Browsing by Subject "High performance computing"
Now showing 1 - 11 of 11
Results Per Page
Sort Options
Item A high performance class-D amplifier with cascaded sigma-delta modulators(Texas Tech University, 2004-05) Trehan, ChintanThe focus of this thesis is on analysis, simulation and board level implementation of the proposed Class-D power amplifier architecture. The structural design consists of two Sigma-Delta Modulator (SDM) stages in cascade with an intermediate decimation-filter between them. Noise and high tone introduced at the first- stage is filter out through the decimation filter. The signal is converted to a 1-bit Pulse Duration Modulation (PDM) signal by the second stage SDM. The H-Bridge is made part of the SD loop, which enables not only the noise shaping of the quantization noise but also stabilizes the output power switching stage. Output of the H-Bridge is converted to a digital signal using a comparator and latch circuitry and is fed back. To further increase the linearity and performance, high frequency ripples introduced at the H-Bridge is quantized by using a 4-bit SD Analog-to-Digital Converter (ADC) in the feedback loop. Due to the intermediate digital stage and the feedback control at the output stage, the proposed structure has high efficiency and linearity and still is very compact making it possible for wide range of applications.Item A petri net based tool for the analysis of task graph systems(Texas Tech University, 1998-05) Dova, HemchandThe goal of this research is to develop a software tool for the analysis of task graph systems with a graphical user interface. PN's directly incorporate the topological information of the input task graph and also accommodate means to include parameters such as the processor heterogeneity, allocation schemes, communication costs, and random execution times. A technique for the analysis of a task graph systems using PN's has been reported in [6]. The methodology described in this thesis provides a graphical user interface tool for the analysis of the task graph system. The technique reported in [6] and [14] has been used in this thesis for the numerical and simulation analysis of the input task graph systenL The limitation of the numerical method of analysis is that tasks having only exponentially distributed execution times can be analyzed. This restriction is overcome in the simulation method of analysis where tasks having normal distributions can also be analyzed. The results from the numerical and simulation analysis are used to validate the tool. The software is implemented using JAVA. In this thesis three allocation heuristics are also implemented. The allocation heuristics are part of the topology of task graph systems. The Petri net tool generates the allocation of the tasks onto the processors. The software tool is called TASK GRAPH SYSTEM ANALYZER (TGSA). Furthermore, the implementation methodology of the present thesis can also be easily adapted to accommodate additional methods of analysis that use the same topology for the task graph system. Additional factors that could help in the study of task graph systems can also be incorporated.Item A computational framework for the solution of infinite-dimensional Bayesian statistical inverse problems with application to global seismic inversion(2015-08) Martin, James Robert, Ph. D.; Ghattas, Omar N.; Biros, George; Demkowicz, Leszek; Fomel, Sergey; Marzouk, Youssef; Moser, RobertQuantifying uncertainties in large-scale forward and inverse PDE simulations has emerged as a central challenge facing the field of computational science and engineering. The promise of modeling and simulation for prediction, design, and control cannot be fully realized unless uncertainties in models are rigorously quantified, since this uncertainty can potentially overwhelm the computed result. While statistical inverse problems can be solved today for smaller models with a handful of uncertain parameters, this task is computationally intractable using contemporary algorithms for complex systems characterized by large-scale simulations and high-dimensional parameter spaces. In this dissertation, I address issues regarding the theoretical formulation, numerical approximation, and algorithms for solution of infinite-dimensional Bayesian statistical inverse problems, and apply the entire framework to a problem in global seismic wave propagation. Classical (deterministic) approaches to solving inverse problems attempt to recover the “best-fit” parameters that match given observation data, as measured in a particular metric. In the statistical inverse problem, we go one step further to return not only a point estimate of the best medium properties, but also a complete statistical description of the uncertain parameters. The result is a posterior probability distribution that describes our state of knowledge after learning from the available data, and provides a complete description of parameter uncertainty. In this dissertation, a computational framework for such problems is described that wraps around the existing forward solvers, as long as they are appropriately equipped, for a given physical problem. Then a collection of tools, insights and numerical methods may be applied to solve the problem, and interrogate the resulting posterior distribution, which describes our final state of knowledge. We demonstrate the framework with numerical examples, including inference of a heterogeneous compressional wavespeed field for a problem in global seismic wave propagation with 10⁶ parameters.Item Distributed selective re-execution for EDGE architectures(2005) Desikan, Rajagopalan; Burger, Douglas C., Ph. D.Item Jack Rabbit : an effective Cell BE programming system for high performance parallelism(2011-05) Ellis, Apollo Isaac Orion; Lin, Yun Calvin; Fussell, Donald S., 1951-The Cell processor is an example of the trade-offs made when designing a mass market power efficient multi-core machine, but the machine-exposing architecture and raw communication mechanisms of Cell are hard to manage for a programmer. Cell's design is simple and causes software complexity to go up in the areas of achieving low threading overhead, good bandwidth efficiency, and load balance. Several attempts have been made to produce efficient and effective programming systems for Cell, but the attempts have been too specialized and thus fall short. We present Jack Rabbit, an efficient thread pool work queue implementation, with load balancing mechanisms and double buffering. Our system incurs low threading overhead, gets good load balance, and achieves bandwidth efficiency. Our system represents a step towards an effective way to program Cell and any similar current or future processors.Item The Lagniappe programming environment(2008-08) Riché, Taylor Louis, 1978-; Vin, Harrick M.Multicore, multithreaded processors are rapidly becoming the platform of choice for designing high-throughput request processing applications. We refer to this class of modern parallel architectures as multi-[star] systems. In this dissertation, we describe the design and implementation of Lagniappe, a programming environment that simplifies the development of portable, high-throughput request-processing applications on multi-[star] systems. Lagniappe makes the following four key contributions: First, Lagniappe defines and uses a unique hybrid programming model for this domain that separates the concerns of writing applications for uni-processor, single-threaded execution platforms (single-[star]systems) from the concerns of writing applications necessary to efficiently execute on a multi-[star] system. We provide separate tools to the programmer to address each set of concerns. Second, we present meta-models of applications and multi-[star] systems that identify the necessary entities for reasoning about the application domain and multi-[star] platforms. Third, we design and implement a platform-independent mechanism called the load-distributing channel that factors out the key functionality required for moving an application from a single-[star] architecture to a multi-[star] one. Finally, we implement a platform-independent adaptation framework that defines custom adaptation policies from application and system characteristics to change resource allocations with changes in workload. Furthermore, applications written in the Lagniappe programming environment are portable; we separate the concerns of application programming from system programming in the programming model. We implement Lagniappe on a cluster of servers each with multiple multicore processors. We demonstrate the effectiveness of Lagniappe by implementing several stateful request-processing applications and showing their performance on our multi-[star] system.Item Memory management for high-performance applications(2002) Berger, Emery David; McKinley, Kathryn S.Item Performance enhancement of grid-based applications(Texas Tech University, 2004-05) Pathak, SameerNot availableItem Petrophysical modeling and simulatin study of geological CO₂ sequestration(2014-05) Kong, Xianhui; Delshad, Mojdeh; Wheeler, Mary F. (Mary Fanett)Global warming and greenhouse gas (GHG) emissions have recently become the significant focus of engineering research. The geological sequestration of greenhouse gases such as carbon dioxide (CO₂) is one approach that has been proposed to reduce the greenhouse gas emissions and slow down global warming. Geological sequestration involves the injection of produced CO₂ into subsurface formations and trapping the gas through many geological mechanisms, such as structural trapping, capillary trapping, dissolution, and mineralization. While some progress in our understanding of fluid flow in porous media has been made, many petrophysical phenomena, such as multi-phase flow, capillarity, geochemical reactions, geomechanical effect, etc., that occur during geological CO₂ sequestration remain inadequately studied and pose a challenge for continued study. It is critical to continue to research on these important issues. Numerical simulators are essential tools to develop a better understanding of the geologic characteristics of brine reservoirs and to build support for future CO₂ storage projects. Modeling CO₂ injection requires the implementation of multiphase flow model and an Equation of State (EOS) module to compute the dissolution of CO₂ in brine and vice versa. In this study, we used the Integrated Parallel Accurate Reservoir Simulator (IPARS) developed at the Center for Subsurface Modeling at The University of Texas at Austin to model the injection process and storage of CO₂ in saline aquifers. We developed and implemented new petrophysical models in IPARS, and applied these models to study the process of CO₂ sequestration. The research presented in this dissertation is divided into three parts. The first part of the dissertation discusses petrophysical and computational models for the mechanical, geological, petrophysical phenomena occurring during CO₂ injection and sequestration. The effectiveness of CO₂ storage in saline aquifers is governed by the interplay of capillary, viscous, and buoyancy forces. Recent experimental data reveals the impact of pressure, temperature, and salinity on interfacial tension (IFT) between CO₂ and brine. The dependence of CO₂-brine relative permeability and capillary pressure on IFT is also clearly evident in published experimental results. Improved understanding of the mechanisms that control the migration and trapping of CO₂ in the subsurface is crucial to design future storage projects for long-term, safe containment. We have developed numerical models for CO₂ trapping and migration in aquifers, including a compositional flow model, a relative permeability model, a capillary model, an interfacial tension model, and others. The heterogeneities in porosity and permeability are also coupled to the petrophysical models. We have developed and implemented a general relative permeability model that combines the effects of pressure gradient, buoyancy, and capillary pressure in a compositional and parallel simulator. The significance of IFT variations on CO₂ migration and trapping is assessed. The variation of residual saturation is modeled based on interfacial tension and trapping number, and a hysteretic trapping model is also presented. The second part of this dissertation is a model validation and sensitivity study using coreflood simulation data derived from laboratory study. The motivation of this study is to gain confidence in the results of the numerical simulator by validating the models and the numerical accuracies using laboratory and field pilot scale results. Published steady state, core-scale CO₂/brine displacement results were selected as a reference basis for our numerical study. High-resolution compositional simulations of brine displacement with supercritical CO₂ are presented using IPARS. A three-dimensional (3D) numerical model of the Berea sandstone core was constructed using heterogeneous permeability and porosity distributions based on geostatistical data. The measured capillary pressure curve was scaled using the Leverett J-function to include local heterogeneity in the sub-core scale. Simulation results indicate that accurate representation of capillary pressure at sub-core scales is critical. Water drying and the shift in relative permeability had a significant impact on the final CO₂ distribution along the core. This study provided insights into the role of heterogeneity in the final CO₂ distribution, where a slight variation in porosity gives rise to a large variation in the CO₂ saturation distribution. The third part of this study is a simulation study using IPARS for Cranfield pilot CO₂ sequestration field test, conducted by the Bureau of Economic Geology (BEG) at The University of Texas at Austin. In this CO₂ sequestration project, a total of approximately 2.5 million tons supercritical CO₂ was injected into a deep saline aquifer about ~10000 ft deep over 2 years, beginning December 1st 2009. In this chapter, we use the simulation capabilities of IPARS to numerically model the CO₂ injection process in Cranfield. We conducted a corresponding history-matching study and got good agreement with field observation. Extensive sensitivity studies were also conducted for CO₂ trapping, fluid phase behavior, relative permeability, wettability, gravity and buoyancy, and capillary effects on sequestration. Simulation results are consistent with the observed CO₂ breakthrough time at the first observation well. Numerical results are also consistent with bottomhole injection flowing pressure for the first 350 days before the rate increase. The abnormal pressure response with rate increase on day 350 indicates possible geomechanical issues, which can be represented in simulation using an induced fracture near the injection well. The recorded injection well bottomhole pressure data were successfully matched after modeling the fracture in the simulation model. Results also illustrate the importance of using accurate trapping models to predict CO₂ immobilization behavior. The impact of CO₂/brine relative permeability curves and trapping model on bottom-hole injection pressure is also demonstrated.Item Using high performance computing and visualization to enhance risk assessment methodology: case study with perchlorate(Texas Tech University, 2004-05) Albers, Eric PeterSite-specific risk assessments commonly result in large amounts of information that needs to be processed for a wide, often non-scientific, audience consisting of risk managers, regulators, and other decision makers. For this work we combined a series of models into a large virtual representation of the study system. By using a location-based approach, we were able to arrive at a more accurate determination of risk compared to just a maximum-dose approach. Caddo Lake at Longhom Army Ammunition Plant was used to study the impacts of perchlorate (C104') on thyroid hormone secretion in the channel catfish (Ictalurus punctatus). Two hypothetical contaminant plumes were modeled accounting for groundwater upwelling into the lake and effluent discharge near the surface. Results were compared between environmental systems and the three dosing techniques; maximum dose, time-lapsed maximum dose, and location-based dose. Perchlorate tissue concentrations for liver, kidney, gill, skin, muscle, GI tract, and thyroid, as well as thyroid hormone levels and secretion rates were simulated. We have shown that a standard maximum dose approach vastly overestimates exposure for individuals and populations. By simulating large numbers of individuals we are able to achieve low probability extreme events, thereby limiting the need for uncertainty factors. Through the use of commercially available graphics software Maya®, we were able to generate 3- dimensional visualizations of our study site, PBTK model, thyroid hormone secretion, catfish movement, and contaminant plumes, further aiding in data comprehension. This is the first study to generate a 3-dimensional PBTK with commercially available software, as well as use grid computing and 3-d visualization for risk assessment.Item xBFT : Byzantine fault tolerance with high performance, low cost, and aggressive fault isolation(2008-05) Kotla, Ramakrishna Rao, 1976-; Dahlin, MichaelWe are increasingly relying on online services to store, access, share, and disseminate critical information from anywhere and at all times. Such services include email, digital storage, photos, video, health and financial services, etc. With increasing evidence of non-fail-stop failures in practical systems, Byzantine fault tolerant state machine replication technique is becoming increasingly attractive for building highlyreliable services in order to tolerate such failures. However, existing Byzantine fault tolerant techniques fall short of providing high availability, high performance, and long-term data durability guarantees with competitive replication cost. In this dissertation, we present BFT replication techniques that facilitate the design and implementation of such highly-reliable services by providing high availability, high performance and high durability with competitive replication cost (hardware, software, network, management). First, we propose CBASE, a BFT state machine replication architecture that leverages application-level parallelism to improve throughput of the replicated system by identifying and executing independent requests concurrently. Traditional state machine replication based Byzantine fault tolerant (BFT) techniques provide high availability and security but fail to provide high throughput. This limitation stems from the fundamental assumption of generalized state machine replication techniques that all replicas execute requests sequentially in the same total order to ensure consistency across replicas. Our architecture thus provides a general way to exploit application parallelism in order to provide high throughput without compromising correctness. Second, we present Zyzzyva, an efficient BFT agreement protocol that uses speculation to significantly reduce the performance overhead and replication cost of BFT state machine replication. In Zyzzyva, replicas respond to a client’s request without first running an expensive three-phase commit protocol to reach agreement on the order in which the request must be processed. Instead, they optimistically adopt the order proposed by the primary and respond immediately to the client. Replicas can thus become temporarily inconsistent with one another, but clients detect inconsistencies, help correct replicas converge on a single total ordering of requests, and only rely on responses that are consistent with this total order. This approach allows Zyzzyva to reduce replication overheads to near their theoretical minima. Third, we design and implement SafeStore, a distributed storage system designed to maintain long-term data durability despite conventional hardware and software faults, environmental disruptions, and administrative failures caused by human error or malice. The architecture of SafeStore is based on fault isolation, which SafeStore applies aggressively along administrative, physical, and temporal dimensions by spreading data across autonomous storage service providers (SSPs). SafeStore also performs an efficient end-to-end audit of SSPs to detect data loss quickly and improve data durability by reducing MTTR. SafeStore offers durable storage with cost, performance, and availability competitive with traditional storage systems. We evaluate these techniques by implementing BFT replication libraries and further demonstrate the practicality of these approaches by implementing an NFS based replicated file system(CBASE-FS) and a durable storage system (SafeStore-FS).