Embeddings, communication and performance of algorithms in faulty hypercubes
In this research, embedding and global data communication schemes are developed for faulty hypercubes and studied in the context of algorithms. Algorithm LNRCUBE finds a maximum fault-free subcube first, then embeds a linear array in the maximum fault-free subcube. Algorithm LNRLATIN finds a Gray code sequence in a faulty hypercube by making use of a modification of the Latin Multiplication. Algorithm PLNRLATIN is a parallel version of Algorithm LNRLATIN. Algorithm LNRCSQ embeds a linear array using a modification of Coordinate sequence approach. Algorithm PLNRCSQ is a parallel version of Algorithm LNRCSQ. Algorithm LNRDCC embeds a linear array in a faulty hypercube that has Direct-Connect Capability. Algorithms of embedding 2-D meshes in a faulty hypercube are developed. Finally, the embedding and data communication schemes developed are applied to the parallel Conjugate Gradient algorithm and the parallel sorting algorithm.
- Research Organization:
- Ohio State Univ., Columbus, OH (United States)
- OSTI ID:
- 6090965
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallel-sorting algorithms for hypercube multiprocessors
Generalized gray codes and embedding in hypercubes