### Abstract

A parallel computing system becomes increasingly prone to failure as the number of processing elements in it increases. In this paper, we describe a completely general strategy that takes an arbitrary step of an ideal CRCW PRAM and automatically translates it to run efficiently and robustly on a PRAM in which processors are prone to failure. The strategy relies on efficient robust algorithms for solving a core problem, the Certified Write-All Problem. This problem characterizes the core of robustness, because, as we show, its complexity is equal to that of any general strategy for realizing robustness in the model. We analyze the expected parallel time and work of various algorithms for solving this problem. Our results are a non-trivial generalization of R.P. Brent's Lemma. We consider the case where the number of the available processors decreases dynamically over time, whereas Brent's Lemma is only applicable in the case where the processor availability pattern is static.

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

Title of host publication | Proc 22nd Annu ACM Symp Theory Comput |

Publisher | Publ by ACM |

Pages | 138-148 |

Number of pages | 11 |

ISBN (Print) | 0897913612 |

State | Published - 1990 |

Event | Proceedings of the 22nd Annual ACM Symposium on Theory of Computing - Baltimore, MD, USA Duration: May 14 1990 → May 16 1990 |

### Other

Other | Proceedings of the 22nd Annual ACM Symposium on Theory of Computing |
---|---|

City | Baltimore, MD, USA |

Period | 5/14/90 → 5/16/90 |

### Fingerprint

### ASJC Scopus subject areas

- Engineering(all)

### Cite this

*Proc 22nd Annu ACM Symp Theory Comput*(pp. 138-148). Publ by ACM.

**Efficient robust parallel computations.** / Kedem, Zvi; Palem, Krishna V.; Spirakis, Paul G.

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

*Proc 22nd Annu ACM Symp Theory Comput.*Publ by ACM, pp. 138-148, Proceedings of the 22nd Annual ACM Symposium on Theory of Computing, Baltimore, MD, USA, 5/14/90.

}

TY - GEN

T1 - Efficient robust parallel computations.

AU - Kedem, Zvi

AU - Palem, Krishna V.

AU - Spirakis, Paul G.

PY - 1990

Y1 - 1990

N2 - A parallel computing system becomes increasingly prone to failure as the number of processing elements in it increases. In this paper, we describe a completely general strategy that takes an arbitrary step of an ideal CRCW PRAM and automatically translates it to run efficiently and robustly on a PRAM in which processors are prone to failure. The strategy relies on efficient robust algorithms for solving a core problem, the Certified Write-All Problem. This problem characterizes the core of robustness, because, as we show, its complexity is equal to that of any general strategy for realizing robustness in the model. We analyze the expected parallel time and work of various algorithms for solving this problem. Our results are a non-trivial generalization of R.P. Brent's Lemma. We consider the case where the number of the available processors decreases dynamically over time, whereas Brent's Lemma is only applicable in the case where the processor availability pattern is static.

AB - A parallel computing system becomes increasingly prone to failure as the number of processing elements in it increases. In this paper, we describe a completely general strategy that takes an arbitrary step of an ideal CRCW PRAM and automatically translates it to run efficiently and robustly on a PRAM in which processors are prone to failure. The strategy relies on efficient robust algorithms for solving a core problem, the Certified Write-All Problem. This problem characterizes the core of robustness, because, as we show, its complexity is equal to that of any general strategy for realizing robustness in the model. We analyze the expected parallel time and work of various algorithms for solving this problem. Our results are a non-trivial generalization of R.P. Brent's Lemma. We consider the case where the number of the available processors decreases dynamically over time, whereas Brent's Lemma is only applicable in the case where the processor availability pattern is static.

UR - http://www.scopus.com/inward/record.url?scp=0025028518&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025028518&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0025028518

SN - 0897913612

SP - 138

EP - 148

BT - Proc 22nd Annu ACM Symp Theory Comput

PB - Publ by ACM

ER -