### Abstract

In this paper, we address multiway network partitioning problem of dividing the cells of network into multiple blocks so as to minimize the number of nets interconnecting cells in different blocks while balancing the blocks' sizes. The sequential iterative improvement algorithm for the problem consists of several passes each of which is performed by repeatedly iterating the move operation. Therefore, the whole execution time taken by the algorithm is greatly affected by the number of the move operations and the execution time for each move operation. We present a fast parallel algorithm for solving the multiway network partiotioning problem by reducing both of them. We propose a new dynamic iterative method which reduces the number of move operations executed at each pass dynamically, and hence speed up the whole algorithm sharply. Moreover, we reduced the execution time of each move by its parallelization using the proper cell distribution.

Original language | English |
---|---|

Title of host publication | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |

Publisher | Springer Verlag |

Pages | 81-88 |

Number of pages | 8 |

Volume | 1823 |

ISBN (Print) | 9783540675532 |

Publication status | Published - 2000 |

Event | 8th International Conference on High Performance Computing and Networking, HPCNEurope 2000 - Amsterdam, Netherlands Duration: 2000 May 8 → 2000 May 10 |

### Publication series

Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

Volume | 1823 |

ISSN (Print) | 03029743 |

ISSN (Electronic) | 16113349 |

### Other

Other | 8th International Conference on High Performance Computing and Networking, HPCNEurope 2000 |
---|---|

Country | Netherlands |

City | Amsterdam |

Period | 00/5/8 → 00/5/10 |

### Fingerprint

### ASJC Scopus subject areas

- Computer Science(all)
- Theoretical Computer Science

### Cite this

*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)*(Vol. 1823, pp. 81-88). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1823). Springer Verlag.

**Dynamic iterative method for fast network partitioning.** / Jeong, Chang-Sung; Song, Young Min; Jo, Sung Up.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).*vol. 1823, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1823, Springer Verlag, pp. 81-88, 8th International Conference on High Performance Computing and Networking, HPCNEurope 2000, Amsterdam, Netherlands, 00/5/8.

}

TY - GEN

T1 - Dynamic iterative method for fast network partitioning

AU - Jeong, Chang-Sung

AU - Song, Young Min

AU - Jo, Sung Up

PY - 2000

Y1 - 2000

N2 - In this paper, we address multiway network partitioning problem of dividing the cells of network into multiple blocks so as to minimize the number of nets interconnecting cells in different blocks while balancing the blocks' sizes. The sequential iterative improvement algorithm for the problem consists of several passes each of which is performed by repeatedly iterating the move operation. Therefore, the whole execution time taken by the algorithm is greatly affected by the number of the move operations and the execution time for each move operation. We present a fast parallel algorithm for solving the multiway network partiotioning problem by reducing both of them. We propose a new dynamic iterative method which reduces the number of move operations executed at each pass dynamically, and hence speed up the whole algorithm sharply. Moreover, we reduced the execution time of each move by its parallelization using the proper cell distribution.

AB - In this paper, we address multiway network partitioning problem of dividing the cells of network into multiple blocks so as to minimize the number of nets interconnecting cells in different blocks while balancing the blocks' sizes. The sequential iterative improvement algorithm for the problem consists of several passes each of which is performed by repeatedly iterating the move operation. Therefore, the whole execution time taken by the algorithm is greatly affected by the number of the move operations and the execution time for each move operation. We present a fast parallel algorithm for solving the multiway network partiotioning problem by reducing both of them. We propose a new dynamic iterative method which reduces the number of move operations executed at each pass dynamically, and hence speed up the whole algorithm sharply. Moreover, we reduced the execution time of each move by its parallelization using the proper cell distribution.

UR - http://www.scopus.com/inward/record.url?scp=84944048003&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944048003&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84944048003

SN - 9783540675532

VL - 1823

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 81

EP - 88

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -