Scientific and engineering problems are often large, complex, irregular and data-parallel. The performance of many parallel applications is affected by factors such as irregular nature of the problem, the difference in processor characteristics and runtime loads, the non-uniform distribution of data, and the unpredictable system behavior. These factors give rise to load imbalance. In general, in order to achieve high performance, dynamic load balancing strategies are embedded into solution algorithms. Over time, a number of dynamic load balancing algorithms have been implemented into software tools and successfully used in scientific applications. However, most of these dynamic load balancing tools use an iterative static approach that does not address irregularities during the application execution, and the scheduling overhead incurred is high. During the last decade, a number of dynamic loop scheduling strategies have been proposed to address causes of load imbalance in scientific applications running in parallel and distributed environments. However, there is no single strategy that works well for all scientific applications, and it is up to the user to select the best strategy and integrate it into the application. In most applications using dynamic load balancing, the load balancing algorithm is directly embedded in the application, with close coupling between the data structures of the application and the load balancing algorithm. This typical approach leads to two disadvantages. First, the integration of each newly developed load balancing algorithm into the application needs to be performed from scratch. Second, it is unlikely that the user has incorporated the optimal load balancing algorithm into the application. Moreover, in a certain application (of various problem sizes and number of processors), it is difficult to assess in advance the advantage of incorporating one load balancing algorithm versus another. To overcome these drawbacks, there is a need for developing an application programming interface (API) for dynamic load balancing scientific applications using the recently developed dynamic loop scheduling algorithms. This thesis describes the design and development of such an API, called ProLAS, which is scalable, and independent of data structures of a host application. ProLAS performance is evaluated theoretically and experimentally (after being used in scientific applications). A qualitative and quantitative analysis of ProLAS is presented by comparing its performance with the state of the art technology in dynamic load balancing tools (e.g. CHARM++ library) for parallel applications. The analysis of the experimental results of using ProLAS in a few scientific aplications indicate that it consistently outperforms the existing technology in dynamic load balancing.

Parallel and distributed environments are used to solve large scientific and engineering problems that often are irregular and data parallel. However, performance of many parallel applications is affected by computation overheads, communication time and load imbalance. Among these factors, load imbalance is caused by the irregular nature of the problem, its algorithm, the difference in processor characteristics, and runtime loads. A number of applications achieve load balancing by one-time assignment of task. However, a number of applications have workloads that are unpredictable, and vary over the course of their execution. For such type of applications, load balancing is achieved by dynamic assignment of tasks at runtime. A large group of scientific applications has parallel loops as major source of concurrency. However, due to the irregular execution times of the loops, it is difficult to achieve optimal performance without dynamic load balancing. There are number of dynamic load balancing tools and libraries have been developed for different kind of applications. However these libraries fail to address all three degradation factors i.e. problem, algorithmic, and systemic. In this thesis a dynamic load balancing library called LB_Migrate is presented which addresses the degradation factors in application with parallel loops. The library provides a range of dynamic scheduling techniques and data migration strategies to achieve effective load balancing. It is designed to be independent of the host application data structure hence providing the flexibility to be used with different applications. The analysis of the experimental results using LB_Migrate with different applications indicates consistent performance improvement, and low overhead cost by the use of the library.

No description available.

Denna studie behandlar de statliga intressen som bidrar till att Nato fortsätter att vara en attraktiv institution för stater, då hotet den bildades för att möta har försvunnit. Jag har utgått ifrån min hypotes att det finns varierande intressen hos staterna beroende på deras kapacitet och maktposition i systemet, som gör det lönsamt för staterna att fortsätta verka inom institutionen. I studien används realismen för att finna och identifiera dessa intressen hos USA, Frankrike och Polen som alla är medlemmar av Nato, men även den liberala institutionalismen avvänds som ett komplement till realismen. Mina resultat bekräftade min hypotes om staternas egenintressen i organisationen, och visade att Nato främst är en arena för staterna att utöva och utöka sin makt och sitt inflytande över andra stater. I USA:s fall handlar det om att fördröja och kanske förhindra att ett europeiskt försvarssamarbete tar form och utvecklar kapacitet att utmana deras unipolära ställning. De två andra staterna har valt att två olika hållningar mot USA, där Frankrike använder sig utav balancing för att begränsa USA:s inflytande och Polen använder sig utav en bangwagoning strategi gentemot USA, för att vinna fördelar. / This essay considers the national interests of the states who are members in NATO and their incitement to continue their membership, though the threat that the alliance was formed to counter has dissolved. With a starting point in my theory I have derived a hypothesis that there are self – centred interests of the states that would explain the persistence of Nato, and these interests varies between the states because of their different capacities and power position. For this study I have used the realist theory to find and identify these self – centred interests within USA, France and Poland who all are members of NATO. I have also used the liberal institutional theory to complement the realist theory. As a result of this essay I found that my hypothesis was valid, and showed that NATO was primarily an arena for the states to exercise and pursue power and influence. I found that the US interest in NATO primary concerns the possibility to delay or prevent the Europeans to adapt a European security organisation that might challenge the US unipolar position. France and Poland has adopted different strategies against the US. France has adopted a balancing strategy to oppose the US, while Poland is bandwagoning with the US to gain some advantages.

Current feeding strategies aim to maximize efficiency at the pen level. However, feed intake varies across animals and in response to diet composition, making it difficult to capture these variations and control feeding effectively. A precision feeding system is required to feed animals individually, continuously monitor responses, and make timely adjustments to feed tailoring. Such a system would efficiently integrate dairy operations to enhance profitability and reduce their environmental footprint. Thus, the objectives of this dissertation were to build, test, and apply a precision feeding system able to tailor feeding strategies to animals more precisely and closely match their individual requirements. In Chapter 3, we describe the precision feeding system framework using directional data streams. The system integrates real-time farm data, segmented into data-analytic modules for independent testing and troubleshooting. It provides feeding instructions to automatic feeders and generates animal and financial monitoring reports. In Chapter 4, we describe the "Animal Performance" system module. This study developed a predictive model to estimate individual dry matter intake (DMI) by integrating markers, animal characteristics, dietary nutrient concentrations, and chewing sensor data. The performance of the developed model was then assessed and contrasted with the NASEM (2021) DMI equations. By incorporating covariates derived from short-term use of external and internal markers we demonstrated a greater accuracy of DMI predictions when using a fixed effects model, supporting its predictive capabilities for further application. In Chapter 5, we describe the "Diet Optimization" systems module, used to maximize profit by optimizing rations using a developed compact-vectorized version of NASEM (2021). The study aimed to simulate optimized diets, evaluate the economic impact of feeding individual diet, compare feed costs and income over feed cost (IOFC) for optimized group diet, and compare optimized diets against pen-averages (PEN). The results showed that IND diets had lower costs, higher milk production, and increased IOFC compared to CLU diets. Additionally, both IND and CLU diets outperformed PEN solutions. This work established methods for deriving efficient diet solutions for individual animals and using clustering techniques for more precise pen-level feeding. In Chapter 6, we describe the application of "Animal Performance", "Diet Optimization", and "Nutrient Titration" system modules. The former DMI model described in Chapter 4 was applied to the experimental data. The middle utilized optimized diets generated by the optimizer developed in Chapter 5, with additional algorithm updates. The latter aimed to investigate individual milk true protein production responses of dairy cows to varying levels of metabolizable protein (MP) and rumen-protected amino acids (RPAA) using automatic feeding systems and rank animals based on their individual gross milk protein efficiencies. Results demonstrated heterogeneous animal responses across MP and RPAA levels, ranging from linear, and quadratic to no response, emphasizing the necessity of addressing individual variability within a common pen. High-efficiency animals behaved consistently across MP treatments with lower variability, while low-efficiency animals showed high variability but consistently remained in the bottom efficiency rank. In conclusion, the precision feeding system underscores true capabilities to tailor nutrient delivery to individual cows, maximizing economic and environmental benefits, and sets the stage for future research focused on further refinement and automation of these technologies / Doctor of Philosophy / Feeding practices for dairy cattle have evolved significantly from manual grain mix offering to group feeding. While pen-level feeding has its benefits, it overlooks opportunities to maximize efficiency and minimize feed waste and nutrient excretion by not using individual animal variation to apply control feeding. With modern farms and increased technology adoption, feeding animals while being individually milked, even when group-housed, is now possible, leveraging this variability to apply precision feeding. In Chapter 3, we described the development of a precision feeding system that leverages technological advancements on dairy farms to gather and analyze data, supporting informed decision-making. This system includes various modules for testing and adjusting feeding strategies according to animal needs, providing feeding instructions to automatic feeders, and generating reports to help farmers monitor their animals and manage costs. Recognizing that precision feeding relies on quality data and accurate predictions of crucial metrics such as dry matter intake (DMI), Chapter 4 focused on developing a mathematical equation to predict DMI on an individual animal basis. This model demonstrated potential for commercial dairy operations due to its use of readily available farm-level predictors and its adequate performance compared to gold-standard field equations. Given the lack of efficient optimizers that incorporate individual animal data, in Chapter 5, we described the development of a new optimizer incorporated into the system to maximize profits. We simulated different feeding strategies, including optimized individual and group diets, and demonstrated that these tailored diets were more cost-effective and led to higher milk production compared to pen-average diets. To complete the development, testing, and application cycle, in Chapter 6, we applied the precision feeding system to determine the metabolizable protein (MP) requirements of dairy cows and assess milk protein production responses to different levels of MP and rumen-protected amino acids (RPAA). The results indicated varied responses among cows, highlighting the importance of individualized feeding to account for animal-to-animal differences within the same pen. Top-efficient animals were consistent in their responses across treatments, whereas bottom-efficient animals exhibited greater variability and consistently underperformed. In conclusion, the precision feeding system demonstrated significant potential to improve the efficiency of dairy farming by more accurately meeting the specific needs of dairy cattle. Future research will focus on refining this system and further automating the process for broader farm applications.

GPU computing can significantly improve performance by taking advantage of massive parallelism of GPUs for data parallel applications. Computation in visualization applications is suitable for parallelization on the GPU, which can improve performance and interactivity in these applications. If used effectively, multiple GPUs can lead to a significant speedup over a single GPU. However, the use of multiple GPUs requires memory management, scheduling, and load balancing to ensure that a program takes full advantage of available processors. This work presents methods for data-driven and dynamic multi-GPU load balancing using a pipelined approach and a framework for use with different applications. Data-driven load balancing can improve utilization for applications by taking into account past performance for different combinations of input parameters. The dynamic load balancing method based on buffer fullness can adjust to workload changes at runtime to gain an additional performance improvement. This work provides a framework for load balancing to account for differing characteristics of applications. Implementation of a multi-GPU data structure allows for use of these load balancing methods in the framework. The effectiveness of the framework is demonstrated with performance results from interactive visualization that shows a significant speedup due to load balancing. / Master of Science

En établissant une relation entre une stratégie d’alliance et la mise en œuvre d’une politique étrangère dans un espace géographique limité, l’auteur entend interpréter des intentions (géo)politiques et place son étude en discussion avec les travaux théoriques centrés sur les rapports de force. Il reconstruit le contexte de la décision polonaise de faire alliance avec les Etats-Unis et montre l’importance de savoir si les comportements d’alliance et d’alignement sont connectés dans la logique des alliances ou si ce sont deux catégories exclusives l’une de l’autre. Il propose ensuite une relecture de la politique orientale polonaise en prenant notamment en compte l’histoire de la Pologne ainsi que les conceptions de ses élites dirigeantes afin de montrer que la stratégie d’alliance avec les Etats-Unis peut être comprise comme l’instrument privilégié de la politique orientale polonaise. Enfin, il identifie les paramètres qui permettent à la fois de corroborer l’existence d’une alliance polono-américaine opérante dans l’espace post-soviétique et d’estimer sa durabilité. Le cas polonais montre qu’il ne faut pas considérer tous les Etats comme s’ils étaient des puissances satisfaites qui cherchent surtout à maximiser leur sécurité plutôt que leur puissance, et qu’il est indispensable de penser le phénomène des alliances (géo)politiquement, historiquement, mais aussi culturellement pour être en mesure d’éviter le piège des faux paradigmes dont sont porteurs des discours construits a posteriori. / By establishing a connection between an alliance strategy and the implementation of a foreign policy in a limited geographical area, the author means to interpret (geo)political intentions and place his study in the context of theoretical works focussing on balance of power. He recontextualises the Polish decision to establish an alliance with the United States and shows the importance of knowing whether alliance and alignment behaviours are connected in the logic of alliances or whether they are two mutually exclusive categories. He then proposes to reconsider Polish eastern policy, paying particular attention to the history of Poland and the thinking of its ruling elites in order to demonstrate that the strategy of alliance with the United States can be regarded as the main instrument of Polish eastern policy. Finally, he identifies the parameters that make it possible both to confirm the existence of a Polish-American alliance operating within the post-Soviet area and to estimate its durability. The example of Poland shows that not all States can be regarded as self-satisfied powers mainly striving to maximise their security rather than their power and that the phenomenon of alliances must be considered in (geo-)political and historical, but also cultural terms in order to avoid the trap of false paradigms to which post facto discourses are prone.

A single-chip, hybrid, heterogeneous, and dynamic shared memory multiprocessor architecture is being developed which may be used for real-time and non-real-time applications. This architecture can execute any application described by a dataflow (process flow) graph of any topology; it can also dynamically reconfigure its structure at the node and processor architecture levels and reallocate its resources to maximize performance and to increase reliability and fault tolerance. Dynamic change in the architecture is triggered by changes in parameters such as application input data rates, process execution times, and process request rates. The architecture is a Hybrid Data/Command Driven Architecture (HDCA). It operates as a dataflow architecture, but at the process level rather than the instruction level. This thesis focuses on the development, testing and evaluation of a new graphic software (hdca) developed to first do a static resource allocation for the architecture to meet timing requirements of an application and then hdca simulates the architecture executing the application using statically assigned resources and parameters. While simulating the architecture executing an application, the software graphically and dynamically displays parameters and mechanisms important to the architectures operation and performance. The new graphical software is able to show system and node level dynamic capability of the HDCA. The newly developed software can model a fixed or varying input data rate. The model also allows fault tolerance analysis of the architecture.

Thesis (Ph.D.) -- McMaster University, 1997. / Includes bibliographical references. Also available via World Wide Web.

The purpose of this thesis is determinate the forces in crankshaft mechanism acting on new design engine from company Zetor, assessment of the state of balance, design modifications balancing units, calculation of modal analysis, stress analysis and strain control. The content of the work is the creation of technical documentation for the newly modified balancer shafts.