Dimension Ordering and Broadcast Algorithms in Faulty SIMD Hypercubes

Improved one-to-all broadcasting algorithms on faulty SIMD hypercubes

“…The sequence of dimensions used for broadcasting in each algorithm is the same regardless of which node is the source. The proposed one-to-all broadcasting algorithms can tolerate n/2 more faulty nodes than Raghavendra and Sridhar's algorithms [8] although 8 extra steps are needed. The fault-tolerance improvement of this paper is about 50%.…”

“…The fault model used in this paper follows the fault model [8]. In the fault model, it assumes that the functions for computation and communication in the faulty node are all lost, where such faulty nodes are called truly faulty (TF) nodes.…”

“…In an SIMD hypercube, all nodes can exchange messages with their neighbors along a specific dimension in each step. Specifically, one advantage is that the sequence of dimensions used for broadcasting in each of Raghavendra and Sridhar's algorithms [8] is the same regardless of which node is the source. In addition, they also claim two advantages of their broadcasting algorithms: programs are independent of process ID and very simple to develop.…”

“…With n+3 log n, n+2 log n+2, n+log n+O(log log n), n + log n + 5, and n + 12 steps, Raghavendra and Sridhar [8] presented five broadcasting algorithms on an * Fax: +886-2-26205236.…”

“…Considering faulty hypercubes, it is an important issue to design fault-tolerant broadcasting algorithms. Many fault-tolerant broadcasting algorithms [1,3,6,[8][9][10]12,13], most of which are implemented on MIMD hypercubes, have been presented.…”