Welcome to Runner Research Press, a showcase of the University's research and scholarly works.

 

Communities in Runner Research Press

Select a community to browse its collections.

Recent Submissions

Item
Management of shared resources in multi-threading / multi-core systems
(2014) Zhang, Yilin
Based on the traditional superscalar processors, Simultaneous Multi-Threading (SMT) offers an improved mechanism to enhance the overall performance by exploiting Thread-Level Parallelism (TLP) to overcome the limits of Instruction-Level Parallelism (ILP), and a multi-core system with multiple independent processors is capable in utilizing job-level parallelism by allowing multiple jobs to be processed currently. The most common characteristic of parallel systems is the sharing of key datapath components among multiple independent threads/jobs in order to better utilize the resources. In an SMT system, due to the various characteristics of each thread, the occupation of the shared resources can be very unbalanced. Our research is aiming to solve this problem and to make efficient resource allocation among threads. Our investigation shows that among the resources in an SMT system, physical register file, Issue Queue (IQ) and write buffer are the most critical resources that are shared among threads. There are several approaches proposed in this dissertation: Register File Allocation, Instruction Recalling, Speculative Trace Control, Autonomous IQ Usage Control, Write Buffer Capping and Integrated Shared Resources Control. To better utilize the physical register file, we limit the maximal number of physical registers that a thread is allowed to occupy at any time, so as to eliminate the overwhelming occupation caused by a single thread. Several techniques have been proposed in order to improve the utilization of IQ: (1) to reduce the IQ occupation of the inactive thread, we introduce Instruction Recalling to remove those long-latency instructions; (2) to reduce the wastes of resources caused by the wrong-way trace due to a branch miss prediction, we propose an algorithm to control the amount of speculative instructions from a thread to be dispatched and executed in the pipeline, the so-called Speculative Trace Control technique; and (3) to remove the environment dependency of a technique, we introduced Autonomous Control to adjust the IQ distribution based on the real-time performance output. The write buffer is another shared resource which is easily unfairly occupied. Write Buffer Capping is a technique which prevents any threads from overwhelmingly occupying the write buffer by setting a cap value on the maximal amount of write buffer entries that a thread is allowed to take. The Integrated Shared Resource Management takes the above factors into consideration and manages the usage of the most critical shared resources (physical register file, IQ and write buffer) simultaneously for each thread, providing even significant enhancement with relative small hardware investment. In a multi-core system, memory and the interconnection network are shared among processors and their performances are key to the overall throughput of the system. In the last chapter we further extend our analysis on the impact that different interconnection networks impose on the whole system's overall performance. We show that the tradeoff between latency and concurrent access capacity may become a critical deciding factor in choosing the correct size of network for applications with different memory traffic demands.
Item
Essays on state pension plans and trading in bankrupt stocks
(2014) Zhang, Hongxian
This dissertation consists of two essays on suboptimal behavior of financial markets. Essay I examines stocks of bankrupt firms after the court confirms they will receive nothing. While trading volume is negligible for most worthless stocks, some have sizable trading volume, indicating investor ignorance of their zero intrinsic value. Prices respond irrationally to news in several instances, and they are higher for more liquid worthless stocks, which are more likely to attract uninformed investors. Our analysis includes the first empirical examination of short-selling in bankrupt firms. Short-covering cannot account for the anomalous price and trading volume. Short-sellers are active in these stocks and play a useful role in pushing prices down toward intrinsic value. Essay II examines the effects of state corruption as well as political and governance factors on U.S. public pension funds. We find that pension funds in states with more corruption have lower performance; a one standard deviation increase in corruption is associated with a decrease in annual returns between 17 and 21 basis points, and this relation is robust to state-level and pension-level fixed effects. Pensions located in more corrupt jurisdictions also invest a larger fraction of their assets in equities. We find that having a new treasurer decreases the negative effects of corruption, suggesting that frequent changes in administrations are beneficial in corrupt jurisdictions. Governance-related variables and political affiliation variables are by themselves not significantly related to pension returns, although these variables are associated with differences in asset allocation.
Item
Bayesian power prior analysis and its application to operational risk and Rasch model
(2010) Zhang, Honglian
When sample size is small, informative priors can be valuable in increasing the precision of estimates. Pooling historical data and current data with equal weights under the assumption that both of them are from the same population may be misleading when heterogeneity exists between historical data and current data. This is particularly true when the sample size of historical data is much larger than that of the current data. One way of constructing an informative prior in the presence of the historical data is the power prior, which is realized by raising the likelihood of the historical data to a fractional power. In this dissertation, we extend the power prior by considering the existence of nuisance parameters. When historical information is used as priors, we assume that the parameters of interest have not changed, while the nuisance parameter may change. The properties of power prior methods with nuisance parameters and its posterior distributions are examined for normal populations. The power prior approaches, with or without nuisance parameters, are compared empirically in terms of the mean squared error (MSE) of the estimated parameter of interest as well as the behavior of the power parameter. To illustrate the implementation of the power prior with nuisance parameter approach, we apply it to lognormal models for operational risk data and the Rasch model for item response theory (IRT). In the application to the Rasch model, we extend the power prior with nuisance parameter approach further by incorporating it with the hierarchical Bayes model.
Item
A sublexical unit based hash model approach for spam detection
(2009) Zhang, Like
This research introduces an original anomaly detection approach based on a sublexical unit hash model for application level content. This approach is an advance over previous arbitrarily defined payload keyword and 1-gram frequency analysis approaches. Based on the split fovea theory in human recognition, this new approach uses a special hash function to identify groups of neighboring words. The hash frequency distribution is calculated to build the profile for a specific content type. Examples of utilizing the algorithm for detecting spam and phishing emails are illustrated in this dissertation. A brief review of network intrusion and anomaly detection will first be presented, followed by a discussion of recent research initiatives on application level anomaly detection. Previous research results for payload keyword and byte frequency based anomaly detection will also be presented. The drawback in using N-gram analysis, which has been applied in most related research efforts, is discussed at the end of chapter 2. The importance of text content analysis to application level anomaly detection will also be explained. After a background introduction of the split fovea theory in psychological research, the proposed sublexical unit hash frequency distribution based method will be presented. How human recognition theory is applied as the fundamental element for a proposed hashing algorithm will be examined followed by a demonstration of how the hashing algorithm is applied to anomaly detection. Spam email is used as the major example in this discussion. The reason spam and phishing emails are used in our experiments includes the availability of detailed experimental data and the possibility of conducting an in-depth analysis of the test data. An interesting comparison between the proposed algorithm and several popular commercial spam email filters used by Google and Yahoo is also presented. The outcome shows the benefits of the proposed approach. The last chapter provides a review of the research and explains how the previous payload keyword approach evolved into the hash model solution. The last chapter discusses the possibility of extending the hash model based anomaly detection to other areas including Unicode applications.
Item
The influence of electronic word-of-mouth (eWOM) skepticism on perceptions toward message credibility and beneficiary organization
(2016) Zhang, Xiao Jerry
Electronic word-of-mouth (eWOM) has been cited as a significant factor influencing Internet users' perceptions in various situations, sectors and industries (Lee et al. 2009; Chatterjee 200; Dellarocas et al. 2007; Cox et al. 2008). However, as more evidence demonstrating the pervasive use of fake eWOM has been exposed (Forrest and Cao 2010; Malbon 2013), Internet users' confidence regarding the truthfulness and genuineness of eWOM may have been severely undermined. Different from most existing online trust and online information credibility literature, this research assumes that Internet users may have already developed a certain level of skepticism toward all eWOM messages (eWOM skepticism). In this study, our research focuses on how eWOM skepticism is influenced by personal and environmental factors, and how eWOM skepticism influences Internet users' message credibility assessment and their attitudes toward the organization that may get benefits from fake eWOM propaganda. To achieve this goal, first, we created the new measurement items for eWOM skepticism and validated them. Then, using control experiment, we collected the data, which were analyzed using Multivariate Analysis of Covariance (MANCOVA) and Partial Least Squares (PLS). The results revealed that dispositional trust, structural assurance and negative experience of eWOM significantly influence eWOM skepticism, and that eWOM skepticism is likely to influence Internet users' judgments and perceptions. Based on our model, we also found that the attributions Internet users made toward the eWOM messages are strong predictors for their attitude toward the eWOM messages and the potential beneficiary organization. This study emphasizes the importance of incorporating eWOM skepticism when investigating eWOM trust scenarios, and supports the argument about the coexistence of trust and distrust. Several theoretical and practical contributions are also discussed.