#### DMCA

## Collision Graph based Communication Scheduling with Re-routing in Parallel Systems (1997)

### Citations

13994 |
Computers and Intractability: A Guide to the Theory of NPcompleteness
- Garey, Johnson
- 1979
(Show Context)
Citation Context ...dressed in a very restricted way in [18--20] but what is developed here is a new scheduling framework able to deal with real-life problems. Just as multi-processor scheduling problems are NP-Complete =-=[21]-=- for most precedence-constrained tasks, the communication scheduling problem or CSP is shown to be NP-Complete as well. Because of this, heuristic methods are employed to arrive at the communication s... |

1311 |
Computer Networks
- Tanenbaum
- 1988
(Show Context)
Citation Context ...cheduling problem [10--12]. There have been several studies related to the problem addressed here. Many of them deal with computer networks and the store-and-forward which occurs using routing tables =-=[13]-=-. Recent efforts have 2 focused on developing communication algorithms for interconnection networks. One such work was done by Bianchini and Shen [14]. There a `traffic scheduling' algorithm for multi... |

531 | A survey of wormhole routing techniques in direct networks.
- Ni, McKinley
- 1993
(Show Context)
Citation Context ...lumns of Table 1 give possible orderings of the message traffic when XY-routing is being used. Messages appearing on the same line may be sent in parallel without collisions occurring. Under wormholes=-=[7]-=- routing assumptions that each message takes the same amount of time, t, to traverse the network, and assuming that the messages are of equal length, these two orderings yield different completion tim... |

497 | Principles of Sequencing and Scheduling - Baker, Trietsch - 2009 |

309 | Theory of Scheduling - Conway, Maxwell, et al. - 1967 |

120 | Practical Multiprocessor Scheduling Algorithms for Efficient Parallel Processing. - Kasahara, Narita - 1984 |

98 |
Partitioning and mapping algorithms into fixed size systolic arrays
- Moldovan, Fortes
- 1986
(Show Context)
Citation Context ...vidual processing elements. Then, communication costs stem from the underlying message passing which occurs. This research assumes that a suitable task allocation scheme, such as the one presented in =-=[3]-=-, has been used and deals specifically with the message transmissions. Therefore, the new scheduling technique is not a type of multi-processor scheduling, but rather deals strictly with determining t... |

44 |
EXECUBE - A new architecture for scalable MPPs,”
- Kogge
- 1994
(Show Context)
Citation Context ...ored by NSF, ARPA, DOD and others, our research group was challenged by the need of obtaining an optimized execution time based on tightly-coupled massively parallel architectures such as the EXECUBE =-=[1]-=-. One application being studied is the implementation of a parallel solution for simulating partial differential equations, representing fluid dynamics problems, by using the multi-dimensional wave di... |

36 |
Numerical integration of partial differential equations by means of multidimensional wave digital filters”, Proc
- Fettweis, Nitsche
- 1990
(Show Context)
Citation Context ...ng studied is the implementation of a parallel solution for simulating partial differential equations, representing fluid dynamics problems, by using the multi-dimensional wave digital filters method =-=[2]-=-. In addition to the large amounts of computation time needed, significant communication time between processors was required by the data volume involved in those simulations. While using multiple pro... |

36 |
Interprocessor Traffic Scheduling Algorithm for Multiple- Processor Networks
- Shen
- 1987
(Show Context)
Citation Context ...ore-and-forward which occurs using routing tables [13]. Recent efforts have 2 focused on developing communication algorithms for interconnection networks. One such work was done by Bianchini and Shen =-=[14]-=-. There a `traffic scheduling' algorithm for multi-processor networks was introduced to try and balance and saturate the links of the network based on the fact that a large number of messages must eve... |

26 | A Compile-time Technique for Controlling Real-time Execution of Task-level Data-flow Graphs," presented at - Shukla, Little, et al. |

22 |
Traffic routing for multicomputer networks with virtual cut-through capability
- Kandlur, Shin
- 1992
(Show Context)
Citation Context ...need to communicate. They also do not consider the scheduling of the individual messages. Our work deals with these individual message transfers and does not dedicate paths to nodes. Kandlur and Shin =-=[15]-=- present a work similar to [8] in that dedicated paths are found. The problem with these techniques is that seldom used dedicated paths can cause other messages to follow longer paths when the needed ... |

18 |
Path selection for communicating tasks in a wormhole-routed multicomputer
- Lee, Kim
- 1994
(Show Context)
Citation Context ...n time that is the problem addressed in this paper. The term used for this research is communication scheduling. Not only does it encompasses routing aspects and path selection issues as discussed in =-=[8, 9]-=-, it also determines the order and timing that the messages in the system should be sent. Thus, this problem is much different from the traditional multi-processor scheduling problem [10--12]. There h... |

13 | Scheduling sequential - Munshi, Simons - 1990 |

12 |
Deadlock-free message routing inmultiprocessor interconnection networks
- Dally, Seitz
- 1987
(Show Context)
Citation Context ...n time that is the problem addressed in this paper. The term used for this research is communication scheduling. Not only does it encompasses routing aspects and path selection issues as discussed in =-=[8, 9]-=-, it also determines the order and timing that the messages in the system should be sent. Thus, this problem is much different from the traditional multi-processor scheduling problem [10--12]. There h... |

12 |
Partitioning and mapping algorithms into xed size systolic arrays
- Moldovan, Fortes
- 1986
(Show Context)
Citation Context ...vidual processing elements. Then, communication costs stem from the underlying message passing which occurs. This research assumes that a suitable task allocation scheme, such as the one presented in =-=[3]-=-, has been used and deals specifically with the message transmissions. Therefore, the new scheduling technique is not a type of multi-processor scheduling, but rather deals strictly with determining t... |

10 |
Contention-free communication scheduling on 2d meshes
- Eberhart, Li
- 1996
(Show Context)
Citation Context ...nger paths when the needed links which are dedicated are not being used. Additionally, they do not use any type of scheduling which can improve the overall performance. Recent work by Eberhart and Li =-=[16]-=- does perform a type of communication scheduling on two-dimensional mesh architectures. However, they use dynamic scheduling and restrict their work to communication patterns that are commonly used in... |

10 | Application specific communication scheduling on parallel systems
- Surma, Sha
- 1995
(Show Context)
Citation Context ...munication patterns that are commonly used in data parallel applications. The work presented here can apply to any type of message-passing activity. By using a new model presented by Surma and Sha in =-=[17]-=- known as a Collision Graph, this paper starts the research on the compile-time analysis of the run-time communication overhead incurred by point-to-point message transmissions. This problem was addre... |

9 | Optimal communication scheduling based on collision graph model - Surma, Tongsima, et al. - 1996 |

9 | Static communication scheduling for minimizing collisions in application-speci c parallel systems - Surma, Sha - 1996 |

8 |
Numerical integration of partial di erential equations using principles of multidimensional wave digital lters
- Fettweis, Nitsche
- 1991
(Show Context)
Citation Context ...n being studied is the implementation of a parallel solution for simulating partial di erential equations, representing uid dynamics problems, by using the multi-dimensional wave digital lters method =-=[2]-=-. In addition to the large amounts of computation time needed, signi cant communication time between processors was required by the data volume involved in those simulations. While using multiple proc... |

6 |
Interprocessor tra c scheduling algorithm for multiple-processor networks
- Bianchini, Shen
- 1987
(Show Context)
Citation Context ...tore-and-forward which occurs using routing tables [13]. Recent e orts have 2focused on developing communication algorithms for interconnection networks. One such work was done by Bianchini and Shen =-=[14]-=-. There a `tra c scheduling' algorithm for multi-processor networks was introduced to try and balance and saturate the links of the network based on the fact that a large number of messages must event... |

4 |
Application speci c communication scheduling on parallel systems
- Surma, Sha
- 1995
(Show Context)
Citation Context ...munication patterns that are commonly used in data parallel applications. The work presented here can apply to any type of message-passing activity. By using a new model presented by Surma and Sha in =-=[17]-=- known as a Collision Graph, this paper starts the research on the compile-time analysis of the run-time communication overhead incurred by point-to-point message transmissions. This problem was addre... |

3 | Hybrid static-dynamic communication scheduling for parallel systems - Surma, Sha - 1997 |