Title :
Transforming arbitrary data flow graphs to synchronous data flow graphs
Author :
Lin, Homg-Dar ; Messerschmit, David G.
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., California Univ., Berkeley, CA, USA
Abstract :
In many practical algorithms data dependencies are necessary, and the question arises as to how to avoid excessive runtime overhead. The authors propose transformations to data-independent graphs for the case where the number of tokens consumed and generated is data-dependent. The proposed approach is consistent with packet communication in multiprocessor systems, and allows static scheduling of a class of algorithms with data dependencies
Keywords :
computerised signal processing; graph theory; packet switching; parallel algorithms; arbitrary data flow graphs; data dependencies; excessive runtime overhead; multiprocessor systems; packet communication; static scheduling; synchronous data flow graphs; tokens; Computer displays; Data flow computing; Delay; Dynamic scheduling; Flow graphs; Processor scheduling; Runtime; Scheduling algorithm; Signal processing algorithms; Switches;
Conference_Titel :
Circuits and Systems, 1989., IEEE International Symposium on
Conference_Location :
Portland, OR
DOI :
10.1109/ISCAS.1989.100355
