Title :
ILP formulation for data path scheduling problem with conditional branches
Author :
Tsuyuno, Kimitake ; Yamada, Shoichiro ; Fujii, Tomoo
Author_Institution :
Dept. of Electr. Eng., Osaka City Univ., Japan
Abstract :
In this paper, we present a new time constraint scheduling method for optimizing data paths with conditional branches. In our problem, an operation in the true path for a conditional branch and the other operation in the false path are mutually exclusive, if the branch is activated before both of the operations will be executed, and operations can share a functional unit. The scheduling problem considering mutual exclusiveness is formulated in an integer linear programming (ILP). An optimal scheduling is obtained quickly by an iterating algorithm based on the ILP formulation
Keywords :
constraint theory; integer programming; linear programming; processor scheduling; ILP; conditional branches; data path scheduling; integer linear programming; iterating algorithm; mutual exclusiveness; optimization; time constraints; Constraint optimization; Cost function; Data engineering; Educational institutions; Hardware; Integer linear programming; Optimal scheduling; Processor scheduling; Scheduling algorithm; Time factors;
Conference_Titel :
VLSI Signal Processing, VIII, 1995. IEEE Signal Processing Society [Workshop on]
Conference_Location :
Sakai
Print_ISBN :
0-7803-2612-1
DOI :
10.1109/VLSISP.1995.527496