Title :
Instruction scheduling using genetic algorithm with taboo search for TTA-like processors
Author :
Zhong, Songyan ; Wei, Jizeng ; Guo, Wei ; Wang, Zhenghua
Author_Institution :
Sch. of Comput. Sci. & Technol., Tianjin Univ., Tianjin, China
Abstract :
This paper presents an efficient instruction scheduling method for configurable processors based on Transport Triggered Architecture (TTA). The genetic algorithm is used to realize the instruction scheduling while the taboo search is integrated to prevent from getting into the locally optimal solution. Experiment on a serial of DSP applications shows that the proposed method can increase instruction-level parallelism (ILP) about 8.3% than the genetic algorithm alone, and 12.0% than the list scheduling method.
Keywords :
genetic algorithms; instruction sets; parallel architectures; processor scheduling; search problems; genetic algorithm; instruction scheduling; instruction-level parallelism; processor scheduling; taboo search; transport triggered architecture; Algorithm design and analysis; Application specific processors; Computer aided instruction; Costs; Genetic algorithms; Job shop scheduling; Processor scheduling; Radio frequency; Registers; Scheduling algorithm; Transport triggered architecture; component; genetic algorithm; instruction scheduling; taboo search;
Conference_Titel :
Computer Design and Applications (ICCDA), 2010 International Conference on
Conference_Location :
Qinhuangdao
Print_ISBN :
978-1-4244-7164-5
Electronic_ISBN :
978-1-4244-7164-5
DOI :
10.1109/ICCDA.2010.5541389
