Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

NP-completeness of linearly-connected multiprocessor scheduling

Book ·
OSTI ID:5090885
This paper analyzes the computational complexity of scheduling for linearly-connected multiprocessor architectures. A program to be scheduled is specified by a directed acyclic graph (DAG), where each vertex of the DAG represents a unit time operation to be performed by one of a fixed number of processors, and each arc of the DAG represents a precedence constraint. Each vertex must be assigned a start time and a processor such that all operations are completed and the constraints imposed by the DAG and the architecture are observed. A scheduling algorithm is a decision procedure to determine whether a given number of time units is sufficient for completion of the operations specified by a given DAG. The authors first show how precedence-constrained scheduling and linearly-connected multiprocessor scheduling can be expressed as subgraph isomorphism problems. They then specify another type of linearly-connected multiprocessor scheduling problem in terms of subgraph isomorphism and prove that it is NP-complete. Finally, they use similar techniques to prove that scheduling for the Aspex Inc. PIPE machine is NP-complete. The NP-completeness results support the development of approximation algorithms and heuristic methods for linearly-connected multiprocessor scheduling.
OSTI ID:
5090885
Country of Publication:
United States
Language:
English

Similar Records

Generalized multiprocessor scheduling for directed acylic graphs
Conference · Fri Dec 30 23:00:00 EST 1994 · OSTI ID:87627

Scheduling algorithms for PIPE (Pipelined Image-Processing Engine)
Journal Article · Thu Mar 31 23:00:00 EST 1988 · J. Parallel Distrib. Comput.; (United States) · OSTI ID:6390162

Performance bounds on multiprocessor schedules
Conference · Sat Dec 31 23:00:00 EST 1983 · OSTI ID:5082359