Asymptotically tight bounds for computing with faulty arrays of processors

C. Kaklamanis, A. R. Karlin, F. T. Leighton, Victor Milenkovic, P. Raghavan, S. Rao, C. Thomborson, A. Tsantilas

Research output: Contribution to journalArticle

43 Citations (Scopus)

Abstract

The computational power of 2-D and 3-D processor arrays that contain a potentially large number of faults is analyzed. Both a random and a worst-case fault model are considered, and it is proved that in either scenario low-dimensional arrays are surprisingly fault tolerant. It is also shown how to route, sort, and perform systolic algorithms for problems such as matrix multiplication in optimal time on faulty arrays. In many cases, the running time is the same as if there were no faults in the array (up to constant factors). On the negative side, it is shown that any constant congestion embedding of an n × n fault-free array on an n × n array with Θ(n2) random faults (or Θ(log n) worst-case faults) requires dilation Θ(log n). For 3-D arrays, knot theory is used to prove that the required dilation is Ω(√log n).

Original languageEnglish (US)
Pages (from-to)285-296
Number of pages12
JournalIEEE Transactions on Industry Applications
Volume27
Issue number1 pt 1
StatePublished - Jan 1991
Externally publishedYes

Fingerprint

Systolic arrays
Parallel processing systems

ASJC Scopus subject areas

  • Engineering (miscellaneous)
  • Electrical and Electronic Engineering

Cite this

Kaklamanis, C., Karlin, A. R., Leighton, F. T., Milenkovic, V., Raghavan, P., Rao, S., ... Tsantilas, A. (1991). Asymptotically tight bounds for computing with faulty arrays of processors. IEEE Transactions on Industry Applications, 27(1 pt 1), 285-296.

Asymptotically tight bounds for computing with faulty arrays of processors. / Kaklamanis, C.; Karlin, A. R.; Leighton, F. T.; Milenkovic, Victor; Raghavan, P.; Rao, S.; Thomborson, C.; Tsantilas, A.

In: IEEE Transactions on Industry Applications, Vol. 27, No. 1 pt 1, 01.1991, p. 285-296.

Research output: Contribution to journalArticle

Kaklamanis, C, Karlin, AR, Leighton, FT, Milenkovic, V, Raghavan, P, Rao, S, Thomborson, C & Tsantilas, A 1991, 'Asymptotically tight bounds for computing with faulty arrays of processors', IEEE Transactions on Industry Applications, vol. 27, no. 1 pt 1, pp. 285-296.
Kaklamanis C, Karlin AR, Leighton FT, Milenkovic V, Raghavan P, Rao S et al. Asymptotically tight bounds for computing with faulty arrays of processors. IEEE Transactions on Industry Applications. 1991 Jan;27(1 pt 1):285-296.
Kaklamanis, C. ; Karlin, A. R. ; Leighton, F. T. ; Milenkovic, Victor ; Raghavan, P. ; Rao, S. ; Thomborson, C. ; Tsantilas, A. / Asymptotically tight bounds for computing with faulty arrays of processors. In: IEEE Transactions on Industry Applications. 1991 ; Vol. 27, No. 1 pt 1. pp. 285-296.
@article{8b5fd86c6d1b4941958315c649f74159,
title = "Asymptotically tight bounds for computing with faulty arrays of processors",
abstract = "The computational power of 2-D and 3-D processor arrays that contain a potentially large number of faults is analyzed. Both a random and a worst-case fault model are considered, and it is proved that in either scenario low-dimensional arrays are surprisingly fault tolerant. It is also shown how to route, sort, and perform systolic algorithms for problems such as matrix multiplication in optimal time on faulty arrays. In many cases, the running time is the same as if there were no faults in the array (up to constant factors). On the negative side, it is shown that any constant congestion embedding of an n × n fault-free array on an n × n array with Θ(n2) random faults (or Θ(log n) worst-case faults) requires dilation Θ(log n). For 3-D arrays, knot theory is used to prove that the required dilation is Ω(√log n).",
author = "C. Kaklamanis and Karlin, {A. R.} and Leighton, {F. T.} and Victor Milenkovic and P. Raghavan and S. Rao and C. Thomborson and A. Tsantilas",
year = "1991",
month = "1",
language = "English (US)",
volume = "27",
pages = "285--296",
journal = "IEEE Transactions on Industry Applications",
issn = "0093-9994",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1 pt 1",

}

TY - JOUR

T1 - Asymptotically tight bounds for computing with faulty arrays of processors

AU - Kaklamanis, C.

AU - Karlin, A. R.

AU - Leighton, F. T.

AU - Milenkovic, Victor

AU - Raghavan, P.

AU - Rao, S.

AU - Thomborson, C.

AU - Tsantilas, A.

PY - 1991/1

Y1 - 1991/1

N2 - The computational power of 2-D and 3-D processor arrays that contain a potentially large number of faults is analyzed. Both a random and a worst-case fault model are considered, and it is proved that in either scenario low-dimensional arrays are surprisingly fault tolerant. It is also shown how to route, sort, and perform systolic algorithms for problems such as matrix multiplication in optimal time on faulty arrays. In many cases, the running time is the same as if there were no faults in the array (up to constant factors). On the negative side, it is shown that any constant congestion embedding of an n × n fault-free array on an n × n array with Θ(n2) random faults (or Θ(log n) worst-case faults) requires dilation Θ(log n). For 3-D arrays, knot theory is used to prove that the required dilation is Ω(√log n).

AB - The computational power of 2-D and 3-D processor arrays that contain a potentially large number of faults is analyzed. Both a random and a worst-case fault model are considered, and it is proved that in either scenario low-dimensional arrays are surprisingly fault tolerant. It is also shown how to route, sort, and perform systolic algorithms for problems such as matrix multiplication in optimal time on faulty arrays. In many cases, the running time is the same as if there were no faults in the array (up to constant factors). On the negative side, it is shown that any constant congestion embedding of an n × n fault-free array on an n × n array with Θ(n2) random faults (or Θ(log n) worst-case faults) requires dilation Θ(log n). For 3-D arrays, knot theory is used to prove that the required dilation is Ω(√log n).

UR - http://www.scopus.com/inward/record.url?scp=0025742359&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025742359&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0025742359

VL - 27

SP - 285

EP - 296

JO - IEEE Transactions on Industry Applications

JF - IEEE Transactions on Industry Applications

SN - 0093-9994

IS - 1 pt 1

ER -