the design of a distributed real-time database based on cloud theory is proposed. A distributed indexing mechanism is researched to achieve more efficient management of industrial data. A real-time transaction schedul...
详细信息
ISBN:
(纸本)9781467347082;9781467347075
the design of a distributed real-time database based on cloud theory is proposed. A distributed indexing mechanism is researched to achieve more efficient management of industrial data. A real-time transaction scheduling algorithm in distributed environment is designed to improve transaction processing efficiency. A network dynamic routing and business related load balancing mechanism is researched to deal with modification of system scale and make full use of system resources. these technologies are utilized to implement distributed real-time database for large-scale process industry. Experimental results show that the cloud-based database has better performance and significantly improves data throughput, fault tolerance and scalability.
parallel programming patterns provide enduring principles that serve as a conceptual framework to orient students when they set out to solve problems. Learning patterns enables students to quickly gain the intellectua...
详细信息
the training of SVM can be viewed as a Convex Quadratic Programming (CQP) problem which becomes difficult to be solved when dealing withthe large scale data sets. Traditional methods such as Sequential Minimal Optimi...
详细信息
ISBN:
(纸本)9781479924189
the training of SVM can be viewed as a Convex Quadratic Programming (CQP) problem which becomes difficult to be solved when dealing withthe large scale data sets. Traditional methods such as Sequential Minimal Optimization (SMO) for SVM training is used to solve a sequence of small scale sub-problems, which costs a large amount of computation time and is hard to be accelerated by utilizing the computation power of GPU. Although Interior Point Method (IPM) such as primal-dual interior point method (PDIPM) can be also addressed SVM training well and has favourable potential for parallelizing on GPU, it contains comparatively high time complexity O(l(3)) and space complexity O(l(2)), where l is the number of training instances. Fortunately, by invoking low-rank approximation methods such as Incomplete Cholesky Factorization (ICF) and Sherman Morrison Woodbury formula (SMW), the requirements of both storage and computation of PDIPM can be reduced significantly. In this paper, a parallel PDIPM method (P-PDIPM) along with a parallel ICF method (P-ICF) is proposed to accelerate the SVM training on GPU. Experimental results indicate that the training speed of P-PDIPM on GPU is almost 40x faster than that of the serial one (S-PDIPM) on CPU. Besides, without extensive optimization, P-PDIPM can obtain about 8x speedup over the state of the art tool LIBSVM while maintaining high prediction accuracy.
Wireless technologies employing small sensors are particularly useful since they allow monitoring of kinetic and physiological data without affecting any individual in motion. the primary objective of this collaborati...
详细信息
ISBN:
(纸本)9780769551043
Wireless technologies employing small sensors are particularly useful since they allow monitoring of kinetic and physiological data without affecting any individual in motion. the primary objective of this collaborative research between the Department of Electrical Engineering and computing Sciences and the Department of Athletes at the University of Cincinnati is to build a wireless system using a wireless body area sensor network (WBASN) for the overhead squat in assessing one element of human movement and its relationship to the kinematic lower extremity movement and injuries. this research can provide an objective analysis of the pilot data for movement of the human body during regular athletic exercises.
Modern day systems are facing an avalanche of data, and they are being forced to handle more and more data intensive use cases. these data comes in many forms and shapes: Sensors (RFID, Near Field Communication, Weath...
详细信息
ISBN:
(纸本)9783642408199;9783642408205
Modern day systems are facing an avalanche of data, and they are being forced to handle more and more data intensive use cases. these data comes in many forms and shapes: Sensors (RFID, Near Field Communication, Weather Sensors), transaction logs, Web, social networks etc. As an example, weather sensors across the world generate a large amount of data throughout the year. Handling these and similar data require scalable, efficient, reliable and very large storages with support for efficient metadata based searching. this paper present Mahasen, a highly scalable storage for high volume data intensive applications built on top of a peer-to-peer layer. In addition to scalable storage, Mahasen also supports efficient searching, built on top of the distributed Hash table (DHT)
We focus on the parallelization of two-dimensional square packing problem. In square packing problem, a list of square items need to be packed into a minimum number of unit square bins. All square items have side leng...
详细信息
ISBN:
(纸本)9781479924189
We focus on the parallelization of two-dimensional square packing problem. In square packing problem, a list of square items need to be packed into a minimum number of unit square bins. All square items have side length smaller than or equal to 1 which is also the side length of each unit square bin. the total area of items that has been packed into one bin cannot exceed 1. Using the idea of harmonic, some squares can be put into the same bin without exceeding the bin limitation of side length 1. We try to concurrently pack all the corresponding squares into one bin by a parallel systerm of computation processing. A 9/4-worst case asymptotic error bound algorithm with time complexity theta(n) is showed. Let OPT(I) and A(I) denote, respectively, the cost of an optimal solution and the cost produced by an approximation algorithm A for an instance I of the square packing problem. the best upper bound of on-line square packing to date is 2.1439 proved by Han et al. [23] by using complexity weighting functions. However the upper bound of our parallel algorithm is a litter worse than Han's algorithm, the analysis of our algorithm is more simple and the time complexity is improved. Han's algorithm needs O(nlogn) time, while our method only needs theta(n) time.
It has become increasingly common to see that supercomputingapplications harness the massive parallelism of graphics cards to speed up computations. In this study, an analysis concerning to the time necessity for fou...
详细信息
ISBN:
(纸本)9781467364195
It has become increasingly common to see that supercomputingapplications harness the massive parallelism of graphics cards to speed up computations. In this study, an analysis concerning to the time necessity for four different implementations of parallel matrix multiplication is presented. the execution time of parallel matrix multiplications in Compute Unified Device Architecture (CUDA) can be increased to about 10 times than Matlab implementation, 100 times than Java thread, 300 times than C++ by using duo core Central Processing Unit (CPU) and 600 times than C++ by using single core CPU respectively by our method, as compared with using the fastest tools of GPU-only case or CPU-only case. the goal of this study is to show how to offload parallel computations to the graphics card, when it is necessary, and to give some idea of how to think about code running in the massively parallel environment.
Todays context frameworks provide solutions for context mechanisms for individual applications only: context aware working spaces, easier mobile development frameworks or higher-level context abstractions. RestContext...
详细信息
ISBN:
(纸本)9789897581281
Todays context frameworks provide solutions for context mechanisms for individual applications only: context aware working spaces, easier mobile development frameworks or higher-level context abstractions. RestContext solves this problem with a service logically separating context as a set of information that can characterize a situation from further context interpretation mechanisms. RestContext is a resource oriented architecture which manages sensors of different types. A context may consist of sub-contexts as well as sensors that are linked to one or many contexts. Withthe help of RestContext it is possible to create topologies of contexts. Sensors can interact with context instances by push and pull mechanisms. We demonstrate how RestContext can be used in a distributed weather forecasting example.
Finding basic laws that govern human crowd behavior is a subject that deserves to study. Crowd behavior is a natural instinct for human, which directly impacts how we form opinions and make decisions. Moreover, it is ...
详细信息
ISBN:
(纸本)9780769551043
Finding basic laws that govern human crowd behavior is a subject that deserves to study. Crowd behavior is a natural instinct for human, which directly impacts how we form opinions and make decisions. Moreover, it is common that people change their behavior in a group. In pervasive computing research, substantial work has been directed towards discovering human movement patterns based on wireless networks. this research has, however, mainly been focused on movements of individuals. Mobile phones offer on-body tracking and they are already deployed on a large scale, allowing the characterization of user behavior through the information related to individual movements. In this paper, we observe and analyze the impact of friendship and location attributes on crowd behavior, using location-based wireless mobility information. these preliminary studies will be a good cornerstone for a crowd behavior prediction.
Wireless sensor networks collecting data to monitor real life processes gain increasing attention from the scientific community. these systems promise ubiquitous computing in a digitalized world. However many problems...
详细信息
ISBN:
(纸本)9789897581304
Wireless sensor networks collecting data to monitor real life processes gain increasing attention from the scientific community. these systems promise ubiquitous computing in a digitalized world. However many problems need to be solved to achieve this dream. One of these problems is the limited battery power of the nodes and the limited bandwidth of the communication. Existing work tackle the problem of limited bandwidth by merging communication packets within the network. Based on this approach we investigate the use of application specific aggregation in publish/subscribe wireless sensor systems. We expect this approach to overcome overload situations in the network and decrease packet loss due to bandwidth exhaustion. this paper describes our architecture and evaluates the properties of our approach. We investigated specific topologies theoretically and through real- life experiments. Our results quantify the achievable event count and loss rate reduction.
暂无评论