Title :
Decoupling-capacitor allocation problem solved by genetic algorithm
Author :
Shibasaka, Kazuma ; Kanazawa, Kenji ; Yasunaga, Moritoshi
Author_Institution :
Grad. Sch. of Syst. & Inf. Eng., Univ. of Tsukuba, Tsukuba, Japan
Abstract :
Decoupling capacitors allocated closely to LSIs work well in order to increase PI (power integrity) in PDNs (power distribution networks). In the higher frequency range more than 5 GHz, however, we have to allocate the decoupling capacitors in the whole power-plane as the allocation target area instead of the vicinity of the LSI because of the power-voltage standing waves shorter than the power plane size. This allocation problem becomes a combinatorial explosion problem, which has no deterministic solving approach. We apply the GA to this problem, and also propose a novel chromosome suits to the power plain structure. In the simulation experiment, GA successfully obtains an effective solution that decrease the strongest impedance less than half of that in the conventional allocation.
Keywords :
capacitors; genetic algorithms; integrated circuit modelling; large scale integration; LSI; PDN; allocation target area; combinatorial explosion problem; decoupling capacitor allocation problem; genetic algorithm; power distribution network; power integrity; power plane; power voltage standing waves; Decision support systems; Packaging; Decoupling capacitor; genetic algorithm; power plane;
Conference_Titel :
Electrical Design of Advanced Packaging and Systems Symposium (EDAPS), 2013 IEEE
Conference_Location :
Nara
Print_ISBN :
978-1-4799-2313-7
DOI :
10.1109/EDAPS.2013.6724430