Subcube Recognition, AllocatiodDeallocation and

This paper investigates four different strategies for subcube recognition, allocationldeallocation and relocation in an ndimensional hypercube multiprocessor: the buddy strategy which recognizes 2"+ k-cubes in 0 (N) time, N = 2", has a relocation cost of 0 (KlogK) process-moves where K = 2' is the size of subcube requested; the Gray code (GC) strategy which recognizes 20J+' k-cubes in O(N) time is shown to have a relocation cost of O(N) process-moves; the Recognizing All (RA) strategy can recognize all C$2"+ k-cubes in O(N'.58) time with a relocation cost of O(N) processmoves; and the Extended Buddy (EB) strategy which recognizes 2"-'+ (n-k)2"4-' k-cubes in 0 (NlogN) time is shown to have a relocation cost of 0 (KlogK) process-moves.

[1]  John C. Peterson,et al.  The Mark III Hypercube-Ensemble Concurrent Computer , 1985, International Conference on Parallel Processing.

[2]  John P. Hayes,et al.  Architecture of a Hypercube Supercomputer , 1986, ICPP.

[3]  Ming-Syan Chen,et al.  Processor Allocation in an N-Cube Multiprocessor Using Gray Codes , 1987, IEEE Transactions on Computers.

[4]  Charles L. Seitz,et al.  The cosmic cube , 1985, CACM.