The effect of bilingual representation in prose reading on explicit and implicit memory

Chinese-English bilingual students were randomly assigned to three reading conditions: In the English-English (E-E) condition (n = 44), a text in English was read twice; in the English-Chinese (E-C) condition (n = 30), the English text was read first and its Chinese translation was read second; in the Chinese-English (C-E) condition (n = 30), the Chinese text was read first and English second. An expected explicit memory test on propositions in the format of sentence verification was given followed by an unexpected implicit memory test on unfamiliar word-forms.^ Analyses of covariance were conducted with explicit and implicit memory scores as the dependent variables, reading condition (bilingual versus monolingual) as the independent variable, and TOEFL reading score as the covariate.^ The results showed that the bilingual reading groups outperformed the monolingual reading group on explicit memory tested by sentence-verification but not on implicit memory tested by forced-choice word-identification, implying that bilingual representation facilitates explicit memory of propositional information but not implicit memory of lexical forms. The findings were interpreted as consistent with separate bilingual memory-storage models and the implications of such models in the study of cognitive structures were discussed in relationship to issues of dual coding theory, multiple memory systems, and the linguistic relativity philosophy. ^

Efficient asynchronous scheduling algorithms for cost-effective buffered crossbar switches

Buffered crossbar switches have recently attracted considerable attention as the next generation of high speed interconnects. They are a special type of crossbar switches with an exclusive buffer at each crosspoint of the crossbar. They demonstrate unique advantages over traditional unbuffered crossbar switches, such as high throughput, low latency, and asynchronous packet scheduling. However, since crosspoint buffers are expensive on-chip memories, it is desired that each crosspoint has only a small buffer. This dissertation proposes a series of practical algorithms and techniques for efficient packet scheduling for buffered crossbar switches. To reduce the hardware cost of such switches and make them scalable, we considered partially buffered crossbars, whose crosspoint buffers can be of an arbitrarily small size. Firstly, we introduced a hybrid scheme called Packet-mode Asynchronous Scheduling Algorithm (PASA) to schedule best effort traffic. PASA combines the features of both distributed and centralized scheduling algorithms and can directly handle variable length packets without Segmentation And Reassembly (SAR). We showed by theoretical analysis that it achieves 100% throughput for any admissible traffic in a crossbar with a speedup of two. Moreover, outputs in PASA have a large probability to avoid the more time-consuming centralized scheduling process, and thus make fast scheduling decisions. Secondly, we proposed the Fair Asynchronous Segment Scheduling (FASS) algorithm to handle guaranteed performance traffic with explicit flow rates. FASS reduces the crosspoint buffer size by dividing packets into shorter segments before transmission. It also provides tight constant performance guarantees by emulating the ideal Generalized Processor Sharing (GPS) model. Furthermore, FASS requires no speedup for the crossbar, lowering the hardware cost and improving the switch capacity. Thirdly, we presented a bandwidth allocation scheme called Queue Length Proportional (QLP) to apply FASS to best effort traffic. QLP dynamically obtains a feasible bandwidth allocation matrix based on the queue length information, and thus assists the crossbar switch to be more work-conserving. The feasibility and stability of QLP were proved, no matter whether the traffic distribution is uniform or non-uniform. Hence, based on bandwidth allocation of QLP, FASS can also achieve 100% throughput for best effort traffic in a crossbar without speedup.

Host and network optimizations for performance enhancement and energy efficiency in data center networks

Modern data centers host hundreds of thousands of servers to achieve economies of scale. Such a huge number of servers create challenges for the data center network (DCN) to provide proportionally large bandwidth. In addition, the deployment of virtual machines (VMs) in data centers raises the requirements for efficient resource allocation and find-grained resource sharing. Further, the large number of servers and switches in the data center consume significant amounts of energy. Even though servers become more energy efficient with various energy saving techniques, DCN still accounts for 20% to 50% of the energy consumed by the entire data center. The objective of this dissertation is to enhance DCN performance as well as its energy efficiency by conducting optimizations on both host and network sides. First, as the DCN demands huge bisection bandwidth to interconnect all the servers, we propose a parallel packet switch (PPS) architecture that directly processes variable length packets without segmentation-and-reassembly (SAR). The proposed PPS achieves large bandwidth by combining switching capacities of multiple fabrics, and it further improves the switch throughput by avoiding padding bits in SAR. Second, since certain resource demands of the VM are bursty and demonstrate stochastic nature, to satisfy both deterministic and stochastic demands in VM placement, we propose the Max-Min Multidimensional Stochastic Bin Packing (M3SBP) algorithm. M3SBP calculates an equivalent deterministic value for the stochastic demands, and maximizes the minimum resource utilization ratio of each server. Third, to provide necessary traffic isolation for VMs that share the same physical network adapter, we propose the Flow-level Bandwidth Provisioning (FBP) algorithm. By reducing the flow scheduling problem to multiple stages of packet queuing problems, FBP guarantees the provisioned bandwidth and delay performance for each flow. Finally, while DCNs are typically provisioned with full bisection bandwidth, DCN traffic demonstrates fluctuating patterns, we propose a joint host-network optimization scheme to enhance the energy efficiency of DCNs during off-peak traffic hours. The proposed scheme utilizes a unified representation method that converts the VM placement problem to a routing problem and employs depth-first and best-fit search to find efficient paths for flows.

Algorithms for efficient transaction management and consistent queries in client -server semantic object-oriented parallel databases

Large read-only or read-write transactions with a large read set and a small write set constitute an important class of transactions used in such applications as data mining, data warehousing, statistical applications, and report generators. Such transactions are best supported with optimistic concurrency, because locking of large amounts of data for extended periods of time is not an acceptable solution. The abort rate in regular optimistic concurrency algorithms increases exponentially with the size of the transaction. The algorithm proposed in this dissertation solves this problem by using a new transaction scheduling technique that allows a large transaction to commit safely with significantly greater probability that can exceed several orders of magnitude versus regular optimistic concurrency algorithms. A performance simulation study and a formal proof of serializability and external consistency of the proposed algorithm are also presented.^ This dissertation also proposes a new query optimization technique (lazy queries). Lazy Queries is an adaptive query execution scheme which optimizes itself as the query runs. Lazy queries can be used to find an intersection of sub-queries in a very efficient way, which does not require full execution of large sub-queries nor does it require any statistical knowledge about the data.^ An efficient optimistic concurrency control algorithm used in a massively parallel B-tree with variable-length keys is introduced. B-trees with variable-length keys can be effectively used in a variety of database types. In particular, we show how such a B-tree was used in our implementation of a semantic object-oriented DBMS. The concurrency control algorithm uses semantically safe optimistic virtual "locks" that achieve very fine granularity in conflict detection. This algorithm ensures serializability and external consistency by using logical clocks and backward validation of transactional queries. A formal proof of correctness of the proposed algorithm is also presented. ^

Host and Network Optimizations for Performance Enhancement and Energy Efficiency in Data Center Networks

Modern data centers host hundreds of thousands of servers to achieve economies of scale. Such a huge number of servers create challenges for the data center network (DCN) to provide proportionally large bandwidth. In addition, the deployment of virtual machines (VMs) in data centers raises the requirements for efficient resource allocation and find-grained resource sharing. Further, the large number of servers and switches in the data center consume significant amounts of energy. Even though servers become more energy efficient with various energy saving techniques, DCN still accounts for 20% to 50% of the energy consumed by the entire data center. The objective of this dissertation is to enhance DCN performance as well as its energy efficiency by conducting optimizations on both host and network sides. First, as the DCN demands huge bisection bandwidth to interconnect all the servers, we propose a parallel packet switch (PPS) architecture that directly processes variable length packets without segmentation-and-reassembly (SAR). The proposed PPS achieves large bandwidth by combining switching capacities of multiple fabrics, and it further improves the switch throughput by avoiding padding bits in SAR. Second, since certain resource demands of the VM are bursty and demonstrate stochastic nature, to satisfy both deterministic and stochastic demands in VM placement, we propose the Max-Min Multidimensional Stochastic Bin Packing (M3SBP) algorithm. M3SBP calculates an equivalent deterministic value for the stochastic demands, and maximizes the minimum resource utilization ratio of each server. Third, to provide necessary traffic isolation for VMs that share the same physical network adapter, we propose the Flow-level Bandwidth Provisioning (FBP) algorithm. By reducing the flow scheduling problem to multiple stages of packet queuing problems, FBP guarantees the provisioned bandwidth and delay performance for each flow. Finally, while DCNs are typically provisioned with full bisection bandwidth, DCN traffic demonstrates fluctuating patterns, we propose a joint host-network optimization scheme to enhance the energy efficiency of DCNs during off-peak traffic hours. The proposed scheme utilizes a unified representation method that converts the VM placement problem to a routing problem and employs depth-first and best-fit search to find efficient paths for flows.

Improving caches in consolidated environments

Memory (cache, DRAM, and disk) is in charge of providing data and instructions to a computer's processor. In order to maximize performance, the speeds of the memory and the processor should be equal. However, using memory that always match the speed of the processor is prohibitively expensive. Computer hardware designers have managed to drastically lower the cost of the system with the use of memory caches by sacrificing some performance. A cache is a small piece of fast memory that stores popular data so it can be accessed faster. Modern computers have evolved into a hierarchy of caches, where a memory level is the cache for a larger and slower memory level immediately below it. Thus, by using caches, manufacturers are able to store terabytes of data at the cost of cheapest memory while achieving speeds close to the speed of the fastest one.^ The most important decision about managing a cache is what data to store in it. Failing to make good decisions can lead to performance overheads and over-provisioning. Surprisingly, caches choose data to store based on policies that have not changed in principle for decades. However, computing paradigms have changed radically leading to two noticeably different trends. First, caches are now consolidated across hundreds to even thousands of processes. And second, caching is being employed at new levels of the storage hierarchy due to the availability of high-performance flash-based persistent media. This brings four problems. First, as the workloads sharing a cache increase, it is more likely that they contain duplicated data. Second, consolidation creates contention for caches, and if not managed carefully, it translates to wasted space and sub-optimal performance. Third, as contented caches are shared by more workloads, administrators need to carefully estimate specific per-workload requirements across the entire memory hierarchy in order to meet per-workload performance goals. And finally, current cache write policies are unable to simultaneously provide performance and consistency guarantees for the new levels of the storage hierarchy.^ We addressed these problems by modeling their impact and by proposing solutions for each of them. First, we measured and modeled the amount of duplication at the buffer cache level and contention in real production systems. Second, we created a unified model of workload cache usage under contention to be used by administrators for provisioning, or by process schedulers to decide what processes to run together. Third, we proposed methods for removing cache duplication and to eliminate wasted space because of contention for space. And finally, we proposed a technique to improve the consistency guarantees of write-back caches while preserving their performance benefits.^

Curricular perceptions of adult education at National Chun -Chin University, Chia -Yi, Taiwan

This research was conducted in Chia-Yi, Taiwan to study the needs of adult education participants to determine the factors necessary to provide direction for the development of university adult education curriculum and supportive government educational policies. Factors researched were characteristics of the adult learner, theories of adult learning, demands of adult education, and implications of university adult education as the theoretical foundation for the development of specific curriculum development efforts. The study investigated adult learning needs and their relationship with demographic variable. Analyzing the needs of adult education participant and the relative factors through a survey resulted in recommendations for the development of adult education program plans, content of curriculum, and teaching. Research questions were analyzed using descriptive statistics, frequencies, chi square, one-way analysis of variance (ANOVA), and post hoc analysis. ^ The study showed that most participants in these adult education activities were under forty, middle class, of above average educational levels, and either living or working in the city. People who were older, of lesser social and economic positions, with lower educational standards, and living/working in the country, did not participate as much in adult education opportunities. Recommendations included that in the planning or setting up of adult education activities, attention be given to all the possible barriers or problems that are likely to occur in people's participation, e.g., motivation, interests, content of courses, teaching methods, willingness of participation, qualification of teachers, time, funds, locations, and so on. It is suggested that the resolution of these problems can significantly increase the participation of adult education. ^

Size-based variation in intertissue comparisons of stable carbon and nitrogen isotopic signatures of bull sharks (Carcharhinus leucas) and tiger sharks (Galeocerdo cuvier)

Stable isotopes are important tools for understanding the trophic roles of elasmobranchs. However, whether different tissues provide consistent stable isotope values within an individual are largely unknown. To address this, the relationships among carbon and nitrogen isotope values were quantified for blood, muscle, and fin from juvenile bull sharks (Carcharhinus leucas) and blood and fin from large tiger sharks (Galeocerdo cuvier) collected in two different ecosystems. We also investigated the relationship between shark size and the magnitude of differences in isotopic values between tissues. Isotope values were significantly positively correlated for all paired tissue comparisons, but R2 values were much higher for δ13C than for δ15N. Paired differences between isotopic values of tissues were relatively small but varied significantly with shark total length, suggesting that shark size can be an important factor influencing the magnitude of differences in isotope values of different tissues. For studies of juvenile sharks, care should be taken in using slow turnover tissues like muscle and fin, because they may retain a maternal signature for an extended time. Although correlations were relatively strong, results suggest that correction factors should be generated for the desired study species and may only allow coarse-scale comparisons between studies using different tissue types.

Large-Scale Testing to Study the Effects of Critical Parameters on the Aerodynamic Behavior of Long Span Bridges

Long-span bridges are flexible and therefore are sensitive to wind induced effects. One way to improve the stability of long span bridges against flutter is to use cross-sections that involve twin side-by-side decks. However, this can amplify responses due to vortex induced oscillations. Wind tunnel testing is a well-established practice to evaluate the stability of bridges against wind loads. In order to study the response of the prototype in laboratory, dynamic similarity requirements should be satisfied. One of the parameters that is normally violated in wind tunnel testing is Reynolds number. In this dissertation, the effects of Reynolds number on the aerodynamics of a double deck bridge were evaluated by measuring fluctuating forces on a motionless sectional model of a bridge at different wind speeds representing different Reynolds regimes. Also, the efficacy of vortex mitigation devices was evaluated at different Reynolds number regimes. One other parameter that is frequently ignored in wind tunnel studies is the correct simulation of turbulence characteristics. Due to the difficulties in simulating flow with large turbulence length scale on a sectional model, wind tunnel tests are often performed in smooth flow as a conservative approach. The validity of simplifying assumptions in calculation of buffeting loads, as the direct impact of turbulence, needs to be verified for twin deck bridges. The effects of turbulence characteristics were investigated by testing sectional models of a twin deck bridge under two different turbulent flow conditions. Not only the flow properties play an important role on the aerodynamic response of the bridge, but also the geometry of the cross section shape is expected to have significant effects. In this dissertation, the effects of deck details, such as width of the gap between the twin decks, and traffic barriers on the aerodynamic characteristics of a twin deck bridge were investigated, particularly on the vortex shedding forces with the aim of clarifying how these shape details can alter the wind induced responses. Finally, a summary of the issues that are involved in designing a dynamic test rig for high Reynolds number tests is given, using the studied cross section as an example.