Scribd
Upload
13 views

Physical Design Part - 1

VLSI Physical design

Uploaded by

adarshnani.1089
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
13 views

Physical Design Part - 1

VLSI Physical design

Uploaded by

adarshnani.1089
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 52
i Qiypiicad phiign thc Ss eta are : agn Hicance af PRI- . : " > Pt technology evolving , a, Sirqle chip. nevenutng : ree oe Bop TO Matt Our peGuaqerney + Pesfoumant. _eplreigarian (bau 6A daxpours , signe " pansutoy 00 nls ! is placement 1 ROUTAN can effect peopagarion : Astauys 5. power ee ges ce : oe oS : 1 ogi ¢ Se oeia _formicatd ena wacetet)., 8s s +. Manufactu yet 47 signok fr ogi nee 5 fee a Be seo NIST Resign * “styler - on be ely ey Feutt Qutom Paign Oyle F _ Fever constraints dusing tayout “generation = Becks ON be Plaid “anywhere onthe Che ~ Wayout — are Axa.00 * ppranuall =o coniteaint tn Sige Of tte ale1igM —No plactmenr comiaint . ~ Viet a very compact cnip conrts — reyhheg Oplimired etectvical propesttits * ph erro prone. - laborious , -tme ~ conuming 4 ~ Ulefut fos rotcroproeitoras , TP bf) © eh Blocks L ome Gt the Parts Nike 2 L, APC, PLL 4 i Ks ise fall Guione . ertircat path 4 fyicubs : "3 dara path - cLatg 0 eal), FEaNA — = @suiteh cotati @ Lut Hull, CLevtD a4) 6 ete Ee A fel dy free > Cools are Roane I aows, with Fatereopne cttory Fo-the channeb bho APOVOS + . -y conne ctfons > BIO -GkL can ge “Uyrugh channels 4 oy oves tb Y thayhe hot th “acy ous © Cheedtlracugh cel) ; : simples than utl-curtorn datgn ¢ oct ts well with mecewn ain: tools: > Oceupie move Chip. ove than full caustom | deign bur as geting. clover, WEP. advance ment fia technology + — Placro coh 5 f ened : — lor get. dogic component h wt st ‘ 2 “etandard i 25 can. oveT conurrainte cl C Atke _ manolotory a plang enue y porialitT ) TARGING tore mi 4 wu) ~e Rewabl ancl? to nighly rae P © tere uy Daed = @ pre -dittued(6atr aw 5 @ Gote tray: —* Strmpltties — chip desqn by euing sao el on qates throughout the Chip: : Me Chakt fs diviced Fnto edanttcat’ blocky, S Pach ever to & ‘ek Of the gaa On “et : > Blocks ove rropped yy laud Bee pre-Piar. — cate atl oe perttening ; Bes = are eftabltthed wing AO Zo & Vexttcal’ Channels: oe) ep teiiten Bote Sirags: S Conpersine } AGoeel ved Use Circuit Simusation tools yike SPL pa veaity Corgecthess a => Oonsices factors Ite traniutoy SH ¢ pond B manurtaccutiney ventatelity. UN, Col Devas: Caspur 3 Taitantian design comperanits wth. gre : deprtientaions => pssign gpatiol Locattons cplocrnents te Routing connectfony (outing) to ares | | | dhe to dvign ques ayperencing ” _ Fainication Snirraint 4 Coniiola ‘the ape en” se 7008 0) riektobsth ee at Rae 5 rade ca check 11 potenital Antenna effects & eoeey pours (Grund — connections. » erie inimal datyout _Ghange ak ees arage ' H Fommcarten® ler wafer) pre ~chan layout tr went wa foundory fy pmiuclactusing (ape oul) ; : a -3 le pholamak fo define eae Op ae ae Goromo nly ued — coceea she (eq. nob? ee ‘ 4 7) ; . # Pee g ‘Testing! a choaing packaging yp such al Pr anid pera , Galt Grid ‘nay. base on appiicekton es POU Tey = Dacie patnaging meted ie oF Goleta bumps f ane in fea ‘Dertgn gq We RC - Qeign) Be Ail jeg ees “Berg | ‘feeh wt ; Les = Beitah|. ect (Syren Se Thete x age j dy thes saa - ee ae Clow f- : 2) 3+ hae a poati a tiitoninc Keats the ORE TAO [em eee Te Salar rnoaul one Ng |-saetrrs nrochuts shapers orrvourig mens o MuutA proniked Skuyo pg delay {ou - Gecpaiaererd ao [piss €-| Clock wee 5 aga TL poclute | Graing Shave: bufering, il & geting’ [eqn oui — + power? g ground g Aoutng toe qouting dtitrbuly a aecigne LArotngy iq Clasuae] pouns & grdund, 9 = fo veduy Stee ( : +Qtobal wouttng, : copttmiges, aoc ate! acto | ay conneettory~ Cxt performantr 2 Cramp AAOUTy Hough plore ss oe Fout try ie _ fnclualting aout RElign Rute Cheer COec)?- i—» Geomelrrc @niparnts on ae We we auccegsfia faint cotton: : = Mut “he followed — (0, ayia ant manufacturing defects 2 Minimum | aivoudatsl Poni elemenstons 3 Ensuse layout data corpltey + nay ‘he : juts es »! bea pe pa “@ Iie ps le L udettes ry SRO fosgourm ries by the deitgney soatcbe) re “aeltna by the: aes : Pe Minimum Chcloserse JOVEN) - a? 7 . opp rtkad ¥ ah BS —o checks that antitors ip a fayout ane connected . 10 the same | (oo ay to the CKr AChsmasie- a ; —> Compote exhacted netlist For the (aja xo the “onigtnal cehamatu KeH tut Lospout BAS & Schumovec nels, Grape Goons * Kevice cctsraself Soe) 2) eipaction GOs it Bs qolderm geferente: a SAMA Eleental Ra Chuck CERe):- owes —NeTRY te“ congaines ob, pemer & gd @onectfont 5 : # ARTY thot signal tanition tre: (slo eapactitve ivads , % Fanouls Orie prey hone : Pietra | a check Cae): . CorgegLeneay OF antenna effects- iy >? Qate Leakage - 1) _ change the threrhold wottege — Rduey the Hk pape of the rant Rech lenge? wre acctumulak more charge & ae 5 moe Aral: to darnog & the gates: Long routar WIT ‘a { : Brac -Algoiiibms:: of taut to Solve | ni pracy 1 Sroite Set Procter, an dip whhid a fie them Fear “ Seute ‘Ring Pigoh~" + Greaty “hls “Kinet so} of ae which & Jotaty oplimun) - lero Cie Jocol Ss ae Ay ou — > Divide ¢ Conguea Algona i ‘py divids tro fubpobleros » 4uth the “they should be As jou fo notuye - Go overlap bre — Pynromic pregram ie Algo ts mean problem . Pato sub probumt . overtop oho Rub pyopaU: Ce “ae wd — Mrvat / Integer pro thy Mme g Spau Conymeatte ? “ya a pete eee Evaluation Cby cong q — alge Seucetton foal sie = Apo ere ( by compoairg. a nie by eh Sie i optimgation : e 7 wR : ; gunding ens Ot ans * Algortins a Repercu oP the “type Of | lp. ‘ myachsiney prog lan Find unique athe tae: we Can find ol the algo’ wnith not ck RU VWnu all & oy sate oe OG nikey) “doy 7=0 ren de Sint) ea: Rede os O(nlogn) "ek gnepantie —» ln?) > an®logn+to — oh? leg 2) Rayreptottc notations: Par eu ud “the aucrorithnss ot ame oe 3 aps ee py The — > won't dopends on Mmathiny 06 Oleg oh) “ Notation ; i bout O- Netoton Cha ga ckeviZes eo Lipper on the eu ptour behaviour a} dune. © + $(n) = O(qr) a) there cats He a oe Ap Auch thee Os ty) 2 eg) $n 246N4100N $20 Fi een iS eae Rowton - notartOn Chacracte ri p7e4 @ Howes ba ana on the ayy mptore beh QvtOr Rh ee _a(gn) no of @ TUnction fir) 3 there; eat ye. Cone Cae h 1D 15) o< egtn) € Hiv) ASN. B> Sine $n? 4 6nr* fe 6n 4109P #20 rae , qo) = +n) = agen) Ce No oa | 30 > FI £4851 S Pe: spe46n > 90n +207 co” s nequsly Wolds when » MoD! Cevt Si) aigeo)- 2007 & emea) _noraner) ~ a taht bound op the Ga —func- : cy) eg ge) O- notation chosen ge mptot? Lehontou 6 (qtr) shy FO). = 00" . » Jo) = sete fg) Crp Cay no 7 wee comer F ae CiCa 118 5 ; > ro : [oe on4i24 Gapy me cirlaihereal: s NLOS po. > Any provide a Versari powerful 40 } ~ Aeprerenting, canalysieg g optintning oO Application oO} Gpraphy 8h pos _aflextist esentotion ! Graphs reepoerent Connecti birvo, elecrohic = comperents fh a Vist cr Flom Hip Planning?- Graphs depict elattonships: : _Consrsaints betuxen block ee Routing : Graphs help ee algontbrny © Optima: wire - = connections r + Cartical Poth Rapaliyty + 2 Gnaph- based . agit herp gdentity the Jongeit (Catteal ore eA OSE Ca eet Regrecotarie? of ee “the Ange ore ye fabs cde VG ‘ e =f cow), Coie, ody © lcae), teal (a (a9), Ch HE _ Hypeaguaph: Céncists Of nedes % hyperedges, with cach — hyperedge ding a cubsek OF 5 ot rooIe ‘node, | These ave ured 10. | cepnetent mutt pin) nett 7, mu tAPolnk : Connections : j ta ~ “Rees ‘of he tfodet~ Mow fs, fnetdunt eclge — path: © Poth between @ hodes a. air endeved “unedire cleat ce Cirwctel G-4ph | Qycvic Graph: ~ 4 dtaected ae hag. at hash ore. divecrd oycle - etherioie PL i an Rogerio’ r0ph. +? Vr: ‘design sdata wed Rp Several ERD aly AS. ‘eeu en rected royctre Gre ~ cyete each 4 = Complete. eu ap & a | giaps “ conneds each pede me exert Fo Connected graph “Trees A geeph wits pt a ‘ Sper edges Lapctree does net Contato any bi se oe ieee ge a | Madital ocyerie grr hs vobich mean thar “adding an edge. bre a node that ‘ane | inot coprecttd- We creatts: O ‘ cy cles i — Each ae = nodes ff Or by eqoctty ” eclge pe f ae ud (poo type of ares Tse UNE ah seach ature pt 5 al Oe wi gong upd 1 a at a Rea a ea ju50 5 Stondond et oe} epiesenting. a. graph’ ne) Chatenty Ltete 4ox 6 parse Graphs ~ fox Leni ete fs rouch _ less tha Wl" ot aidlfatsry ust nepreseoation es prefered an thay provide eprepact cep POL acne: Manices: for dense sens = UAEl te close to ie wl? adjauend epres sentarton °S piefenred Pagonerty ~ ~\tst Mepreseptarien of a graph ‘ B te) gneiss CF an onway ndy oe Avr each ‘Neste. 19 Ve \V| ItstS, one. d tet nay (ul < fx cach yey, adjouinny | _tondains a the yerlias ¥. dueh - that these § can edge tury), Be. % ob PY ; eK ai os (WE) be o weighted graph cotth voetght ~funetion Wo ae weight wU/¥) of the verter V edge Curvy) ee be Stored with tr v's ady Hit ot 1 Hi thease: 5 Jet, ? > 1° 6 = (WE) BO | weighed graph | with edge -weight function: to 4 She woeight weu Vv) “ot the. edge (uv) € con be stored gy he’ enmy) WP ap Uh | Column y of: ae adjacency marr E Gb NE en i ae ERY ene oe Re de eae o = oprimizg. | the ortrangement at Mow ct eer in 207 4 dfrunion. “prranious pices one -aesined to Neppeient ae na and colj atunetes: i = saps guloliet too vat ‘et IMpe 23) Bet ed) Mectanglss : : Gnaphi qelattcl to @_ Set es: ines - a type of MING hie ele a aly [za = Not asrgned & \iqued to the on. 4 9 Mntewval : PA. TokeMal 7% nepsaentia Jit aight end poiny dopoted by 4 4 ae e ag ona to ia ey = hy a5 ‘Ve presents Erne ay ag : 1 cw) | apeay cucn). ‘=o edge 31 detind penseén VF 6 vy atk Mterval 45 overlaps with 14. aug doa not completely Contasn: of reas ' 8thin 4y- SI a eb edge bE D reedge ° oN aes mnrewal Grae (v€2) ey Acdirucl a ; i Gr = Sots { ve [Ve Oepienb inkeavad ayy Ee Eyes Ue an «dye sto MW Ny 4 preewal Jy by © with 4 Prem ecto? GeO se mply E00 €& permution Graph: Matehing diagram: A diagnam roheve a hee lint, bea Aro dujiqratid: cepa G end oF another Specttied y-cooidinats * sey coue. typically —occus channe) FOLTINg)- oe ee as bo = (wee). ie a Ne Ave} sa aepiesenh anrewval’ 7 4 ; piateisech Mn Tae 1G, ¥) \ 4 am ee. Roe loc x e 3 Permutetiog Gin th os epee a ae 5 4 7 og a Jaryout wna —- Ghe RR neighbsshood — pe ued ate Global v routing Pee of Ae auronnttoo : —Cach chanmt fr depicted a a rectang channels ade pod netghbo us to). | Graph ‘Searching Nlgowithmas es 5 me Ws) BES Ped Suyse mati cally follow ~the edge o} the DAPh to vecse the Ventrces. Of the gr ee Wed “lo. Atstoves’ pe Shutters of “Whe gig 7? Wear are t BES + DES a popular algos fos grey | ‘ t SOCH Depth Frost Seavch Mr Obeaw “a- dtack to Keop track of Vertfe to Vesity em? 2: Choole O etodting vented ¢ push tt ‘onto “th Stack: : 4 phTu ‘the otack 7% nol empty , do the j \ following : a: Check ff the Vetter on the top of Stack has any unvisiied adjacent ve bp Sf Yee piuh an unvisi td adjacent onto the Stack ee o. SJ no UNVIed Adjacrt vevtrey i soy the cuwenr Nestea, ‘Pop:t Rua QES:- Dlgostioms ACV 6) — RFS(A) vor each verter ucgu paint alle Veattcy —~Urcotour = WoT Te } wht é tntt tallge SEU aT SANE the ja AA Tle! ta e time <0 NTL o dos’ eath veiter UCG-V Chee, cach Veale — PE ourddlour = = WHETE fh Vy & when >> prs-vrst tau) ) o cohtie, yentt ty found, Wier «OFS -vys77(6,u) ie by calling lime = time +) GOS ta ‘fh ~ Uid= time ~ Uccolour = GRAY «dor each wevtta Vln 6 najlul eee tf Me Colour = >VOHRTE ; yh Se : i —pFs ~VIstT Co) © time = time +1 ; + urf = time gales ‘ . yecoloury = BLACK N fab cde} LIf0 Qs: Sroat time, Atrph time , aa * 3, 0 yee ae la eee fc hae Breadth Trst Search (BFS) * BESCG)S) : ee ~ for each verter ucav— {SF 4 * Udolouy = WHITE TC haetipys 7) oe ape Nee —8:alor= ARAN —Ssd=0 = — 8:7 =-NEL oe a-g . iste i ; — €NHVELE (aS) work OF 6. * Us DEQVELE CO) 4 * Foi each. Verter V TP e-nagll Gioun = WHITE voto = GON ey See ts : €xonghs i ~. rs io ee ~ | 0, j {s] y oe Fut 4IEO ar at 1 ‘ Of iy} ak S Rody 1 i i se +47 ® iY va vi es ta et. 2 © of ov sited ~7 oe (3) s tuts | ea al TST 90 rod 8 rag t eed “Spanning @ree Algorithm 7 Slechonic crue « duignd Yeql¥e Conneag Mulipn Component — pfs ~ > quar & to mininnige wire wage while | all pins Me Connecti, bo Fh ts modiulid ay AN aundme cted “With gins a vestry g- wre CONS | ‘ edge weights : "2 GAaph veptesentation: + Refine graph G-=(v,6) 2 + cepiaents fins,and E epretentt . Potential “aonnections: : + Each: edge Curvy) hoa voeig ht oC oaneting Wire doit ote pins UE Ve > Objectie: pees ‘ : 7G an aayele Subset TCE that ao Minimige the, total WeIGht | (s LOBy Ga ube g ooo. OES Me problem Se Cat the. *: "Minimum -Apanning, Free" Geist)” —y ubhfob “ate! inp to chat design to opttmrge wine lage Pards Plgoatthen: ie ; «NAST perm (6/0) hy a —4toy each Veattr UEGV yUN Key 2 bo 7 1 ue NIL —- Key =6 ‘ ty ; , t -p-p peas Lg ~~ doy each verter. ueey | -rNseRt (au) ~— whi ged eee f ress exmencr- sence) — Kouskal’s Algoaithm » MST_ KRUSKAL(G)) ape ~ fot each vated VEGV : * MAKE - se) = (uote a — Afr tea weight hom, Lowest to pighut? Me °p checks, for each ealge: Curvy) -whethes th cndpoins ug iN) belong 0 Oy Wee a ee ee 744 “thay belong be ditrsent, trees, adel the edge Cusy) ‘to fh ¢ merge. the werttay fn the a thew Ps fr for Loop cramming edge 40 ota OF: tothemotie tgnme «the cage: : - ; - Ghertest Path Dlqor thmst— Dijkstra’s ee gov the ot oust: Shoitert paths psoklems 6 -& Ae Ae ctia graph ~— 6 (we) bug equines non- negative wetgnes on al ri tages (uv) 20 Jot ect ea cus) Algotitinm: DrrKstea(4 is) ~ INI TIALTSE - -S3N6le- covect Cars) 23-3 Moe ef (ee = tor each: Neater’ ve «Gert (@)u) = i ” —whil org cae 2 ed cerree.” Mente) oe 52 60tuy 3 i ) 6 a don each” yet VI RELAY Cav)

You might also like