### Abstract

We study the Distributed Memory Machine (DMM) with faults in memory. The DMM consists of n synchronized processors together with n memory units (MUs). A MU can be accessed by at most one processor at a time. The total number of memory faults is assumed to be at most a fixed fraction of the total number of words. We develop two fast randomized simulations of the PRAM on such a faulty DMM. A simulation consists of two phases: the preprocessing is followed by the simulation proper done in a step-by-step fashion. One simulation is of an n log n-processor PRAM and it operates with the optimal expected slowdown O(log n), the other is of a PRAM with n/log n processors and has the slowdown O(log log n).

Original language | English (US) |
---|---|

Title of host publication | Automata, Languages and Programming - 23rd International Colloquium, ICALP 1996, Proceedings |

Editors | Friedhelm Meyer auf der Heide, Burkhard Monien |

Publisher | Springer Verlag |

Pages | 586-597 |

Number of pages | 12 |

ISBN (Print) | 3540614400, 9783540614401 |

State | Published - Jan 1 1996 |

Externally published | Yes |

Event | 23rd International Colloquium on Automata, Languages, and Programming, ICALP 1996 - Paderborn, Germany Duration: Jul 8 1996 → Jul 12 1996 |

### Publication series

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

Volume | 1099 |

ISSN (Print) | 0302-9743 |

ISSN (Electronic) | 1611-3349 |

### Conference

Conference | 23rd International Colloquium on Automata, Languages, and Programming, ICALP 1996 |
---|---|

Country | Germany |

City | Paderborn |

Period | 7/8/96 → 7/12/96 |

### Fingerprint

### ASJC Scopus subject areas

- Theoretical Computer Science
- Computer Science(all)

### Cite this

*Automata, Languages and Programming - 23rd International Colloquium, ICALP 1996, Proceedings*(pp. 586-597). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1099). Springer Verlag.

**Shared-memory simulations on a faulty-memory DMM.** / Chlebus, Bogdan S.; Gambin, Anna; Indyk, Piotr.

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

*Automata, Languages and Programming - 23rd International Colloquium, ICALP 1996, Proceedings.*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1099, Springer Verlag, pp. 586-597, 23rd International Colloquium on Automata, Languages, and Programming, ICALP 1996, Paderborn, Germany, 7/8/96.

}

TY - GEN

T1 - Shared-memory simulations on a faulty-memory DMM

AU - Chlebus, Bogdan S.

AU - Gambin, Anna

AU - Indyk, Piotr

PY - 1996/1/1

Y1 - 1996/1/1

N2 - We study the Distributed Memory Machine (DMM) with faults in memory. The DMM consists of n synchronized processors together with n memory units (MUs). A MU can be accessed by at most one processor at a time. The total number of memory faults is assumed to be at most a fixed fraction of the total number of words. We develop two fast randomized simulations of the PRAM on such a faulty DMM. A simulation consists of two phases: the preprocessing is followed by the simulation proper done in a step-by-step fashion. One simulation is of an n log n-processor PRAM and it operates with the optimal expected slowdown O(log n), the other is of a PRAM with n/log n processors and has the slowdown O(log log n).

AB - We study the Distributed Memory Machine (DMM) with faults in memory. The DMM consists of n synchronized processors together with n memory units (MUs). A MU can be accessed by at most one processor at a time. The total number of memory faults is assumed to be at most a fixed fraction of the total number of words. We develop two fast randomized simulations of the PRAM on such a faulty DMM. A simulation consists of two phases: the preprocessing is followed by the simulation proper done in a step-by-step fashion. One simulation is of an n log n-processor PRAM and it operates with the optimal expected slowdown O(log n), the other is of a PRAM with n/log n processors and has the slowdown O(log log n).

UR - http://www.scopus.com/inward/record.url?scp=72849110811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=72849110811&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:72849110811

SN - 3540614400

SN - 9783540614401

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

SP - 586

EP - 597

BT - Automata, Languages and Programming - 23rd International Colloquium, ICALP 1996, Proceedings

A2 - Meyer auf der Heide, Friedhelm

A2 - Monien, Burkhard

PB - Springer Verlag

ER -