Breuer Algorithm

Breuer Algorithm - A CLASS OF MI N CUT PLACEMENT ALGORI...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: A CLASS OF MI N- CUT PLACEMENT ALGORI THMS % by Mel vi n A. Br euer Depar t ment s of El ect r i cal Engi neer i ng g~ Comput er Sci ence Uni ver s i t y of Sout her n Cal i f or ni a Los Angel es , Cal i f or ni a 90007 Summa r y In t hi s paper we pr es ent a cl as s of r ai n- cut pl ace- ment al gor i t hms for s ol vi ng s ome as s i gnment pr ob- l ems r el at ed to the phys i cal i mpl ement at i on of el ec- t r i cal ci r cui t s . We di s cus s t he need f or abandoni ng cl as s i cal obj ect i ve f unct i ons bas ed upon di s t ance, and i nt r oduce new obj ect i ve f unct i ons bas ed upon "s i gnal s - cut. " The number of s i gnal s cut by a line c is a l ower bound on the number of r out i ng t r acks whi ch mus t cr os s c in r out i ng t he ci r cui t . Thr ee s peci f i c obj ec- t i ve f unct i ons ar e i nt r oduced and the r el at i ons hi p be- t ween one of t hes e and a cl as s i cal di s t ance meas ur e bas ed upon hal f - per i met er is pr es ent ed. Two r ai n- cut pl acement al gor i t hms ar e pr es ent ed. They ar e r ef er r ed to as C) uadr at ur e and Sl i ce/ Bi s ec- tion. The concept s of a bl ock and cut line ar e i nt r o- duced. Thes e two ent i t i es ar e the maj or cons t r uct s in devel opi ng any new mi n- cut pl acement al gor i t hm. Mos t of t he concept s pr es ent ed have been i mpl e- ment ed, and s ome exper i ment al r es ul t s ar e gi ven. I. I nt r oduct i on Thi s paper deal s wi t h a cl as s i cal pr obl em encoun- t er ed in the phys i cal i mpl ement at i on of ci r cui t car ds or chi ps , r ef er r ed to as the pl acement pr obl em. The pr obl em is def i ned, s emi f or mal l y, as f ol l ows . Gi v- en a s et of el ement s 8 = {el, e~ . . . . . e~} and a s et of s i gnal s ,2= [ s u sm . . . . . s , }. We as s oci at e wi t h each el ement e E ~ a s et of s i gnal s ,2e, wher e J~ = J. Si mi - l ar l y wi t h each s i gnal s E • we as s oci at e a s et of el e- ment s e,, where $ ={else ,/,}. d~, is said to be a signal net. We are also given a set of slots or loca- tions ~=[LI, L m ..... L ], where pgn. The place- ment problem is to assign each e l~ ~ to a unique lo- cation Lj such that some objective is optimized. Nor- mally each element is considered to be a point, and if e t is assigned to location L~ then its position is de- fined by the coordinate values (x],y~). Usually a sub- set of the elements in d~ are fixed, i.e., pre-assigned to locations, and only the remaining elements can be assigned to the remaining unassigned locations. Non fixed elements are called moveable elements, and those slots not pre-assigned elements are called open slots. Example placement problems deal with polycell LSI chip design [2,4,5], PC card design, and assign- ing functions to previously placed IC's....
View Full Document

This note was uploaded on 11/21/2009 for the course ESE 556 taught by Professor Doboli,a during the Fall '08 term at SUNY Stony Brook.

Page1 / 7

Breuer Algorithm - A CLASS OF MI N CUT PLACEMENT ALGORI...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online