Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242750
A. Reddy, R. Freitas
An evaluation of the effectiveness of adaptive routing techniques in tolerating failures is presented. It is shown that adaptive routing techniques yield gracefully degradable systems for the workloads considered. For medium to large communication granularity and the workloads considered in this study, if adaptive-routing is used, it is shown that the problem completion time does not increase drastically due to failures. When node failures were considered, it was observed that the mismatch of problem communication structure with the physical communication structure did not result in significant loss of performance. Since adaptive routing techniques are warranted for performance reasons, it is argued that making use of this adaptive routing hardware to tolerate failures is a favorable option.<>
{"title":"Fault tolerance of adaptive routing algorithms in multicomputers","authors":"A. Reddy, R. Freitas","doi":"10.1109/SPDP.1992.242750","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242750","url":null,"abstract":"An evaluation of the effectiveness of adaptive routing techniques in tolerating failures is presented. It is shown that adaptive routing techniques yield gracefully degradable systems for the workloads considered. For medium to large communication granularity and the workloads considered in this study, if adaptive-routing is used, it is shown that the problem completion time does not increase drastically due to failures. When node failures were considered, it was observed that the mismatch of problem communication structure with the physical communication structure did not result in significant loss of performance. Since adaptive routing techniques are warranted for performance reasons, it is argued that making use of this adaptive routing hardware to tolerate failures is a favorable option.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128874784","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242717
K. Makki, P. Banta, K. Been, N. Pissinou, E. Park
The authors present an algorithm for solving the k mutual exclusion problem in a distributed system. The algorithm is token based, whereby a token is passed among sites. Only sites that either receive the token with a nonzero semaphore or receive the token with a zero semaphore and later receive a release message from a previous site are allowed to enter their critical sections. Attached to the token is a queue which lists the sites scheduled to receive the token and a general semaphore. In all cases except extremely light token request traffic, the number of messages per critical section execution can be expressed as a small constant. This constant approaches three in an extremely heavy token request environment.<>
{"title":"A token based distributed k mutual exclusion algorithm","authors":"K. Makki, P. Banta, K. Been, N. Pissinou, E. Park","doi":"10.1109/SPDP.1992.242717","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242717","url":null,"abstract":"The authors present an algorithm for solving the k mutual exclusion problem in a distributed system. The algorithm is token based, whereby a token is passed among sites. Only sites that either receive the token with a nonzero semaphore or receive the token with a zero semaphore and later receive a release message from a previous site are allowed to enter their critical sections. Attached to the token is a queue which lists the sites scheduled to receive the token and a general semaphore. In all cases except extremely light token request traffic, the number of messages per critical section execution can be expressed as a small constant. This constant approaches three in an extremely heavy token request environment.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133806928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242742
R. M. Lee, V. Allan
Architectures such as horizontal microengines, multiple reduced instruction set computer (RISC) architectures, very long instruction word (VLIW), and long instruction word (LIW) machines benefit from the utilization of low-level parallelism. The authors propose a novel technique, advanced software pipelining, which solves known problems of other pipelining techniques. Implementation on the program dependence graph (PDG) benefits from the data dependence information and the control structure of the program in a single representation. Software pipelines are constructed on the PDG without imposing a fixed schedule; instead the proposed method groups concurrent operations in anticipation of the needs of the final schedule.<>
{"title":"Advanced software pipelining and the program dependence graph","authors":"R. M. Lee, V. Allan","doi":"10.1109/SPDP.1992.242742","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242742","url":null,"abstract":"Architectures such as horizontal microengines, multiple reduced instruction set computer (RISC) architectures, very long instruction word (VLIW), and long instruction word (LIW) machines benefit from the utilization of low-level parallelism. The authors propose a novel technique, advanced software pipelining, which solves known problems of other pipelining techniques. Implementation on the program dependence graph (PDG) benefits from the data dependence information and the control structure of the program in a single representation. Software pipelines are constructed on the PDG without imposing a fixed schedule; instead the proposed method groups concurrent operations in anticipation of the needs of the final schedule.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114815334","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242762
S. Miguet, Virginie Poty
The authors study the allocation of regular data arrays like images or matrices onto a network of processors. They investigate how to minimize the total communication time, while taking into account the time of constructing the messages. They show how to find the optimal mesh parameters as a function of problem-dependent and machine-dependent parameters. The main result is that, because of the reindexing time, the optimal mesh does not partition the data into square subdomains. It is also shown that the linear network (pipeline) very often performs better than a real 2-D mesh, since it avoids the reindexing of messages. They report numerical experiments performed on the TNode, a reconfigurable network of Transputers, and on the iPSC/860 hypercube, in which arbitrary grids can be embedded.<>
{"title":"Revisiting the allocation of regular data arrays to a mesh of processors","authors":"S. Miguet, Virginie Poty","doi":"10.1109/SPDP.1992.242762","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242762","url":null,"abstract":"The authors study the allocation of regular data arrays like images or matrices onto a network of processors. They investigate how to minimize the total communication time, while taking into account the time of constructing the messages. They show how to find the optimal mesh parameters as a function of problem-dependent and machine-dependent parameters. The main result is that, because of the reindexing time, the optimal mesh does not partition the data into square subdomains. It is also shown that the linear network (pipeline) very often performs better than a real 2-D mesh, since it avoids the reindexing of messages. They report numerical experiments performed on the TNode, a reconfigurable network of Transputers, and on the iPSC/860 hypercube, in which arbitrary grids can be embedded.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122491868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242734
T. Lai, Li-Fen Wu
The authors propose a fault-tolerant algorithm for termination detection that can tolerate any number of stop failures. This algorithm improves on existing ones in many aspects, including worst-case message complexity, average-case message complexity and storage overhead. Most important, it runs as efficiently as the best non-fault-tolerant algorithm available if no process fails during the computation, and incurs only a reasonable amount of cost for each process failure that actually occurs. The main idea of the algorithm is a barrier set at a particular process. It is possible that other termination detection algorithms can be made fault-tolerant using similar techniques.<>
{"title":"An (N-1)-resilient algorithm for distributed termination detection","authors":"T. Lai, Li-Fen Wu","doi":"10.1109/SPDP.1992.242734","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242734","url":null,"abstract":"The authors propose a fault-tolerant algorithm for termination detection that can tolerate any number of stop failures. This algorithm improves on existing ones in many aspects, including worst-case message complexity, average-case message complexity and storage overhead. Most important, it runs as efficiently as the best non-fault-tolerant algorithm available if no process fails during the computation, and incurs only a reasonable amount of cost for each process failure that actually occurs. The main idea of the algorithm is a barrier set at a particular process. It is possible that other termination detection algorithms can be made fault-tolerant using similar techniques.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122550728","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-12-01DOI: 10.1109/SPDP.1992.242741
A. Robison, D. Soroker
The operation of permuting data among the vertices of a hypercube computer induces a set of paths from senders to receivers. Permutations with edge-disjoint paths are desirable for efficient communication. The authors give simple algebraic descriptions for large classes of permutations that induce edge-disjoint paths for the commercially popular 'e-cube' routing algorithm. The descriptions cover most useful edge-disjoint permutations, and are easily applied in practice. Many previous proofs in the literature that specific permutations are edge-disjoint fall out as simple corollaries of the present work. Some new applications of this framework are presented. The first application considered concerns Gray code embeddings: the others are motivated by the connection of the present results to switching networks.<>
{"title":"An algebraic framework for edge-disjoint permutations on hypercubes","authors":"A. Robison, D. Soroker","doi":"10.1109/SPDP.1992.242741","DOIUrl":"https://doi.org/10.1109/SPDP.1992.242741","url":null,"abstract":"The operation of permuting data among the vertices of a hypercube computer induces a set of paths from senders to receivers. Permutations with edge-disjoint paths are desirable for efficient communication. The authors give simple algebraic descriptions for large classes of permutations that induce edge-disjoint paths for the commercially popular 'e-cube' routing algorithm. The descriptions cover most useful edge-disjoint permutations, and are easily applied in practice. Many previous proofs in the literature that specific permutations are edge-disjoint fall out as simple corollaries of the present work. Some new applications of this framework are presented. The first application considered concerns Gray code embeddings: the others are motivated by the connection of the present results to switching networks.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115307301","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors consider problems whose size varies with time, and discuss the organization and analysis of algorithms for them. They prove that the benefits of a parallel solution may be greater than usual, they also prove that the standard lower and upper bounds on parallel speed-up do not hold in this case, and must be reformulated.<>
{"title":"The p-shovelers problem-computing with time-varying data","authors":"F. Luccio, L. Pagli","doi":"10.1145/130956.130960","DOIUrl":"https://doi.org/10.1145/130956.130960","url":null,"abstract":"The authors consider problems whose size varies with time, and discuss the organization and analysis of algorithms for them. They prove that the benefits of a parallel solution may be greater than usual, they also prove that the standard lower and upper bounds on parallel speed-up do not hold in this case, and must be reformulated.<<ETX>>","PeriodicalId":265469,"journal":{"name":"[1992] Proceedings of the Fourth IEEE Symposium on Parallel and Distributed Processing","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127972429","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：