Results 1 
2 of
2
Deterministic Computations on a PRAM with Static Faults
"... We develop a deterministic simulation of fully operational Parallel Random Access Machine (PRAM) on a PRAM with some faulty processors and memory cells. The faults considered are static, i.e., once the machine starts to operate, the operational/faulty status of PRAM components does not change. The s ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
We develop a deterministic simulation of fully operational Parallel Random Access Machine (PRAM) on a PRAM with some faulty processors and memory cells. The faults considered are static, i.e., once the machine starts to operate, the operational/faulty status of PRAM components does not change. The simulating machine can tolerate a constant fraction of faults among processors and memory cells. The simulating PRAM has n processors and m memory cells, and simulates a PRAM with n processors and m) memory cells. The simulation is in three phases: (1) preprocessing, followed by (2) retrieving the input by the processors active in the simulation, followed by (3) the proper part of the simulation performed in a stepbystep fashion. Preprocessing is performed in time O(( m n + log n) log n). The input is retrieved in time O(log 2 n). The slowdown of the proper part of the simulation is O(log m).
Fullyscalable faulttolerant simulations for BSP and CGM
 in: Proceedings of the 13th International Parallel Processing Symposium & 10th Symposium on Parallel Distributed Processing
, 1999
"... In this paper we consider general simulations of algorithms designed for fully operational BSP and CGM machines on machines with faulty processors. The faults are deterministic (i.e., worstcase distributions of faults are considered) and static (i.e., they do not change in the course of computati ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
In this paper we consider general simulations of algorithms designed for fully operational BSP and CGM machines on machines with faulty processors. The faults are deterministic (i.e., worstcase distributions of faults are considered) and static (i.e., they do not change in the course of computation). We assume that a constant fraction of processors are faulty. We present a deterministic simulation (resp. a randomized simulation) that achieves constant slowdown per local computations and O((log h p)