R. Fatoohi and S. Weeratunga, Performance evaluation of three distributed computing environments for scientific applications, in Proc. of Supercomputing '94," pp. 400#408.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....recent years. A cluster of computers is a federation of computers linked by an interconnection network where the computers run integration software to support collaborative computations. Many positive results on using clusters of computers for load sharing and parallel computing have been reported [10, 13, 2]. A preliminary version of this paper appears in Proc. of 1999 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA) June 1999. Correspondences should be sent to M.J.M. Ma, Department of Computer Science and Information Systems, The University of ....

R. Fatoohi and S. Weeratunga, "Performance Evaluation of Three Distributed Computing Environments for Scientific Applications", Supercomputing 9, pp. 400-408. 37

....on networks of workstation, since a higher cost effectiveness can be achieved, as compared to current parallel supercomputers. In order to show the benefits of this approach, many studies have been carried out which evaluate the performance over several distributed environments and networks [2, 4], including ATM networks. A review of the state of the art in approaches to efficiently support parallel computing over ATM networks can be found in [10] and is summarized in the following. Replacing traditional LAN technologies: Traditional technologies such as Ethernet and FDDI are replaced by ....

R. Fatoohi and S. Weeratunga. Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of Supercomputing'94, pages 400--409, 1994.

....itself. In order to achieve the desired performance, this architecture has to be implemented in such a manner that the performance degradation necessarily introduced be as low as possible. In HPDC systems, the main challenge is to minimize the impact of the network on the global performance [4]. The increasing bandwidth provided by emerging networking technologies is an importantstep towards this goal; however, other factors contributing to communications time like transport protocols and network interfaces have not improved at the same rate [13] Therefore, the network technology is no ....

....in Section 4. Finally, Section 5 summarizes the paper. 2. Replacing traditional LAN technologies The attractiveness of parallel computing over LANs has lead to some studies of both the performance of workstation clusters over legacy networks and their limitations due to this assumption, as in [4], where several cluster environments, based on legacy LANs, are compared to multicomputers. The achieved performance can potentially be improved with the introduction of a high speed network technology such as ATM. In [10] the performance achieved with traditional LAN technologies has been ....

[Article contains additional citation context not shown here]

R. Fatoohi and S. Weeratunga. Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of Supercomputing'94, pages 400--409, 1994.

....recent years. A cluster of computers is a federation of computers linked by an interconnection network where the computers run integration software to support collaborative computations. Many positive results on using clusters of computers for load sharing and parallel computing have been reported [10, 13, 2]. The Java programming language [5] has received unprecedented acceptance for a programming language since its introduction in late 1994. Java provides a multi threaded model for concurrent programming as a built in feature, which could have great potentials for parallel processing. However, in ....

R. Fatoohi and S. Weeratunga, "Performance Evaluation of Three Distributed Computing Environments for Scientific Applications", Proc. of Supercomputing '94, pp. 400--408.

....benchmarks are implemented in Fortran 77 with a few common extensions that are also part of Fortran 90. They use MPI for communication but can be run on one processor with a dummy MPI library that performs negligible work. Anumber of NPB implementationshave been produced by research projects [13, 14, 15, 16, 17, 18]. What distinguishes NPB 2.0 from other publicly available source code implementations of NPB 1.0 First, NPB 2.0 codes were designed as benchmarks, and are not a byproduct of a related research project. They are carefully coded to use modern algorithms, avoid unnecessary computation, and ....

Fatoohi, R.; and Weeratunga, S.: "Performance Evaluation of Three Distributed Computing Environments for Scientific Applications," Supercomputing '94, IEEE 1063-9535/94.

....recent years. A cluster of computers is a federation of computers linked by an interconnection network where the computers run integration software to support collaborative computations. Many positive results on using clusters of computers for load sharing and parallel computing have been reported [10, 13, 2]. A preliminary version of this paper appears in Proc. of 1999 International Conference on Parallel and Distributed Processing Techniques and Applications (PDPTA) June 1999. Correspondences should be sent to M.J.M. Ma, Department of Computer Science and Information Systems, The University of ....

R. Fatoohi and S. Weeratunga, \Performance Evaluation of Three Distributed Computing Environments for Scientic Applications", Supercomputing '94, pp. 400-408. 37

....A cluster of computers is a federation of computers linked by an interconnection network where the computers are running integration software for performing collaborative computation. Many promising results have been reported on using clusters of computers for load sharing and parallel computing [13, 18]. With the advent of high speed networking and microprocessor technologies, Cluster Computing has emerged as a favorable alternative to Massively Parallel Machines for high performance and large scale computing. In general, computers belonging to a cluster are loosely coupled and they do not share ....

R. Fatoohi and S. Weeratunga, `Performance Evaluation of Three Distributed Computing Environments for Scientific Applications', Supercomputing `94, pp. 400-408.

....itself. In order to achieve the desired performance, this architecture has to be implemented in such a manner that the performance degradation necessarily introduced be as low as possible. In HPDC systems, the main challenge is to minimize the impact of the network on the global performance [4]. The increasing bandwidth provided by emerging networking technologies is an important step towards this goal; however, other factors contributing to communications time like transport protocols and network interfaces have not improved at the same rate [13] Therefore, the network technology is ....

....are described in Section 4. Finally, Section 5 summarizes the paper. 2 HPDC over Local Area Networks The attractiveness of parallel computing over LANs has lead to some studies of both the performance of workstation clusters over legacy networks and their limitations due to this assumption. In [4] several cluster environments, based on legacy LANs, are compared to multicomputers. In [10] the performance of such technologies has also been measured and compared to that of ATM over an equivalent environment. 2.1 HPDC with legacy LAN technologies The measurements in [4] include both ....

[Article contains additional citation context not shown here]

R. Fatoohi and S. Weeratunga. Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of Supercomputing'94, pages 400--409, 1994.

....have triggered the research in high performance distributed computing (HPDC) using workstation clusters. In order to show the benefits of this approach to parallel computing, several studies have been carried out which evaluate the performance over several distributed environments and networks [1, 5]. Many works focus on the particular case of ATM networks. A review of the state of the art in mechanisms to efficiently support HPDC over ATM networks can be found in [6] In the following we summarize the strategies than have been undertaken in the literature. Replacing traditional LAN ....

R. Fatoohi and S. Weeratunga. Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of Supercomputing'94, pages 400--409, 1994.

....in network performance have triggered the research in parallel computing using workstation clusters. In order to show the benefits of this approach to parallel computing, several studies have been carried out which evaluate the performance over several distributed environments and networks [5]. This idea of using networked workstations for high performance computing is the basis of the NOW (Networks of Workstations) project in Berkeley [1] It focus on three directions: low overhead communications, operating systems and file systems. As far as communications are concerned, the goal is ....

R. Fatoohi and S. Weeratunga. Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of IEEE Supercomputing, pages 400--409, 1994.

....In many cases, a code will perform significantly better on one system than another given where the strengths of that system lie and the demands on resources placed by the executing program. This is of particular importance when evaluating different architectures for specific applications [4] [7] [12] Adaptive parallel finite element codes represent a class of applications that could benefit significantly from the analysis approach we propose. Typically these codes impose significant computational demand, hence the need for parallelism. Analysis is difficult from an experiment management ....

R. Fatoohi and S. Weeratunga. Performance evaluation of three distributed computing environments for scientific applications. In Proceedings Supercomputing '94, pages 400--409, Washington, DC, USA, November 1994.

....availability of programming tools, such as the Parallel Virtual Machine (PVM) have made distributed computing more appealing for scientists and engineers. However, the conventional network technologies may not be quite adequate for such applications due to their high latency and low throughput [6]. This might change with the emerging technologies that promise higher bandwidth, lower latency and richer connectivity. The emerging high speed networks can theoretically transmit data per link at a hundred megabytes per second. Combined with high connectivity (switched network) the aggregate ....

....are not conclusive for the bidirectional test where the switched networks showed some performance degradations, but by less than 50 . 7 Communication tests: PVM Two Fortran programs were used to measure the throughput and latency of these networks under the PVM message passing library; see [6] for more details. In the first program, called ring, the processors form a ring where each processor receives a message of prescribed length from a previous processor and sends the same message to the next processor. Only one message goes around the ring at any given time. This program measures ....

[Article contains additional citation context not shown here]

Fatoohi, R. and Weeratunga, S., Performance Evaluation of Three Distributed Computing Environments for Scientific Applications. In Proceedings of Supercomputing '94, (Washington, November 1994), IEEE Computer Society Press, pp. 400 -- 409.

No context found.

R. Fatoohi and S. Weeratunga, Performance evaluation of three distributed computing environments for scientific applications, in Proc. of Supercomputing '94," pp. 400#408.

No context found.

R. Fatoohi and S. Weeratunga, `Performance Evaluation of Three Distributed Computing Environments for Scientific Applications', Supercomputing `94, pp.400-408.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC