This paper covers the fast solution of large acoustic problems on low-resources parallel platforms.A domain decomposition method is coupled with a dynamic load balancing scheme to efficiently accelerate a geometrical ...
详细信息
This paper covers the fast solution of large acoustic problems on low-resources parallel platforms.A domain decomposition method is coupled with a dynamic load balancing scheme to efficiently accelerate a geometrical acoustic *** geometrical method studied implements a beam-tracing method where intersections are handled as in a ray-tracing *** the distribution of the global processing upon multiple sub-domains,a second parallelization level is operated by means of multi-threading and shared memory *** experiments show that this method allows to handle large scale open domains for parallelcomputing purposes on few *** acoustic pollution arrising from car traffic was simulated on a large model of the Shinjuku district of Tokyo,*** good speed-up results illustrate the performance of this new domain decomposition method.
Particle swarm optimization is a powerful technique for computer aided prediction of proteins' three-dimensional structure. In this work, employing an all-atom force field and the standard algorithm, as implemente...
详细信息
Particle swarm optimization is a powerful technique for computer aided prediction of proteins' three-dimensional structure. In this work, employing an all-atom force field and the standard algorithm, as implemented in the ArFlock library in previous work, the low-energy conformations of several peptides of different sizes in vacuum starting from completely extended conformations are investigated. The computed structures are in good overall agreement with experimental data and results from other computer simulations. Periodic boundary conditions applied to the search space improve the performance of the method dramatically, especially when the linear velocity update rule is used. It is also shown that asynchronous parallelization speeds up the simulation better than the synchronous one and reduces the effective time for predictions significantly.
Current undergraduate parallel and distributed computing course faces several problems such as neglecting the importance of the course, lack of programming practice, etc. As multi-core computers spreads, it is necessa...
详细信息
ISBN:
(纸本)9781424441983
Current undergraduate parallel and distributed computing course faces several problems such as neglecting the importance of the course, lack of programming practice, etc. As multi-core computers spreads, it is necessary to improve the course. The multi-core technology is included as an indispensable teaching material, parallel programming is enhanced and several practical projects are designed to help students in programming training. After the improvement, the course is more effective and suitable for the demands of the students.
In the last decade, agent-based modeling and simulation (ABMS) has been applied to a variety of domains, demonstrating the potential of this technique to advance science, engineering, and policy analysis. However, rea...
详细信息
In the last decade, agent-based modeling and simulation (ABMS) has been applied to a variety of domains, demonstrating the potential of this technique to advance science, engineering, and policy analysis. However, realizing the full potential of ABMS to find breakthrough research results requires far greater computing capability than is available through current ABMS tools. The Repast for High Performance computing (Repast HPC) project addresses this need by developing a useful and useable next-generation ABMS system explicitly focusing on larger-scale distributedcomputing platforms. Repast HPC is intended to smooth the path from small-scale simulations to large-scale distributed simulations through the use of a Logo-like system. This article's contribution is its detailed presentation of the implementation of Repast HPC as a useful and usable framework, a complete ABMS platform developed explicitly for larger-scale distributedcomputing systems that leverages modern C++ techniques and the ReLogo language.
Nowadays, data-intensive problems are so prevalent that numerous organizations in various industries have to face them in their business operation. It is often crucial for enterprises to have the capability of analyzi...
详细信息
Nowadays, data-intensive problems are so prevalent that numerous organizations in various industries have to face them in their business operation. It is often crucial for enterprises to have the capability of analyzing large volumes of data in an effective and timely manner. MapReduce and its open-source implementation Hadoop dramatically simplified the development of parallel data-intensive computing applications for ordinary users, and the combination of Hadoop and cloud computing made large-scale parallel data-intensive computing much more accessible to all potential users than ever before. Although Hadoop has become the most popular data management framework for parallel data-intensive computing in the clouds, the Hadoop scheduler is not a perfect match for the cloud environments. In this paper, we discuss the issues with the Hadoop task assignment scheme, and present an improved scheme for heterogeneous computing environments, such as the public clouds. The proposed scheme is based on an optimal minimum makespan algorithm. It projects and compares the completion times of all task slots' next data block, and explicitly strives to shorten the completion time of the map phase of MapReduce jobs. We conducted extensive simulation to evaluate the performance of the proposed scheme compared with the Hadoop scheme in two types of heterogeneous computing environments that are typical on the public cloud platforms. The simulation results showed that the proposed scheme could remarkably reduce the map phase completion time, and it could reduce the amount of remote processing employed to a more significant extent which makes the data processing less vulnerable to both network congestion and disk contention.
This paper proposes parallel and distributed algorithms for solving very large-scale sparse optimization problems on computer clusters and clouds. Modern datasets usually have a large number of features or training sa...
详细信息
ISBN:
(纸本)9781479923908
This paper proposes parallel and distributed algorithms for solving very large-scale sparse optimization problems on computer clusters and clouds. Modern datasets usually have a large number of features or training samples, and they are usually stored in a distributed manner. Motivated by the need of solving sparse optimization problems with large datasets, we propose two approaches including (i) distributed implementations of prox-linear algorithms and (ii) GRock, a parallel greedy block coordinate descent method. Different separability properties of the objective terms in the problem enable different data distributed schemes along with their corresponding algorithm implementations. We also establish the convergence of GRock and explain why it often performs exceptionally well for sparse optimization. Numerical results on a computer cluster and Amazon EC2 demonstrate the efficiency and elasticity of our algorithms.
This paper presents design, implementation, and performance evaluation results of a parallel particle filter (PF) and a particle flow filter (PFF) using a Graphics Processing Unit (GPU) as a parallelcomputing environ...
详细信息
ISBN:
(纸本)9786058631113
This paper presents design, implementation, and performance evaluation results of a parallel particle filter (PF) and a particle flow filter (PFF) using a Graphics Processing Unit (GPU) as a parallelcomputing environment to speedup the computation. Simulation results from a high dimensional nonlinear filtering problem show that, for the considered example, the parallel PFF implementation is significantly superior to the parallel PF implementation in both estimation accuracy and computational performance. It is demonstrated that using GPU can markedly accelerate both particle filters and particle flow filters through parallelization.
In today’s digital environment, distributed systems are increasingly present in a wide variety of environments, ranging from public software applications to critical systems. distributed Systems introduces the underl...
详细信息
ISBN:
(数字)9781118601365
ISBN:
(纸本)9781848212503
In today’s digital environment, distributed systems are increasingly present in a wide variety of environments, ranging from public software applications to critical systems. distributed Systems introduces the underlying concepts, the associated design techniques and the related security issues. distributed Systems: Design and Algorithms, is dedicated to engineers, students, and anyone familiar with algorithms and programming, who want to know more about distributed systems. These systems are characterized by: several components with one or more threads, possibly running on different processors; asynchronous communications with possible additional assumptions (reliability, order preserving, etc.); local views for every component and no shared data between components. This title presents distributed systems from a point of view dedicated to their design and their main principles: the main algorithms are described and placed in their application context, i.e. consistency management and the way they are used in distributed file-systems.
Current undergraduate parallel and distributed computing course faces several problems such as neglecting the importance of the course, lack of programming practice, etc. As multi-core computers spreads, it is necessa...
详细信息
Current undergraduate parallel and distributed computing course faces several problems such as neglecting the importance of the course, lack of programming practice, etc. As multi-core computers spreads, it is necessary to improve the course. The multi-core technology is included as an indispensable teaching material, parallel programming is enhanced and several practical projects are designed to help students in programming training. After the improvement, the course is more effective and suitable for the demands of the students.
We present a parallel algorithm that solves a time-domain non-linear mathematical model of the cochlea. The previously known serial solution of the cochlear model is recursive in the longitudinal dimension and iterati...
详细信息
We present a parallel algorithm that solves a time-domain non-linear mathematical model of the cochlea. The previously known serial solution of the cochlear model is recursive in the longitudinal dimension and iterative in the time dimension. These two characteristics of the serial solution limit parallelism and prevent efficient computations on a massively parallel processor. We introduce a novel parallel algorithm that successfully overcomes the challenges posed by the cochlear model. We present performance results of a parallel implementation of the algorithm that shortens the computation time by a typical factor of 160 – 180, which makes the proposed algorithm of practical value for applications such as clinical audiological diagnosis.
暂无评论