This dissertation presents and evaluates a methodology for scheduling medical application workloads in virtualized computing environments. Such environments are being widely adopted by providers of “cloud computing” services. In the context of provisioning resources for medical applications, such environments allow users to deploy applications on distributed computing resources while keeping their data secure. Furthermore, higher level services that further abstract the infrastructure-related issues can be built on top of such infrastructures. For example, a medical imaging service can allow medical professionals to process their data in the cloud, easing them from the burden of having to deploy and manage these resources themselves. In this work, we focus on issues related to scheduling scientific workloads on virtualized environments. We build upon the knowledge base of traditional parallel job scheduling to address the specific case of medical applications while harnessing the benefits afforded by virtualization technology. To this end, we provide the following contributions: An in-depth analysis of the execution characteristics of the target applications when run in virtualized environments. A performance prediction methodology applicable to the target environment. A scheduling algorithm that harnesses application knowledge and virtualization-related benefits to provide strong scheduling performance and quality of service guarantees. In the process of addressing these pertinent issues for our target user base (i.e. medical professionals and researchers), we provide insight that benefits a large community of scientific application users in industry and academia. Our execution time prediction and scheduling methodologies are implemented and evaluated on a real system running popular scientific applications. We find that we are able to predict the execution time of a number of these applications with an average error of 15%. Our scheduling methodology, which is tested with medical image processing workloads, is compared to that of two baseline scheduling solutions and we find that it outperforms them in terms of both the number of jobs processed and resource utilization by 20-30%, without violating any deadlines. We conclude that our solution is a viable approach to supporting the computational needs of medical users, even if the cloud computing paradigm is not widely adopted in its current form.

<p>This thesis gives a partial answer to the question: “Can type systems detect modeling errors in scientific computing, particularly for inverse problems derived from physical models?” by considering, in detail, the major aspects of inverse problems in Magnetic Resonance Imaging (MRI). We define a type-system that can capture all correctness properties for MRI inverse problems, including many properties that are not captured with current type-systems, e.g., frames of reference. We implemented a type-system in the Haskell language that can capture the errors arising in translating a mathe- matical model into a linear or nonlinear system, or alternatively into an objective function. Most models are (or can be approximated by) linear transformations, and we demonstrate the feasibility of capturing their correctness at the type level using what is arguably the most difficult case, the (discrete) Fourier transformation (DFT). By this, we mean that we are able to catch, at compile time, all known errors in ap- plying the DFT. The first part of this thesis describes the Haskell implementation of vector size, physical units, frame of reference, and so on required in the mathemat- ical modelling of inverse problems without regularization. To practically solve most inverse problems, especially those including noisy data or ill-conditioned systems, one must use regularization. The second part of this thesis addresses the question of defining new regularizers and identifying existing regularizers the correctness of which (in our estimation) can be formally verified at the type level. We describe such Bayesian regularization schemes based on probability theory, and describe a novel simple regularizer of this type. We leave as future work the formalization of such regularizers.</p> / Master of Science (MSc)

Scientific datasets associated with a research project can proliferate over time as a result of activities such as sharing datasets among collaborators, extending existing datasets with new measurements, and extracting subsets of data for analysis. As such datasets begin to accumulate, it becomes increasingly difficult for a scientist to keep track of their derivation history, which complicates data sharing, provenance tracking, and scientific reproducibility. Understanding what relationships exist between datasets can help scientists recall their original derivation history. For instance, if dataset A is contained in dataset B, then the connection between A and B could be that A was extended to create B. We present a relationship-identification methodology as a solution to this problem. To examine the feasibility of our approach, we articulated a set of relevant relationships, developed algorithms for efficient discovery of these relationships, and organized these algorithms into a new system called ReConnect to assist scientists in relationship discovery. We also evaluated existing alternative approaches that rely on flagging differences between two spreadsheets and found that they were impractical for many relationship-discovery tasks. Additionally, we conducted a user study, which showed that relationships do occur in real-world spreadsheets, and that ReConnect can improve scientists' ability to detect such relationships between datasets. The promising results of ReConnect's evaluation encouraged us to explore a more automated approach for relationship discovery. In this dissertation, we introduce an automated end-to-end prototype system, ReDiscover, that identifies, from a collection of datasets, the pairs that are most likely related, and the relationship between them. Our experimental results demonstrate the overall effectiveness of ReDiscover in predicting relationships in a scientist's or a small group of researchers' collections of datasets, and the sensitivity of the overall system to the performance of its various components.

High performance and scientific computing take advantage of high-end and high-spec computer architectures. As these architectures evolve, and new architectures are created, applications may be able to run at greater and greater speeds. These changes persent challenges to implementors who wish to take advantage of the newest features and machines. Portability layers such as RAJA and Agency seek to abstract away machine-specific details and allow scientists to take advantage of new features as they become available. We enhance RAJA with a lower-level framework, Agency, to determine if these layered abstractions provide performance or maintainability benefits.

Many quantum information protocols rely on the ability to distinguish between entangled quantum states known as Bell states. However, theoretical limits exist on the maximal distinguishability of these entangled states using linear evolution and local measurement (LELM) devices. In the case of two particles entangled in multiple qubit variables, the maximum number of distinguishable Bell states is known. However, in the more general case of two particles entangled in multiple qudit variables, only an upper bound is known under additional assumptions. I have written software in Matlab and Mathematica to explore computationally the maximum number of Bell states that can be distinguished in the case of two particles entangled in a qutrit variable, and the case of two particles entangled in both a qutrit and qubit variable. Using code I have written in Mathematica, I have reduced the number of cases to check for sets of 9 qubit x qutrit Bell states from 94,143,280 to 10,365. Further work needs to be done to computationally check these cases for distinguishability by an LELM apparatus.

Le code source de la libraire développée accompagne ce dépôt dans l'état où il était à ce moment. Il est possible de trouver une version plus à jour sur github (http://github.com/abergeron). / Le calcul scientifique sur processeurs graphiques (GPU) est en plein essor depuis un certain temps, en particulier dans le domaine de l'apprentissage machine. Cette thèse présente les efforts pour établir une structure de données de table au multidimensionnel de manière efficace sur GPU. Nous commençons par faire une revue de ce qui est actuellement similaire dans le domaine et des désavantages d'avoir une multitude d'approches. Nous nous intéresserons particulièrement aux calculs fait à partir du langage Python. Nous décrirons des techniques intéressantes telles que la réduction d'ordre et le calcul asynchrone automatique. Pour terminer nous présenterons l'utilisation du module développé dans le cadre de cette thèse. / Scientific computing on GPU (graphical processing units) is on the rise, specifically in machine learning. This thesis presents the implementation of an efficient multidimensional array on the GPU. We will begin by a review of what currently implements similar functionality and the disadvantage of a fragmented approach. We will focus on packages that have a Python interface. We will explain techniques to optimize execution such as order reduction and automatic asynchronous computations. Finally, we will present the functionality of the module developed for this thesis.

Dans le cadre du génie logiciel en traitement d'images (TDI), nous nous intéressons à la notion de réutilisabilité des algorithmes. Dans de nombreux outils logiciels, l'implémentation d'un algorithme est souvent dépendante du type des données traitées. Au sens le plus général, les formes que peuvent prendre les images numériques discrètes sont nombreuses (image 2D classiques, volumes 3D, graphes non réguliers, complexes cellulaires, etc.) conduisant à une explosion combinatoire du nombre théorique d'implémentations. La programmation générique (PG) est un cadre adapté au développement d'outils logiciels réutilisables. Nous présentons un paradigme de programmation basé sur la PG conçu pour la création de logiciels scientifiques tels ceux dédiés au TDI. Cette approche concilie réutilisabilité, puissance d'expression, extensibilité et performance. Nous proposons ensuite une architecture logicielle pour le TDI basée sur ce paradigme de programmation, s'appuyant sur une bibliothèque générique de TDI. Les fondations de ce cadre définissent des concepts fondamentaux du TDI, qui permettent l'écriture d'algorithmes réutilisables sur de nombreux types d'images. Nous présentons enfin une stratégie pour construire des outils haut niveau au dessus de cette bibliothèque tels que des ponts vers des langages dynamiques ou des interfaces graphiques. Ce mécanisme est conçu pour préserver la généricité et la performance des outils logiciels sous-jacents, tout en permettant un usage plus simple et plus flexible de ceux-ci / In the context of software engineering for image processing (IP), we consider the notion of reusability of algorithms. In many software tools, an algorithm's implementation often depends on the type of processed data. In a broad definition, discrete digital images may have various forms : classical 2D images, 3D volumes, non-regular graphs, cell complexes, and so on : thus leading to a combinatorial explosion of the theoretical number of implementations. Generic programming (GP) is a framework suited to the development of reusable software tools. We present a programming paradigm based on GP designed for the creation of scientific software such as IP tools. This approach combines the benefits of reusability, expressive power, extensibility, and efficiency. We then propose a software architecture for IP using this programming paradigm based on a generic IP library. The foundations of this framework define essential IP concepts, enabling the development of algorithms compatible with many image types. We finally present a strategy to build high-level tools on top of this library, such as bridges to dynamic languages or graphical user interfaces. This mechanism has been designed to preserve the genericity and efficiency of the underlying software tools, while making them simpler to use and more flexible

The United States Army Corp of Engineers relies on accurate and detailed surface models for various construction projects and preventative measures. To aid in these efforts, it is necessary to work for advancements in surface model creation. Current methods for model creation include Delaunay triangulation, raster grid interpolation, and Hydraulic Spline grid generation. While these methods produce adequate surface models, attempts for improved methods can still be made. A method for raster based spline creation is presented as a variation of the Hydraulic Spline algorithm. By implementing Hydraulic Splines in raster data instead of vector data, the model creation process is streamlined. This method is shown to be more efficient and less computationally expensive than previous methods of surface model creation due to the inherent advantages of raster data over vector data.

In this thesis, we propose an ε-Stokes problem connecting the Stokes problem and the corresponding pressure-Poisson equation using one pa- rameter ε &gt; 0. We prove that the solution to the ε-Stokes problem, converges as ε tends to 0 or ∞ to the Stokes and pressure-Poisson prob- lem, respectively. Most of these results are new. The precise statements of the new results are given in Proposition 3.5, Theorem 4.1, Theorem 5.2, and Theorem 5.3. Numerical results illustrating our mathematical results are also presented. / STINT (DD2017-6936) "Mathematics Bachelor Program for Efficient Computations"

A computação em nuvem tem recebido um grande destaque, ao propor um novo e eficiente mecanismo para disponibilizar recursos computacionais. Dos investimentos cada vez maiores nessa plataforma, inclusive pela academia, surge a oportunidade de compartilhar estes recursos computacionais entre diferentes instituições. As grades computacionais são um mecanismo bem estabelecido para o compartilhamento de alguns tipos de recursos computacionais. Através do entendimento de como isso é feito nestas estruturas, esse trabalho avalia as soluções existentes e propõe um arquitetura alternativa que permite a criação das federações de nuvens computacionais. / Cloud computing is a new model to provide computing resources. The growing interest and investments on this platform creates an opportunity to share this resources across different institutions. The grid computing is the standard way of achieving this. Using grid as reference, this work survey current technologies and present an alternative design to allow the development of academic cloud computing federations.