<p>Cloud computing appears to be a paradigm shift in service oriented computing. Massively scalable Cloud architectures are spawned by new business and social applications as well as Internet driven economics. Besides being inherently large scale and highly distributed, Cloud systems are almost always virtualized and operate in automated shared environments. The deployed Cloud services are still in their infancy and a variety of research challenges need to be addressed to predict their long-term behavior. Performance and dependability of Cloud services are in general stochastic in nature and they are affected by a large number of factors, e.g., nature of workload and faultload, infrastructure characteristics and management policies. As a result, developing scalable and predictive analytics for Cloud becomes difficult and non-trivial. This dissertation presents the research framework needed to develop high fidelity stochastic models for large scale enterprise systems using Cloud computing as an example. Throughout the dissertation, we show how the developed models are used for: (i) performance and availability analysis, (ii) understanding of power-performance trade-offs, (ii) resiliency quantification, (iv) cost analysis and capacity planning, and (v) risk analysis of Cloud services. In general, the models and approaches presented in this thesis can be useful to a Cloud service provider for planning, forecasting, bottleneck detection, what-if analysis or overall optimization during design, development, testing and operational phases of a Cloud.</p> / Dissertation

abstract: User satisfaction is pivotal to the success of mobile applications. At the same time, it is imperative to maximize the energy efficiency of the mobile device to ensure optimal usage of the limited energy source available to mobile devices while maintaining the necessary levels of user satisfaction. However, this is complicated due to user interactions, numerous shared resources, and network conditions that produce substantial uncertainty to the mobile device's performance and power characteristics. In this dissertation, a new approach is presented to characterize and control mobile devices that accurately models these uncertainties. The proposed modeling framework is a completely data-driven approach to predicting power and performance. The approach makes no assumptions on the distributions of the underlying sources of uncertainty and is capable of predicting power and performance with over 93% accuracy. Using this data-driven prediction framework, a closed-loop solution to the DEM problem is derived to maximize the energy efficiency of the mobile device subject to various thermal, reliability and deadline constraints. The design of the controller imposes minimal operational overhead and is able to tune the performance and power prediction models to changing system conditions. The proposed controller is implemented on a real mobile platform, the Google Pixel smartphone, and demonstrates a 19% improvement in energy efficiency over the standard frequency governor implemented on all Android devices. / Dissertation/Thesis / Doctoral Dissertation Computer Engineering 2017

Proton Exchange Membrane Fuel Cell (PEMFC) is the most commonly used type of hydrogen fuel cell and a promising solution for vehicular and stationary power applications. This research starts with an extensive review of the PEMFC research, including experimental testing, and performance modeling, and performance degradation modeling using relatively accurate and easy-to-use mechanistic models. Next, a new PEMFC performance degradation model is introduced by amending the semi-empirical, mechanistic performance model to support the design and control of PEMFC systems and fuel cell electric vehicles (FCEVs). The new model takes into account critical factors impacting PEMFC performance. The performance degradation due to the oxidation of catalyst platinum (Pt) and loss of active surface area is captured by fitting the degradation model parameters using experimental data to capture the observed PEMFC performance fading. The new performance degradation model is then tested and further improved under the four typical load modes that a PEMFC system experiences in a vehicular application under regular driving cycles. The model is also fitted with PEMFC experimental degradation data under different load modes to improve modeling accuracy. The new model is applied and tested using simulations of a representative FCEV. The actual power load on an 80 kW PEMFC system in the modeled FCEV was obtained using the Advanced Vehicle Simulator (ADVISOR) under the US EPA Urban Dynamometer Driving Schedule (UDDS). With the ability to predict the operation life of the PEMFC, the appropriate sizes of the PEMFC system and the energy storage system (ESS) can be determined. Improved power control and energy management can be developed to extend the operation life of the PEMFC and lower the lifecycle cost of the FCEV. / Graduate

The computational Grid has become an important infrastructure as an execution environment for scientific applications that require large amount of computing resources. Applications which would otherwise be unmanageable or take a prohibitively longer execution time under previous computing paradigms can now be executed efficiently on the Grid within a reasonable time. Multi-agent based simulation (MABS) is a methodology used to study and understand the dynamics of real world phenomena in domains involving interaction and/or cooperative problem solving where the participants are characterized by entities having autonomous and social behaviour. For certain domains the size of the simulation is extremely large, intractable without employing adequate computing resources such as the Grid. Although the Grid has come with immense opportunities to resource demanding applications such as MABS, it has also brought with it a number of challenges related to performance. Performance problems may have their origins either on the side of the computing infrastructure or the application itself, or both. This thesis aims at improving the performance of MABS applications by overcoming problems inherent to the behaviour of MABS applications. It also studies the extent to which the MABS technologies have been exploited in the field of simulation and find ways to adapt existing technologies for the Grid. It investigates performance monitoring and prediction systems in the Grid environment and their implementation for MABS application with the purpose of identifying application related performance problems and their solutions. Our research shows that large-scale MABS applications have not been implemented despite the fact that many problem domains that cannot be studied properly with only partial simulation. We assume that this is due to the lack of appropriate tools such as MABS platforms for the Grid. Another important finding of this work is the improvement of application performance through the use of MABS specific middleware.

By 2004, microprocessor design focused on multicore scaling"increasing the number of cores per die in each generation "as the primary strategy for improving performance. These multicore processors typically equip multiple memory subsystems to improve data throughput. In addition, these systems employ heterogeneous processors such as GPUs and heterogeneous memories like non-volatile memory to improve performance, capacity, and energy efficiency. With the increasing volume of hardware resources and system complexity caused by heterogeneity, future systems will require intelligent ways to manage hardware resources. Early research to improve performance and energy efficiency on heterogeneous, multi-core, multi-memory systems focused on tuning a single primitive or at best a few primitives in the systems. The key limitation of past efforts is their lack of a holistic approach to resource management that balances the tradeoff between performance and energy consumption. In addition, the shift from simple, homogeneous systems to these heterogeneous, multicore, multi-memory systems requires in-depth understanding of efficient resource management for scalable execution, including new models that capture the interchange between performance and energy, smarter resource management strategies, and novel low-level performance/energy tuning primitives and runtime systems. Tuning an application to control available resources efficiently has become a daunting challenge; managing resources in automation is still a dark art since the tradeoffs among programming, energy, and performance remain insufficiently understood. In this dissertation, I have developed theories, models, and resource management techniques to enable energy-efficient execution of parallel applications through thread and data management in these heterogeneous multi-core, multi-memory systems. I study the effect of dynamic concurrent throttling on the performance and energy of multi-core, non-uniform memory access (NUMA) systems. I use critical path analysis to quantify memory contention in the NUMA memory system and determine thread mappings. In addition, I implement a runtime system that combines concurrent throttling and a novel thread mapping algorithm to manage thread resources and improve energy efficient execution in multi-core, NUMA systems. In addition, I propose an analytical model based on the queuing method that captures important factors in multi-core, multi-memory systems to quantify the tradeoff between performance and energy. The model considers the effect of these factors in a holistic fashion that provides a general view of performance and energy consumption in contemporary systems. Finally, I focus on resource management of future heterogeneous memory systems, which may combine two heterogeneous memories to scale out memory capacity while maintaining reasonable power use. I present a new memory controller design that combines the best aspects of two baseline heterogeneous page management policies to migrate data between two heterogeneous memories so as to optimize performance and energy. / Ph. D.

Projecting performance of applications and hardware is important to several market segments—hardware designers, software developers, supercomputing centers, and end users. Hardware designers estimate performance of current applications on future systems when designing new hardware. Software developers make performance estimates to evaluate performance of their code on different architectures and input datasets. Supercomputing centers try to optimize the process of matching computing resources to computing needs. End users requesting time on supercomputers must provide estimates of their application’s run time, and incorrect estimates can lead to wasted supercomputing resources and time. However, application performance is challenging to predict because it is affected by several factors in application code, specifications of system hardware, choice of compilers, compiler flags, and libraries. This dissertation uses statistical techniques to model and optimize performance of scientific applications across different computer processors. The first study in this research offers statistical models that predict performance of an application across different input datasets prior to application execution. These models guide end users to select parameters that produce optimal application performance during execution. The second study offers a suite of statistical models that predict performance of a new application on a new processor. Both studies present statistical techniques that can be generalized to analyze, optimize, and predict performance of diverse computation- and data-intensive applications on different hardware.

The Fukushima Daiichi nuclear accidents in 2011 sparked considerable interest in the U.S. to develop new nuclear fuel with enhanced accident tolerance. Throughout the development of these new fuel concepts they will be extensively modeled using specialized computer codes and experimentally tested for a variety of different postulated accident scenarios. One accident scenario of interest, reactivity-initiated accident, is a nuclear reactor event involving a sudden increase in fission rate that causes a rapid increase in reactor power and temperature of the fuel which can lead to the failure of the fuel rods and are lease of radioactive material. The focus of this work will be on the fuel performance modeling of reactivity-initiated accidents using the BISON code being developed at Idaho National Laboratory. The overall goal of this work is to provide the best possible modeling predictions for future experimental tests. Accurate predictive capability modeling using BISON is important for safe operation of these tests and provides a cheaper alternative to the expensive experiments.

Emergent systems in high-performance computing (HPC) expect maximal efficiency to achieve the goal of power budget under 20-40 megawatts for 1 exaflop set by the Department of Energy. To optimize efficiency, emergent systems provide multiple power-performance control techniques to throttle different system components and scale of concurrency. In this dissertation, we focus on three throttling techniques: CPU dynamic voltage and frequency scaling (DVFS), dynamic memory throttling (DMT), and dynamic concurrency throttling (DCT). We first conduct an empirical analysis of the performance and energy trade-offs of different architectures under the throttling techniques. We show the impact on performance and energy consumption on Intel x86 systems with accelerators of Intel Xeon Phi and a Nvidia general-purpose graphics processing unit (GPGPU). We show the trade-offs and potentials for improving efficiency. Furthermore, we propose a parallel performance model for coordinating DVFS, DMT, and DCT simultaneously. We present a multivariate linear regression-based approach to approximate the impact of DVFS, DMT, and DCT on performance for performance prediction. Validation using 19 HPC applications/kernels on two architectures (i.e., Intel x86 and IBM BG/Q) shows up to 7% and 17% prediction error correspondingly. Thereafter, we develop the metrics for capturing the performance impact of DVFS, DMT, and DCT. We apply the artificial neural network model to approximate the nonlinear effects on performance impact and present a runtime control strategy accordingly for power capping. Our validation using 37 HPC applications/kernels shows up to a 20% performance improvement under a given power budget compared with the Intel RAPL-based method. / Ph. D. / System efficiency on high-performance computing (HPC) systems is the key to achieving the goal of power budget for exascale supercomputers. Techniques for adjusting the performance of different system components can help accomplish this goal by dynamically controlling system performance according to application behaviors. In this dissertation, we focus on three techniques: adjusting CPU performance, memory performance, and the number of threads for running parallel applications. First, we profile the performance and energy consumption of different HPC applications on both Intel systems with accelerators and IBM BG/Q systems. We explore the trade-offs of performance and energy under these techniques and provide optimization insights. Furthermore, we propose a parallel performance model that can accurately capture the impact of these techniques on performance in terms of job completion time. We present an approximation approach for performance prediction. The approximation has up to 7% and 17% prediction error on Intel x86 and IBM BG/Q systems respectively under 19 HPC applications. Thereafter, we apply the performance model in a runtime system design for improving performance under a given power budget. Our runtime strategy achieves up to 20% performance improvement to the baseline method.

With the advent of exascale architectures maximizing performance while maintaining energy consumption within reasonable limits has become one of the most critical design constraints. This constraint is particularly significant in light of the power budget of 20 MWatts set by the U.S. Department of Energy for exascale supercomputing facilities. Therefore, understanding an application's characteristics, execution pattern, energy footprint, and the interactions of such aspects is critical to improving the application's performance as well as its utilization of the underlying resources. With conventional methods of analyzing performance and energy consumption trends scientists are forced to limit themselves to a manageable number of design parameters. While these modeling techniques have catered to the needs of current high-performance computing systems, the complexity and scale of exascale systems demands that large-scale design-space-exploration techniques are developed to enable comprehensive analysis and evaluations. In this dissertation we present research on performance and energy modeling of current high performance computing and future exascale systems. Our thesis is focused on the design space exploration of current and future architectures, in terms of their reconfigurability, application's sensitivity to hardware characteristics (e.g., system clock, memory bandwidth), application's execution patterns, application's communication behavior, and utilization of resources. Our research is aimed at understanding the methods by which we may maximize performance of exascale systems, minimize energy consumption, and understand the trade offs between the two. We use analytical, statistical, and machine-learning approaches to develop accurate, portable and scalable performance and energy models. We develop application and machine abstractions using Aspen (a domain specific language) to implement and evaluate our modeling techniques. As part of our research we develop and evaluate system-level performance and energy-consumption models that form part of an automated modeling framework, which analyzes application signatures to evaluate sensitivity of reconfigurable hardware components for candidate exascale proxy applications. We also develop statistical and machine-learning based models of the application's execution patterns on heterogeneous platforms. We also propose a communication and computation modeling and mapping framework for exascale proxy architectures and evaluate the framework for an exascale proxy application. These models serve as external and internal extensions to Aspen, which enable proxy exascale architecture implementations and thus facilitate design space exploration of exascale systems. / Ph. D.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21280号 / 情博第674号 / 新制||情||116(附属図書館) / 京都大学大学院情報学研究科システム科学専攻 / (主査)教授 田中 利幸, 教授 山下 信雄, 准教授 増山 博之, 教授 笠原 正治 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM