Software-based communication latency hiding for commodity workstation networks

Strumpen, Volker

doi:10.1109/icpp.1996.537155

Cited by 12 publications

(9 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our approach to distributed simulation is novel practically as it does not involve explicit message passing as in [33,34], but rather provides a distributed implementation of the same abstract data types as in the sequential simulator. The only programmer's obligation is to spawn the number of threads sufficient for hiding communication latency caused by the transparent distribution.…”

Section: Related Workmentioning

confidence: 99%

Parallel Agent-Based Simulation on a Cluster of Workstations

Popov

Rafea

Holmgren

et al. 2003

Parallel Process. Lett.

View full text Add to dashboard Cite

We discuss a parallel implementation of an agent-based simulation. Our approach allows to adapt a sequential simulator for large-scale simulation on a cluster of workstations. We target discrete-time simulation models that capture the behavior of Web users and Web sites. Web users are connected with each other in a graph resembling the social network. Web sites are also connected in a similar graph. Users are stateful entities. At each time step, they exhibit certain behaviour such as visiting bookmarked sites, exchanging information about Web sites in the "word-of-mouth" style, and updating bookmarks. The real-world phenomena of emerged aggregated behavior of the Internet population is studied. The system distributes data among workstations, which allows large-scale simulations infeasible on a stand-alone computer. The model properties cause traffic between workstations proportional to partition sizes. Network latency is hidden by concurrent simulation of multiple users. The system is implemented in Mozart that provides multithreading, dataflow variables, component-based software development, and network-transparency. Currently we can simulate up to 10 6 Web users on 10 4 Web sites using a cluster of 16 computers, which takes few seconds per simulation step, and for a problem of the same size, parallel simulation offers speedups between 11 and 14.

show abstract

Section: Related Workmentioning

confidence: 99%

Parallel Agent-Based Simulation on a Cluster of Workstations

Popov

Rafea

Holmgren

et al. 2003

Parallel Process. Lett.

View full text Add to dashboard Cite

show abstract

“…This implementation reduces the total communication latency by a factor of 3-4. In [6,7] new designs using multithread implementation to further hide the communication latency are discussed.…”

Section: Parallelizable Portion Of Ufdamentioning

confidence: 99%

<title>Parallel implementation of a unified approach to image focus and defocus analysis on the Parallel Virtual Machine</title>

Liu

Subbarao

et al. 1998

SPIE Proceedings

View full text Add to dashboard Cite

A unified approach to image focus and defocus analysis (UFDA) was proposed recently for three-dimensional shape and focused image recovery of objects. One version of this approach which yields very accurate results is highly computationally intensive. In this paper we present a parallel implementation of this version of UFDA on the Parallel Virtual Machine (PVM). One of the most computationally intensive parts of the UFDA approach is the estimation of image data that would be recorded by a camera for a given solution for 3D shape and focused image. This computational step has to be repeated once during each iteration of the optimization algorithm. Therefore this step has been sped up by using the Parallel Virtual Machine (PVM). PVM is a software package that allows a heterogeneous network of parallel and serial computers to appear as a single concurrent computational resource. In our experimental environment PVM is installed on four UNIX workstations communicating over Ethernet to exploit parallel processing capability. Experimental results show that the communication overhead in this case is relatively low. An average of 1.92 speedup is attained by the parallel UFDA algorithm running on 2 PVM connected computers compared to the execution time ofsequential processing. By applying the UFDA algorithm on 4 PVM connected machines an average of 3.44 speedup is reached. This demonstrates a practical application of PVM to 3D machine vision.

show abstract

“…To minimize the communication overhead, many softwarebased mechanisms [17,18] 15] have also been proposed to improve the efficiency of complicated communication routines. In addition, effective graph partitioning strategies [6,10] have been developed by the researchers to partition a task into several subtasks among which the communication cost was minimized.…”

Section: Background and Related Workmentioning

confidence: 99%

A Framework for Solving Sequence Problem of Multiple Input Streams

Lin

2007

Int. J. Found. Comput. Sci.

View full text Add to dashboard Cite

In this research, a framework is proposed for answering the question: whether a dominant sequence exists for the multiple self-contained input streams destined for a computation provider running the requested applications concurrently. The framework is extended to find an optimal semidominant sequence for multiple input streams. This paper also shows that an existent universal dominant sequence can be formed by permuting a set of dominant input streams. In this paper, the neighbor-connected graph is first proposed to model the sequence problem of multiple input streams. Then, an algorithm with time complexity 0(nL n )is proposed to determine the existence of a dominant data-receiving sequence, where L is the maximal length of the input streams plus one and n is the number of input streams. Examples of applying the framework to solve the sequence problems for several computation-intensive applications are also given. Furthermore, the framework is extended to find an optimal semidominant sequence using an algorithm with time complexity 0(nL n log5*), where <5* is the maximal degree that the sequence diverges from all the optimal subsequences. Finally, the property of a universal dominant sequence is studied and the performance of the sequence is compared with those of other sequences.

show abstract

Software-based communication latency hiding for commodity workstation networks

Cited by 12 publications

References 27 publications

Parallel Agent-Based Simulation on a Cluster of Workstations

Parallel Agent-Based Simulation on a Cluster of Workstations

<title>Parallel implementation of a unified approach to image focus and defocus analysis on the Parallel Virtual Machine</title>

A Framework for Solving Sequence Problem of Multiple Input Streams

Contact Info

Product

Resources

About