Communication-sensitive loop scheduling for DSP applications

• Published in: IEEE transactions on signal processing Vol. 45; no. 5; pp. 1309 - 1322
• Main Authors: Tongsima, S., Sha, E.H.-M., Passos, N.L.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.1997; Institute of Electrical and Electronics Engineers
• Subjects: Applied sciences; Concurrent computing; Data communication; Data flow computing; Delay; Digital signal processing; Exact sciences and technology; Flow graphs; High performance computing; Operational research and scientific management; Operational research. Management science; Parallel architectures; Processor scheduling; Scheduling algorithm; Scheduling, sequencing; Multiprocessor; Parallel system; Resource allocation; Task scheduling; Signal processing; Digital processing; Scheduling; Algorithm
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/78.575702

The performance of computation-intensive digital signal processing applications running on parallel systems is highly dependent on communication delays imposed by the parallel architecture. In order to obtain a more compact task/processor assignment, a scheduling algorithm considering the communication time between processors needs to be investigated. Such applications usually contain iterative or recursive segments that are modeled as communication sensitive data flow graphs (CS-DFGs), where nodes represent computational tasks and edges represent dependencies between them. Based on the theorems derived, this paper presents a novel efficient technique called cyclo-compaction scheduling, which is applied to a CS-DFG to obtain a better schedule. This new method takes into account the data transmission time, loop carried dependencies, and the target architecture. It implicitly uses the retiming technique (loop pipelining) and a task remapping procedure to allocate processors and to iteratively improve the parallelism while handling the underlying communication and resource constraints. Experimental results on different architectures demonstrate that this algorithm yields significant improvement over existing methods. For some applications, the final schedule length is less than one half of its initial length.