A. Grimshaw et al. Metasystems: An approach combining parallel processing and heterogenous distributed computing systems. In JPDC, 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the hybrid compile and run time system (Section 5) Finally, we present the modeling and experimental results (Section 6) and our conclusions (Section 7) 2. RELATED WORK Compile time static loop scheduling has been well studied [9, 14] Static scheduling for heterogeneous NOWs was proposed in [4, 5, 7]. The task queue model for dynamic loop scheduling has targeted shared memory machines [11, 14] while the diffusion model has been used for distributedmemory machines [10] A method for task level scheduling in heterogeneous programs was proposed in [13] and [2] presents an application specific ....

Grimshaw, A. S., et al. Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems. J. Parallel Distrib. Comput. 21, 3 (1994).

.... networked environment have been called a metacomputer ( 11] E A metasystem is a system composed of heterogeneous hosts (both parallel processors and conventional architectures) possibly controlled by separate organizational entities, and connected by an irregular interconnection network ([8]) EIE Metasystem is a wide area environment in which users operate transparently, consisting in workstations, PCs, graphics rendering engines, supercomputers and nontraditional computing devices such as televisions ( 9] EI E Networked virtual supercomputers, or metacomputers, are ....

Grimshaw, A. S., Weissman, J. B., West, E. A., Loyot Jr., E.: Metasystems: an Approach Combining Parallel Processing and Distributed Heterogeneous Computing System. Parallel and Distributed Computing. 21 (3) (1994) 257-270

....Legion Core Objects that manage the persistent states of inactive objects, and Legion Class Objects which manage the association between their persistent instances and Vault objects. 1 Introduction The basic goal of the Legion project[3] is to construct a wide area, high performance metasystem[2] capable of providing a usable programmer interface to the increasingly complex nation wide computing and communications infrastructure. The basic unit of program composition and scheduling in Legion is the active object, and the programming interfaces supported by Legion are object oriented. ....

A.S. Grimshaw, J.B.Weissman, E.A. West, and E. Loyot, "Meta Systems: An Approach Combining Parallel Processing And Heterogeneous Distributed Computing Systems," Journal of Parallel and Distributed Computing, pp. 257-270, Vol. 21, No. 3, June 1994.

....processors, the scheduling problem can be informally stated as deciding which processor executes which task when. One solution is to statically partition the work among the processors (Static Scheduling) at compile time [Wol96, Pol88, LP93] Such schemes have been implemented on NOWs [CLZ95, CR92, Gea94] Another solution is to hand out tasks one at a time to a free processor requesting work (Simple (Dynamic) Scheduling) Pol88, ML94] More complicated dynamic strategies have also been proposed (in some cases application speci c) LK87, NS93, Bea96] The scheduling policy for the D DOALL speci ....

A. S. Grimshaw and et al. Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Dist. Computing, 21(3), 1994. 27

....from local cluster management, several schemes have been devised for high throughput computing on a supernational scale. They include the Iowa State University s Batrun [35] the Flock of Condors used in the Dutch Po l der initiative [15] the Nimrod project [1] and the object oriented Legion [20] which proved useful in a nation wide cluster. While these schemes emphasize mostly the application support on homogeneous systems, the AppLeS project [8] provides applicationlevel scheduling agents on heterogeneous systems, taking into account their actual resource performance. Another approach ....

A. Grimshaw, J. Weissman, E. West, and E. Loyot. Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems. J. Parallel Distributed Computing, Vol. 21, 1994, pp. 257--270.

....moving data to the nodes of a supercomputer or, more recently,of a network of workstations. Zaki [14] provides an example of a load balancing method that optimizes the eciency of parallel computation on a network of compute nodes. Other examples and motivating discussion can be found in [2] and [4]. These techniques, however, do not directly target the speci c issues that arise in distributed data mining, suchas ways of combining predictive models and the accuracy of the resulting predictivesystem. The work presented in this paper, to our knowledge, is the rst attempt to identify a ....

A. Grimshaw, J. Weissman, E. West, and E. Loyot. Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21(3):257-270, 1994.

....by the granularity of the AP, it is easy to control the granularity by adjusting the total number of abstract processors. 2.2 Abstract Processor Assignment Strategy There exist two fundamental models of execution at the RP when two or more APs are assinged to it. They are the multi thread model[4] and the single thread model. In our compiler, we adopted the singlethread model because the overhead of handling irregular data region is much smaller than the overhead of thread scheduling in multi thread model under currently available operating systems. In the single thread model, APs ....

Grimshaw, A. S., Weissman, J. B. et al.: "Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems," Journal of Parallel and Distributed Computing,21(3).pp.257-270, 1994.

....type prior to its operational debut. An adaptive allocation and scheduling scheme then uses the program s envelope to adjust its execution to prevailing runtime circumstances. A rudimentary application of program supplied information has been successfully demonstrated in the Mentat system [2]. The adaptive moves are guided by information encoded in the envelope at an earlier time. In this role analytic models certainly work, as do empirical ones. Note, however, that the scheduler demands no explanation of why a program behaves as it does. The scheduler wants only an accurate map the ....

GRIMSHAW A. et al., Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems, Jour. Parallel and Dist. Computing, 21, 1994, p. 257-270.

....mostly based on PVM. A multitude of schemes have been devised for high throughput computing on a somewhat larger scale, including the Iowa State University s Batrun[23] the CORBA based Piranha[24] the Dutch Po l der initiative[25] the Nimrod project[26] and the object oriented Legion[27] which proved useful in a nation wide cluster. While these schemes emphasize mostly on application support on homogeneous systems, the AppLeS project[28] provides application level scheduling agents on heterogeneous systems, taking into account their actual resource performance. 6.1. Comparing ....

A. Grimshaw, J. B. Weissman, E. A. West and E. C. Loyot, `Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems', J. Parallel Distrib. Comput., 21, 257--270 (1994).

....by the granularity of the AP, it is easy to control the granularity by adjusting the total number of abstract processors. 2.2. Abstract Processor Assignment Strategy There exists two fundamental models of execution at the RP when two or more APs are assigned to it. They are the multi thread model[8] and the single thread model. In our compiler, we adopted the single thread model because the overhead of handling irregular data region is much smaller than the overhead of thread scheduling in multi thread model under currently available operating systems. In the single thread model, APs ....

....in the HCE however. In our implementation, we provide each real processor with an index table AP RP Map[ which indicates the relation of the AP number to the RP number, and let the real processor refer to this table to determine the owner of a data. For example, if the index table AP RP Map[8]= 0,2,1,2,0,3,3,3 is given, it means that AP 0 and AP 4 are assigned to RP 0 , AP 2 is assigned to RP 1 , AP 1 and AP 3 are assigned to RP 2 and so on, where AP i and RP i stands for abstract processor i and real processor i, respectively. Using this table, the condition of the following if ....

A. S. Grimshaw, J. B. Weissman, et al. Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21(3):257--270, 1994.

....of Sun3 s, Sun4 s, the IBM RS 6000, Silicon Graphics, and Gamma s. Mentat is available from http: www.cs.virginia.edu mentat ftp: uvacs.cs.virginia.edu 28 Email addresses: group Gamma mentat virginia.edu Andrew S. Grimshaw Gamma grimshaw virginia.edu References: 89] 90] 91] 92] [93] [94] 2.67 Meyer s Proposal Developer: Proposal by Betrand Meyer Description: oo. Extension of Eiffel. Meyer considers the Design by Contract [164] as the basic principle of objectoriented programming. He introduces the new keyword separate into Eiffel. That keyword indicates that an object is ....

Andrew S. Grimshaw, Jon B. Weissman, Emily A. West, and Ed C. Loyot, Jr. Metasystems: An approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21(3):257-- 270, June 1994. Also University of Virginia, Charlottesville, VA, Technical Report CS-92-43.

No context found.

Grimshaw AS, Weismann JB, West EA, Loyot Jr EC. MetaSystems: An approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing 1994; 21(3):257--270.

....tool of big and small science, and it is shifting from the single physical component design to the design of a whole physical system. Increasingly, such simulations are being dynamically composed by scientists working collaboratively on the grid, even though they are geographically distributed [6, 7, 9, 10, 11, 12]. The realization of this simulation scenario requires that grid users be able to identify relevant software and its parameters before implementation on a remote computing resource. However, the obstacles to selecting the best solution approach for a particular problem, and subsequently finding ....

A.S. Grimshaw, J.B. Weismann, E.A. West, and Ed. C. Loyot Jr. MetaSystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems. Journal of Parallel and Distributed Computing, Vol. 21(3), June 1994.

....for creating and managing processes, communication, and scheduling. While these systems accommodate heterogeneity to some extent, they do not exploit heterogeneity. A number of systems that provide greater support for heterogeneous parallel processing have emerged over the past few years [27][31][33] 62] 76] These systems may be 19 distinguished by the level of compiler and runtime system support for managing parallelism and scheduling. Mentat [33] is an object oriented parallel processing system. Mentat programs are written in MPL, a high level language based on C . The programmer ....

....class instances. Instances of Mentat classes, known as Mentat objects, are implemented by address space disjoint processes, and communicate via methods. Method invocation is accomplished via an RPC like mechanism. A strategy for supporting data conversion of arbitrary data types is discussed in [31]. Mentat also performs runtime scheduling [37] based on Eager and Lazoswka s adaptive load sharing [21] Support for scheduling data parallel computations in heterogeneous environments has been recently added to the runtime scheduler [31] 91] as part of this thesis. Data parallelism is expressed ....

[Article contains additional citation context not shown here]

A.S. Grimshaw, J.B. Weissman, E.A. West, and E. Loyot, "Metasystems: An Approach Combining Parallel Processing And Heterogeneous Distributed Computing Systems," Journal of Parallel and Distributed Computing, Vol. 21(3), June 1994.

No context found.

Grimshaw, A.S., Weissman, J.B., West, E.A., Loyot, E., "Metasystems: an approach combining parallel processing and heterogeneous distributed computing systems," Journal of Parallel and Distributed Computing, vol. 21, no. 3, pp. 25769, June 1994.

.... and the growing availability of higher performance computing and networking hardware have made commonplace the use of networks of workstations, personal computers, and supercomputers as virtual, distributed memory parallel machines, or metasystems, for solving computationally demanding problems [1,34,38,41]. The combination of heterogeneous architectures and operating system platforms within a high performance distributed metasystem gives rise to a number of problems not present in homogenous distributed systems and parallel computers. The complexity of varying architectural features such as data ....

....strengthening the argument for increased employment of fault tolerance techniques. However, beyond these existing uses for process state capture and recovery, the availability of a such a mechanism in the metasystem context also opens up new possibilities. For example, a metasystem such as Legion[38] could use process state capture and recovery for improved resource management; if the number of active entities in the system became greater than could be efficiently supported, the system could temporarily preempt the execution of some jobs by checkpointing and terminating their processes, then ....

[Article contains additional citation context not shown here]

A.S. Grimshaw, J.B.Weissman, E.A. West, and E. Loyot, "Meta Systems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems," Journal of Parallel and Distributed Computing, pp. 257-270, vol. 21, no. 3, June 1994.

....system is PVM [12] but it is limited to support for communication, heterogeneity, and remote execution. Other projects deal specifically with achieving high performance via scheduling, but do so in a restricted distributed system such as a NOW or metasystem. Metasystem projects include [2] 4][5]. NOW projects include Condor [10] DQS [11] and LSF [18] All of these efforts concentrate on local as opposed to wide area scheduling. Much of the NOW scheduling research is based on classic distributed system scheduling models [1] 3] 3 Legion is a project designed to provide a comprehensive ....

A.S. Grimshaw, J.B. Weissman, E.A. West, and E. Loyot, "Metasystems: An Approach Combining Parallel Processing And Heterogeneous Distributed Computing Systems," Journal of Parallel and Distributed Computing, Vol. 21(3), June 1994. 14

No context found.

A. Grimshaw et al. Metasystems: An approach combining parallel processing and heterogenous distributed computing systems. In JPDC, 1994.

No context found.

Andrew S. Grimshaw, Jon B. Weissman, Emily A. West, and Ed. C. Loyot Jr. Metasystems: an approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21:257-270, 1994.

No context found.

Grimshaw, A., Weissman, J., West, E. and E. Lyot, J. Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems. Journal of Parallel and Distributed Computing, 21 (3). 257-270. 1994.

No context found.

Andrew S. Grimshaw, Jon B. Weissman, Emily A. West, and Ed. C. Loyot Jr. Metasystems: an approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21:257-270, 1994.

No context found.

Grimshaw, A., Weissman, J., West, E. and E. Lyot, J. Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems. Journal of Parallel and Distributed Computing, 21 (3). 257-270. 1994.

No context found.

Grimshaw, A., Weissman, J., West, E., and Loyot Jr., E. Metasystems: an Approach Combining Parallel Processing and Distributed Heterogeneous Computing System. Parallel and Distributed Computing, 21(3):257--270, 1994.

No context found.

Andrew S. Grimshaw, Jon B. Weissman, Emily A. West, and Ed. C. Loyot Jr. Metasystems: an approach combining parallel processing and heterogeneous distributed computing systems. Journal of Parallel and Distributed Computing, 21:257270, 1994. 19

No context found.

Grimshaw, A., Weissman, J., West, E., Loyot, E.: Metasystems: An Approach Combining Parallel Processing and Heterogeneous Distributed Computing Systems. J. Parallel Distributed Computing, Vol. 21, 1994, pp. 257#270.

First 50 documents  Next 50

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