Academic literature on the topic 'Workload'

Thesis (M.S.) -- University of Maryland, College Park, 2005.<br>Thesis research directed by: Dept. of Computer Science. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

The primary contribution of this thesis is the development of a new performance metric, Capacity, which evaluates the performance of Chip Heterogeneous Multiprocessors (CHMs) that process multiple heterogeneous channels. Performance metrics are required in order to evaluate any system, including computer systems. A lack of appropriate metrics can lead to ambiguous or incorrect results, something discovered while developing the secondary contribution of this thesis, that of workload modes for CHMs â or Workload Specific Processors (WSPs). For many decades, computer architects and designers have focused on techniques that reduce latency and increase throughput. The change in modern computer systems built around CHMs that process multi-channel communications in the service of single users calls this focus into question. Modern computer systems are expected to integrate tens to hundreds of processor cores onto single chips, often used in the service of single users, potentially as a way to access the Internet. Here, the design goal is to integrate as much functionality as possible during a given time window. Without the ability to correctly identify optimal designs, not only will the best performing designs not be found, but resources will be wasted and there will be a lack of insight to what leads to better performing designs. To address performance evaluation challenges of the next generation of computer systems, such as multicore computers inside of cell phones, we found that a structurally different metric is needed and proceeded to develop such a metric. In contrast to single-valued metrics, Capacity is a surface with dimensionality related to the number of input streams, or channels, processed by the CHM. We develop some fundamental Capacity curves in two dimensions and show how Capacity shapes reveal interaction of not only programs and data, but the interaction of multiple data streams as they compete for access to resources on a CHM as well. For the analysis of Capacity surface shapes, we propose the development of a demand characterization method in which its output is in the form of a surface. By overlaying demand surfaces over Capacity surfaces, we are able to identify when a system meets its demands and by how much. Using the Capacity metric, computer performance optimization is evaluated against workloads in the service of individual users instead of individual applications, aggregate applications, or parallel applications. Because throughput was originally derived by drawing analogies between processor design and pipelines in the automobile industry, we introduce our Capacity metric for CHMs by drawing an analogy to automobile production, signifying that Capacity is the successor to throughput. By developing our Capacity metric, we illustrate how and why different processor organizations cannot be understood as being better performers without both magnitude and shape analysis in contrast to other metrics, such as throughput, that consider only magnitude. In this work, we make the following major contributions: â ¢ Definition and development of the Capacity metric as a surface with dimensionality related to the number of input streams, or channels, processed by the CHM. â ¢ Techniques for analysis of the Capacity metric. Since the Capacity metric was developed out of necessity, while pursuing the development of WSPs, this work also makes the following minor contributions: â ¢ Definition and development of three foundations in order to establish an experimental foundation â a CHM model, a multimedia cell phone example, and a Workload Specific Processor (WSP). â ¢ Definition of Workload Modes, which was the original objective of this thesis. â ¢ Definition and comparison of two approaches to workload mode identification at run time; The Workload Classification Model (WCM) and another model that is based on Hidden Markov Models (HMMs). â ¢ Development of a foundation for analysis of the Capacity metric, so that the impact of architectural features in a CHM may be better understood. In order to do this, we develop a Demand Characterization Method (DCM) that characterizes the demand of a specific usage pattern in the form of a curve (or a surface in general). By doing this, we will be able to overlay demand curves over Capacity curves of different architectures to compare their performance and thus identify optimal performing designs.<br>Ph. D.

This doctoral thesis focuses on the modification of ACC to include actual driver workload in the context of automatic headway change. ACC is a driver assistance system that automatically maintains a preliminary defined speed and permits a driver to perform manual headway changes. As drivers show worse (brake) reaction times under high workload situations, the system increases headway accordingly. Driver workload is estimated based on physiological data. Here, we investigate the possibilities of such a system, assuming that physiological sensors can be implemented in future vehicles. The thesis consists of three parts: In the first part, the theoretical background is described and a suitable theoretical model is developed; in the second part, experiments are described, and in the last part, results are discussed. Altogether four experiments support this thesis: 1. The first experiment investigates the foundational relationships between physiology, brake reaction time and workload level. The study employs a simulator setting and results show that physiological data, such as heart rate and skin conductance, permit the identification of different workload levels. These findings validate the results of other studies showing that workload leads to an increase in reaction time. These results could only be validated between the extremes “no-workload” and “high workload” situations. 2. The second experiment simulates an ideal workload-adaptive cruise control (WACC) system. In a simulator setting, system acceptance and awareness are studied, with a view toward future implementation in a real car. The results show better acceptance of WACC in comparison with ACC when subjects receive additional information about the new system. This is because subjects do not perceive changes in distance under high workload conditions. 3. The third experiment focuses on acceptance of the simulated system in on-road conditions. In this study, WACC is integrated in the car and is operated using a MATLAB model. The experiment shows that more subjects notice changes in distance in the on-road condition. In general WACC is preferred over ACC; it is especially these subjects who do not notice changes in distance, who value WACC more than ACC. With the aim of implementing an operational WACC that is capable of adjusting distance according to changes in physiological data, a workload algorithm is developed. 4. The fourth experiment validates the workload algorithm. Results of the algorithm are compared with recordings of the activated workload task and detection rate is calculated. The detection of workload periods was feasible in nearly every case and detection rate was favorable, especially if one considers lags due to design-related latency periods. The experiments presented here indicate that workload is detectable in physiological data and that it influences brake reaction time. Further, we provide evidence pointing to the technical possibility of implementing WACC as well as positive acceptance. The results have been published as an article and are part of this thesis. Also, some parts of the thesis are published as a book chapter (see footnotes). Another publication is in preparation, coauthored by diploma thesis students, who are supervised by the author (consult footnotes). This dissertation is composed, in part, of these publications. References to page numbers of the diploma theses are given to ensure correspondence. The author escorted the topic WACC from the beginning to the end. Sometimes students were involved and intensively supervised, from a thematic as well as a personnel guidance perspective. The author planned the whole project and executed studies and calculations. His psychology insights were not only limited to the discipline of psychology but were furthermore, with the help of students, interdisciplinarily expanded to the subject of informatics. Every study and every result which is presented within this work, was conducted or achieved by the author or (if students supported him) was discussed with the author in weekly discussions (and often several times a day). In these discussions the author provided new ideas and corrections if necessary. Apart from that, the author looked after the fulfillment of the central theme, implemented his psychological knowledge on a daily basis and provided his expertise to complement interdisciplinary point of views. He discussed the central theme as well as details with external partners like the MIT AgeLab as well as professors of the European Union from the adaption project (a project aimed at educating future researchers which includes involvement of highly important commercial and educational partners) and beyond. In this time he also visited conferences and accumulated knowledge which led to the successful achievements of the main objective and he was relevant in reaching the common goals of the adaption project. Furthermore he presented the results of the scientific work on a conference, workshops and in written publications. Within BMW Group Research and Technology, he identified important department- and project- partners and combined the knowledge to a result which benefits science and economy<br>In dieser Dissertation wird eine Abwandlung des Active Cruise Control (ACC) untersucht, das zusätzlich die Belastung (Workload) des Fahrers als Parameter betrachtet, um den Abstand zum Vordermann automatisiert zu verändern. Bei diesem ACC handelt es sich um ein Fahrerassistenzsystem, das automatisiert die eingestellte Geschwindigkeit hält und eine manuelle (durch den Nutzer ausgelöste) Abstandsveränderung zum Vordermann ermöglicht. Da sich die Bremsreaktionszeit von Fahrern in hohen Belastungssituationen verschlechtert, soll das entwickelte Workload-adaptive Cruise Control (WACC) in Situationen hoher Belastung den Abstand zum Vordermann automatisiert erhöhen. Die Belastung des Fahrers soll durch physiologische Daten ermittelt werden. Die vorliegende Arbeit untersucht die Möglichkeit eines solchen Systems unter der Annahme, dass in Zukunft geeignete physiologische Sensoren ins Auto eingebaut werden können. Die Arbeit besteht aus drei Teilen: • Im ersten Teil wird der theoretische Hintergrund beschrieben und ein passendes theoretisches Modell entwickelt. • Im zweiten Teil werden die durchgeführten Experimente beschrieben. • Im dritten Teil werden die Ergebnisse diskutiert. Insgesamt wurden im Rahmen dieser Arbeit vier Experimente durchgeführt: Das erste Experiment beschäftigte sich mit den grundlegenden Zusammenhängen zwischen Physiologie, Bremsreaktionszeit und Belastungslevel. Wie die Ergebnisse der im Simulator durchgeführten Studie zeigen, können mit physiologischen Daten wie Herzrate, Herzratenvariabilität und Hautleitfähigkeit unterschiedliche Workloadlevel identifiziert werden. Darüber hinaus wurden die Ergebnisse anderer Studien bestätigt, die belegen, dass Workload die Bremsreaktionszeit erhöht, wobei dies nur im Kontrast zwischen den Extrembereichen „kein Workload“ und „hoher Workload“ nachweisbar ist. Das zweite Experiment diente der Simulierung eines perfekten WACC. Im Simulator wurden Akzeptanz und Systemwahrnehmung getestet, um vor der Implementierung in ein Realfahrzeug weitere Erkenntnisse zu gewinnen. Im Vergleich zum ACC wurde das WACC von den Probanden besser akzeptiert, nachdem sie zusätzliche Informationen zum neuen System erhalten hatten. Der wesentliche Grund dafür ist, dass die Probanden ohne Informationen die Abstandsveränderung bei hohem Workload nicht realisieren. Das dritte Experiment fokussierte auf die Akzeptanz des simulierten Systems unter Realbedingungen. Das WACC wurde in das Auto integriert und durch ein MATLAB Modell gesteuert. Als Ergebnis zeigte sich, dass unter Realbedingungen mehr Probanden die Abstandsveränderung realisieren als im Simulator. Generell wird das WACC präferiert – vor allem jene Probanden, die die Abstandsveränderung nicht realisieren, bewerten das WACC besser als das ACC. Mit den in diesem Teilexperiment erhobenen Daten wurde ein Algorithmus zur Workloaderkennung entwickelt. Auf dieser Basis konnte im letzten Experiment ein Realsystem implementiert werden, das aufgrund physiologischer Daten den Abstand verändert. Das vierte Experiment beschäftigte sich mit der Validierung des Algorithmus zur Workloaderkennung. Die Ergebnisse des Algorithmus wurden mit der aufgezeichneten Aktivierung des Workloadtasks verglichen und eine Detektionsrate ermittelt. Die Detektion der Workloadperioden gelingt in fast allen Fällen und die Detektionsrate ist vielversprechend, gerade wenn man Verzögerungen berücksichtigt, die wegen der Latenzzeit körperlicher Reaktionen nicht verbesserungsfähig sind. In den vorliegenden Experimenten konnte gezeigt werden, dass Workload über die Physiologie messbar ist und sich auf die Bremsreaktionszeit auswirkt. Darüber hinaus wurde gezeigt, dass ein WACC technisch machbar ist und die Ergebnisse lassen außerdem auf eine hohe Akzeptanz schließen. Die Forschungsergebnisse wurden in einem Artikel publiziert, der auch in dieser Dissertation zu finden ist. Teile der vorliegenden Arbeit wurden außerdem als Buchkapitel veröffentlicht (siehe Fußnoten), eine weitere Publikation mit den vom Autor umfassend betreuten Diplomanden ist in Ausarbeitung. Um die Nachvollziehbarkeit zu gewährleisten, wurde auf die Seitenzahlen der entsprechenden Diplomarbeiten verwiesen. Als Autor dieser Dissertation habe ich das Thema WACC von Anfang bis Ende selbst erarbeitet bzw. wurden Studenten eingesetzt und angeleitet, wo es sinnvoll erschien. Neben der fachlichen wie personellen Führung der Studenten umfassten meine Aufgaben die Planung des Gesamtprojekts, das Durchführen der Studien und die Berechnung von Kennzahlen – sowohl in meinem eigenen psychologischen Fachgebiet als auch interdisziplinär mit Hilfe von Experten aus der Informatik. Jedes Ergebnis, das in dieser Arbeit präsentiert wird, wurde entweder von mir selbst erzielt oder – sofern ich dabei von Studenten unterstützt wurde – mit mir in wöchentlichen (oft auch mehrmals täglich stattfindenden Meetings) besprochen. Wichtig war mir als Autor, einen durchgängigen Weg zur Entwicklung eines WACC zu wahren, mein psychologisches Fachwissen täglich anzuwenden und in interdisziplinären Aufgaben und Diskussionen meine Perspektive einzubringen. Besonders wichtig war dabei die Diskussion des Gesamtprojekts und der Details mit externen Partnern wie dem MIT AgeLab oder Professoren aus der Europäischen Union im Rahmen des Adaptation Projekts (ein von der EU gefördertes Projekt zur Ausbildung von Forschern, unter Einbindung wirtschaftlich und wissenschaftlich hochrangiger Partner). Durch die Präsentation der Ergebnisse auf Konferenzen, in Workshops und Publikationen konnte ich einen Beitrag dazu leisten, um die Adaptation-Ziele zu erreichen. Innerhalb der BMW Group Forschung und Technik habe ich darauf geachtet, relevante Schnittstellen- und Projektpartner zu identifizieren und das erlangte Wissen zu einem Ergebnis zu verbinden, das Wissenschaft und Wirtschaft gleichermaßen nützt

The present research developed and assessed the Team Workload Questionnaire (TWLQ). Despite extensive workload studies, little research has been conducted on the workload experienced by teams. Team workload has largely been ignored by research with no validated theory constructed or dedicated team workload measures available to researchers and practitioners. The research was conducted in two studies with study 1 focusing on the development of the TWLQ with 216 members of sports team completing a team workload measure after games or practise. In study 2, 14 dyadic teams performed two sessions of an unmanned aerial vehicle (UAV) search-and-rescue task. The TWLQ was used to measure the team workload demand of the task. Study 1: Principle Axis Factoring method with Direct Oblimin rotation indicated three separate factors for the TWLQ with the factors classified as Task Workload, Team Workload, and Task-Team Balancing. Study 2: The TWLQ exhibited differential sensitivity, with the three factors measuring unique components of the workload demands in teams. The TWLQ is a valid and reliable subjective measure that can be used to assess the workload demand in team tasks. It provides researchers a tool to advance the understating of team workload and gives practitioners the means to assess the workload demands of team tasks in applied settings.

Advanced development in processor design, increasing the heterogeneity of computer system, by involving Graphics Processing Units (GPUs), Field-Programmable Gate Arrays (FPGAs) and custom accelerators, and increasing the number of cores and threads in such systems puts extra pressure on the main memory, demanding a higher performance. Current computing trends are putting ever more pressure on main memory. In modern computer systems, this is generally Dynamic Random Access Memory (DRAM) which consists of a multi-level access hierarchy (e.g. Rank, Bank, Row etc.). This heterogeneity of structure implies different access latencies (and power consumption), resulting in performance differences according to memory access patterns. DRAM controllers manage access and satisfy the timing constraints and now employ complex scheduling and prediction algorithms to mitigate the effect on performance. This complexity can limit the scalability of a controller with the size of memory, while maintaining performance. The focus of this PhD thesis is to improve performance, reliability and scalability (with respect to memory size) of DRAM controllers. To this end, it covers three significant contributors to the performance and reliability of a memory controller: ‘Address Mapping’, ‘Page Closure Policies’ and ‘Reliability Monitoring’. A detailed DRAM simulator is used as an evaluation platform throughout this work. The following contributions are presented in this thesis. Hybrid Address-based Page PolicY (HAPPY): Memory controllers have used static page-closure policies to decide whether a row should be left open (open-page policy) or closed immediately (close-page policy) after use. The appropriate choice can reduce the average memory latency. Since access patterns are dynamic, static page policies cannot guarantee to deliver optimum execution time. Hybrid page policies can cover dynamic scenarios and are now implemented in state-of-the-art processors. These switch between open-page and close-page policies by monitoring the access pattern of row hits/conflicts and predicting future behaviour. Unfortunately, as the size of DRAM memory increases, fine-grain tracking and analysis of accesses does not remain practical. HAPPY proposes a compact, memory address-based encoding technique which can maintain or improve page closure predictor performance while reducing the hardware overhead. As a case study, HAPPY is integrated, with a state-of-the-art monitor – the Intel-adaptive open-page policy predictor employed by the Intel Xeon X5650 – and a traditional Hybrid page policy. The experimental results show that using the HAPPY encoding applied to the Intel-adaptive page closure policy can reduce the hardware overhead by 5× for the evaluated 64 GB memory (up to 40× for a 512 GB memory) while maintaining the prediction accuracy. Dynamic Re-arrangement of Address Mapping (DReAM): The initial location of data in DRAMs is determined and controlled by the ‘address-mapping’ and even modern memory controllers use a fixed and runtime-agnostic address-mapping. On the other hand, the memory access pattern seen at the memory interface level will be dynamically changed at run-time. This dynamic nature of memory access pattern and the fixed behaviour of address mapping process in DRAM controllers, implied by using a fixed address-mapping scheme, means that DRAM performance cannot be exploited efficiently. DReAM is a novel hardware technique that can detect a workload-specific address mapping at run-time based on the application access pattern. The experimental results show that DReAM outperforms the best evaluated baseline address mapping by 5%, on average, and up to 28% across all the workloads. A Run-time Memory hot-row detectOR (ARMOR): DRAM needs refreshing to avoid data loss. Data can also be corrupted within a refresh interval by crosstalk caused by repeated accesses to neighbouring rows; this is the row hammer effect and is perceived as a potentially serious reliability and security threat. ARMOR is a novel technique which improves memory reliability by detecting which rows are potentially being “hammered” within the memory controller, which can then insert extra refresh operations. It can detect (and thus prevent) row hammer errors with minimal execution time overhead and hardware requirements. Alternatively by adding buffers inside the memory controller to cache such hammered rows, execution times are reduced with small hardware costs. The ARMOR technique is now the basis of a patent applicant and under process for commercial exploitation. As a final step of this PhD thesis, an adaptive memory controller was developed integrating HAPPY, DReAM and ARMOR into a standard memory controller. The performance and the implementation cost of such an adaptive memory controller were compared against a state-of-the-art memory controller, as a baseline. The experimental results show that the adaptive memory controller outperforms the baseline, on average by 18%, and up to 35% for some workloads, while requiring around 6 KB-900 KB more storage than the baseline to support a wide range of memory sizes (from 4 GB up to 512 GB).

The growing disparity in data storage and retrieval needs of modern applications is driving the proliferation of a wide variety of storage systems (e.g., key-value stores, cloud storage services, distributed filesystems, and flash cache, etc.). While extant storage systems are designed and tuned for a specific set of applications targeting a range of workload characteristics, they lack the flexibility in adapting to the ever-changing workload behaviors. Moreover, the complexities in implementing modern storage systems and adapting ever-changing storage requirements present unique opportunities and engineering challenges. In this dissertation, we design and develop a series of novel data management and storage systems solutions by applying a simple yet effective rule---workload awareness. We find that simple workload-aware data management strategies are effective in improving the efficiency of modern storage systems, sometimes by an order of magnitude. The first two works tackle the data management and storage space allocation issues at distributed and cloud storage level, while the third work focuses on low-level data management problems in the local storage system, which many high-level storage/data-intensive applications rely on. In the first part of this dissertation (Chapter~ref{ch:mbal}), we propose and develop MBal, a high-performance in-memory object caching framework with adaptive multi-phase load balancing, which supports not only horizontal (scale-out) but vertical (scale-up) scalability as well. MBal is able to make efficient use of available resources in the cloud through its fine-grained, partitioned, lockless design. In the second part of this dissertation (Chapter~ref{ch:cast} and Chapter~ref{ch:pricing}), we design and build CAST (Chapter~ref{ch:cast}), a Cloud Analytics Storage Tiering solution that cloud tenants can use to reduce monetary cost and improve performance of analytics workloads. The approach takes the first step towards providing storage tiering support for data analytics in the cloud. Furthermore, we propose a hybrid cloud object storage system (Chapter~ref{ch:pricing}) that could effectively engage both the cloud service providers and cloud tenants via a novel dynamic pricing mechanism. In the third part of this dissertation (Chapter~ref{ch:offline}), targeting local storage, we explore offline algorithms for flash caching in terms of both hit ratio and flash lifespan. We design and implement a multi-stage heuristic by synthesizing several techniques that manage data at the granularity of a flash erasure unit (which we call a container) to approximate the offline optimal algorithm. In the fourth part of this dissertation (Chapter~ref{ch:turnkey}), we are focused on how to enable fast prototyping of efficient distributed key-value stores targeting a proxy-based layered architecture. In this work, we design and build {con}, a framework that significantly reduce the engineering effort required to build a full-fledged distributed key-value store. Our dissertation shows that simple workload-aware data management strategies can bring huge benefit in terms of both efficiency (i.e., performance, monetary cost, etc.) and flexibility (i.e., ease-of-use, ease-of-deployment, programmability, etc.). The principles of leveraging workload dynamicity and storage heterogeneity can be used to guide next-generation storage system software design, especially when being faced with new storage hardware technologies.<br>Ph. D.