forms of parallelism: Instruction Level Parallelism (ILP), Data Level Parallelism (DLP) and Thread/Task Level Parallelism (TLP). We choose a macro-dataflow based orchestration framework in combination with partial reconfiguration so as to ease exploitation of TLP and DLP. Macro-dataflow serves as a light weight synchronization mechanism. We experiment with two variants of the macro-dataflow orchestration units, namely: hardware controlled orchestration unit and the compiler controlled orchestration unit. We employ a NoC as it helps reduce the reconfiguration overhead. To permit customization of the CGRA for a particular domain through the use of domain-specific custom-Intellectual Property (IP) blocks. This aids in improving both application performance and makes it energy efficient. To develop a CGRA which is completely programmable and accepts any program written using the C89 standard. The compiler and the architecture were co-developed to ensure that every feature of the architecture could be automatically programmed through an application by a compiler. In this CGRA framework, the orchestration mechanism (O) and the host-CGRA coupling (H) are kept fixed and we permit design space exploration of the other terms in the 7-tuple design space. The mode of compilation and execution remains invariant of these changes, hence referred to as a framework. We now elucidate the compilation and execution flow for this CGRA framework. An application written in C language is compiled and is transformed into a set of temporal partitions, referred to as HyperOps in this thesis. The macro-dataflow orchestration unit selects a HyperOp for execution when all its inputs are available. The instructions and operands for a ready HyperOp are transferred to the reconfigurable fabric for execution. Each ALU (in the computation unit) is capable of waiting for the availability of the input data, prior to issuing instructions. We permit the launch and execution of a temporal partition to progress in parallel, which reduces the reconfiguration overhead. We further cut launch delays by keeping loops persistent on fabric and thus eliminating the need to launch the instructions. The CGRA framework has been implemented using Bluespec System Verilog. We evaluate the performance of two of these CGRA instances: one for cryptographic applications and another instance for linear algebra kernels. We also run other general purpose integer and floating point applications to demonstrate the generic nature of these optimizations. We explore various microarchitectural optimizations viz. pipeline optimizations (i.e. changing value of T ), different forms of macro dataflow orchestration such as hardware controlled orchestration unit and compiler-controlled orchestration unit, different execution modes including resident loops, pipeline parallelism, changes to the router etc. As a result of these optimizations we observe 2.5x improvement in performance as compared to the base version. The reconfiguration overhead was hidden through overlapping launching of instructions with execution making. The perceived reconfiguration overhead is reduced drastically to about 9-11 cycles for each HyperOp, invariant of the size of the HyperOp. This can be mainly attributed to the data dependent instruction execution and use of the NoC. The overhead of the macro-dataflow execution unit was reduced to a minimum with the compiler controlled orchestration unit. To benchmark the performance of these CGRA instances, we compare the performance of these with an Intel Core 2 Quad running at 2.66GHz. On the cryptographic CGRA instance, running at 700MHz, we observe one to two orders of improvement in performance for cryptographic applications and up to one order of magnitude performance degradation for linear algebra CGRA instance. This relatively poor performance of linear algebra kernels can be attributed to the inability in exploiting ILP across computation units interconnected by the NoC, long latency in accessing data memory placed at the periphery of the reconfigurable fabric and unavailability of pipelined floating point units (which is critical to the performance of linear algebra kernels). The superior performance of the cryptographic kernels can be attributed to higher computation to load instruction ratio, careful choice of custom IP block, ability to construct large HyperOps which allows greater portion of the communication to be performed directly (as against communication through a register file in a general purpose processor) and the use of resident loops execution mode. The power consumption of a computation unit employed on the cryptography CGRA instance, along with its router is about 76mW, as estimated by Synopsys Design Vision using the Faraday 90nm technology library for an activity factor of 0.5. The power of other instances would be dependent on specific instantiation of the domain specific units. This implies that for a reconfigurable fabric of size 5 x 6 the total power consumption is about 2.3W. The area and power ( 84mW) dissipated by the macro dataflow orchestration unit, which is common to both instances, is comparable to a single computation unit, making it an effective and low overhead technique to exploit TLP.

Coarse-Grained Reconfigurable Architectures(CGRAs) can be employed for accelerating computational workloads that demand both flexibility and performance. CGRAs comprise a set of computation elements interconnected using a network and this interconnection of computation elements is referred to as a reconfigurable fabric. The size of application that can be accommodated on the reconfigurable fabric is limited by the size of instruction buffers associated with each Compute element. When an application cannot be accommodated entirely, application is partitioned such that each of these partitions can be executed on the reconfigurable fabric. These partitions are scheduled by an orchestrator. The orchestrator employs dynamic dataflow execution paradigm. Dynamic dataflow execution paradigm has inherent support for synchronization and helps in exploitation of parallelism that exists across application partitions. In this thesis, we present a compiler that targets such CGRAs. The compiler presented in this thesis is capable of accepting applications specified in C89 standard. To enable architectural design space exploration, the compiler is designed such that it can be customized for several instances of CGRAs employing dataflow execution paradigm at the orchestrator. This can be achieved by specifying the appropriate configuration parameters to the compiler. The focus of this thesis is to provide efficient support for various kinds of parallelism while ensuring correctness. The compiler is designed to support fine-grained task level parallelism that exists across iterations of loops and function calls. Additionally, compiler can also support pipeline parallelism, where a loop is split into multiple stages that execute in a pipelined manner. The prototype compiler, which targets multiple instances of a CGRA, is demonstrated in this thesis. We used this compiler to target multiple variants of CGRAs employing dataflow execution paradigm. We varied the reconfigur-able fabric, orchestration mechanism employed, size of instruction buffers. We also choose applications from two different domains viz. cryptography and linear algebra. The execution time of the CGRA (the best among all instances) is compared against an Intel Quad core processor. Cryptography applications show a performance improvement ranging from more than one order of magnitude to close to two orders of magnitude. These applications have large amounts of ILP and our compiler could successfully expose the ILP available in these applications. Further, the domain customization also played an important role in achieving good performance. We employed two custom functional units for accelerating Cryptography applications and compiler could efficiently use them. In linear algebra kernels we observe multiple iterations of the loop executing in parallel, effectively exploiting loop-level parallelism at runtime. Inspite of this we notice close to an order of magnitude performance degradation. The reason for this degradation can be attributed to the use of non-pipelined floating point units, and the delays involved in accessing memory. Pipeline parallelism was demonstrated using this compiler for FFT and QR factorization. Thus, the compiler is capable of efficiently supporting different kinds of parallelism and can support complete C89 standard. Further, the compiler can also support different instances of CGRAs employing dataflow execution paradigm.

The Aurora Soil Capping study, located in northern Alberta, was constructed to evaluate reclamation practices on lean oil sands dumps. The challenges relating to its success includes determining the appropriate soil cover design(s) for the coarse textured reclamation soil, while utilizing available salvaged natural soils, some of which contain residual bitumen in the form of aggregate oil sand material (AOSM). Limited research on this material raises key questions as to the impact it will play on transport and retention processes, along with potential contamination from hydrocarbon leaching. The research conducted sought to answer these questions. This thesis describes laboratory studies conducted on four soils; the upper organic LFH layer, Bm, BC and subsoil material while varying the amount of AOSM and implementing layering schemes. Material characterization through organic carbon and particle size analysis as well as hydrophobicity studies on AOSM through contact angle analysis were performed. A tension table and pressure plates, along with columns equipped with Time Domain Reflectometry probes, were used for water retention studies. Hydraulic conductivity was measured through constant head methods. To address hydrocarbon leaching concerns, chloride tracer studies were performed and the column outflow was analyzed using Gas Chromatography to detect the hydrocarbon type and concentration. Results from water retention and hydraulic conductivity studies indicated that although the AOSM was hydrophobic, its placement at varying concentrations and forms did not create consistent significant differences in the amount of moisture retained or transported. Results from the column studies showed that under steady state and transient conditions AOSM could result in decreasing infiltration rates and increasing chloride retention. The integration of soil layers further slowed the infiltration rate and delayed chloride transport. Under saturated conditions the presence of higher concentrations of AOSM appeared to increase the rate of water movement. Although these differences were minimal, further studies are required to explore this behavior. Overall, it can be concluded that with appropriate material placement, the addition of layering schemes and hydrocarbon material, the potential exists to increase soil water content in the upper layers of the soil, thereby increasing soil water storage for plant use.

No description available.