Deadlock-Free Fully Adaptive Routing in Irregular Networks without Virtual Channels

Size: px
Start display at page:

Download "Deadlock-Free Fully Adaptive Routing in Irregular Networks without Virtual Channels"

Transcription

1 Dedlock-Free Fully Adptive Routing in Irregulr Networks without Virtul Chnnels Dong Xing School of Softwre Tsinghu University Beijing , Chin Zhigng Yu Dept. of Computer Science nd Tech. Tsinghu University Beijing , Chin Jie Wu Dept. of Computer nd Info. Sciences Temple University Phildelphi, PA Astrct This pper proposes new flow control scheme in VCT-switched irregulr networks. Bsed on the new scheme, novel dedlock-free fully dptive routing lgorithm is introduced. The lgorithm does not need ny virtul chnnel. It requires tht ech input port of switch holds t lest two 1- pcket-sized uffers. The flow control scheme is proposed sed on seline routing scheme, where the downstrem nodes check the numer of sfe uffers t the upstrem nodes. The proposed fully dptive routing lgorithm is on the sis of different seline routing schemes: up*/down*, nd multiple spnning tree sed routing schemes. Extensive simultion results vlidte the effectiveness of the proposed method s compred to well-known existing pproches. Keywords Flow control, irregulr networks, dptive routing, multiple spnning trees. I. INTRODUCTION Networks of worksttions (NOWs) or clusters re recognized s good lterntives for prllel computing due to their competitive cost/performnce rtio nd wire flexiility [19]. Prcticl networks such s Autonet [25], Myrinet [1], nd Server-Net [9] re exmples of high-performnce with irregulr interconnects. The regulr networks, such s meshes/tori/hypercues, re unrelistic due to vritions in module sizes nd shpes, which re not suitle for clusters or NOWs. We consider virtul cut-through (VCT) switched NOWs ecuse we think VCT is more populr for NOWs or clusters. It is essentil to propose n effective dedlock-free dptive routing lgorithm in irregulr networks [33]. Wu nd Sheng [29] proposed dedlock-free routing scheme for irregulr networks using prefix routing. Three different multicst lgorithms were proposed for wormhole-switched irregulr networks in [11], with their respective node orders to reduce contention. Methods in Bolotin, et l. in [2] nd Meji, et l. in [17] minimize the size of the routing tle t ech router for NoCs with irregulr topologies. Flich, et l. in [8] proposed routing scheme LBDR for different vritions of the 2D meshes, which voids ny routing tles. It is found tht the LBDR still cnnot support some topologies. Rodrigo, et l. [24] proposed new mechnism, clled ulbdr, tht dpts to ny irregulr topologies derived from 2D meshes. The method in [24] still does not need ny routing tles. Cno, et l. in [3] proposed new routing methodology nd router implementtion for complex SoCs. This method voided routing tle look-up y mpping the irregulr network into 2D mesh with constnt or reduced logic cost, regrdless of the system size. A generl methodology for the design of dptive routing lgorithms for networks with irregulr topology hs een proposed in [26] y extending the Duto s protocol to the irregulr networks. Puente, et l. in [21] proposed pseudo- Hmiltonin-cycle-sed dptive routing mechnism for irregulr NOWs, which leds to pprent improvement over the clssicl up*/down* routing lgorithm, which comes t n cceptle extr cost. Puente, et l. in [22] proposed n dptive router rchitecture for irregulr NOWs, which improves network cpilities y llocting more resources to the fstest nd most-used virtul network. However, there my still exist trffic congestion sed on the method in [21], [22], ecuse single spnning tree (ST) ws used. Turn models were proposed for high-performnce routing in irregulr networks [10], [13], [18], [27]. These methods usully set the minimum numer of prohiited turns. A zoneordered lel-sed routing scheme ws proposed for irregulr networks with multiple spnning trees (s) in [18]. Any pcket is delivered long the ordered zones in [18]. There my e difficulty in finding the est zones nd zone order, especilly when the numer of STs is lrger. The Min contriutions of this pper include: (1) the generl frmework for fully dptive routing lgorithm in VCT-switched irregulr networks is presented, sed on simple flow control scheme; (2) dedlock-free fully dptive routing lgorithms re presented sed on the flow control scheme for up*/down*, nd routing schemes. In the rest of the pper, Section II presents the preliminries. The new flow control scheme for irregulr networks is in Section III. The dedlock-free dptive routing lgorithms re presented in Section IV. Simultion results re presented in Section V, nd Section VI concludes the pper. II. PRELIMINARIES Some ckground is provided first. The technique to select roots for the s, ST ssignment, nd constrined turn selection proposed in [31] re presented fter tht. A. Bckground The up*/down* routing scheme in irregulr networks ws proposed for Autonet networks in [25] y selecting single root nd then estlishing n ST. Its generl strtegy ws sed on selecting routes in n ST, where the pckets go up the ST on leving the source nd then, come ck down towrd the destintion.

2 () Fig. 1. An irregulr network with single ST: () the irregulr network with eight nodes, nd () the ST for up*/down* routing. Let the r e the root. The distnce dist(u, r) from u to r is greter thn tht from v to r. The link (u, v) is up. Otherwise, the link is down. As the dist(u, r) = dist(v, r), the direction of link (u, v) is set from the higher ID to the lower. The routing pth from source to destintion is estlished in such fshion tht zero or more up links (towrds the root) should e trversed efore zero or more down links cn e trversed (wy from the root) in order to rech the destintion. Fig. 1() presents n irregulr network with eight nodes, nd the ST for the up*/down* routing scheme when the node 1 is selected s the root is shown in Fig. 1(). The drwcks re tht the selected pths my not e the shortest pths nd tht links ner to the root cn e congested nd ecome ottlenecks. It ws shown in [20] tht dedlock-free routing without using virtul chnnels cn e successfully used in cycles of ny size. A restricted injection mechnism is pplied to ny pcket to void dedlocks tht is trying to enter the cycle which is clled Bule Flow Control(BFC). If under no circumstnces the storge spces for pckets in cycle re llowed to ecome full, the pckets trversing long this cycle will lwys e le to dvnce. B. Root Selection for Multiple Spnning Trees The method in [31] selects multiple roots for dptive dedlock-free routing in irregulr networks. It voids congestion t the root of the ST for the trditionl up*/down* routing. We use different metrics for the first root nd the rest roots. This is resonle ecuse selection of the first root is to minimize the verge distnce etween ny pir of nodes. We hve, F R = min r { u,v 3 8 () dist r (u, v)}, (1) where dist r (u, v) is the distnce etween u nd v ccording to the up*/down* routing scheme [25] in the ST with the root r. The lrger the size of cycle, the more influentil it is on the performnce. The remining roots re selected to minimize the numer of constrined turns for removing ll cyclic input port dependencies nd minimize the verge distnce for ll pirs of nodes sed on the metric s presented in Eqution (2), SR = u,v dist(u, v), (2) CT where CT is the numer of incresing constrined turns to remove ll cyclic dependencies in the input port dependency grph. The method in [31] selects the node s the root tht produces the mximum metric s presented in Eqution (2). In Eqution (2), u,v dist(u, v) stnds for reduction for distnce etween ll pirs of nodes. In ll cses, the distnce etween pir of nodes does not increse fter the second root is inserted nd the second ST is estlished. C. Spnning Tree Assignment A simple wy to select the ST for pir of nodes is: determine the ST ccording to the length of the miniml pth for them in the corresponding ST. The ST with shorter pth is selected for the pcket. However, lod-lncing should lso e considered when selecting the ST. Tht is, the numer of pckets with the sme source should e evenly ssigned to ll STs. We cll n ST ccording to its root. Consider delivering pcket from node 5 to node 7 s shown in Fig. 2. The miniml pth selected to deliver the pcket in the ST 1 y the originl up*/down* routing scheme is 5-1-7, where the pcket is delivered long n up link (5,1) first, nd down link (1,7) fter tht. The length of the pth is two. If the pcket is delivered vi the ST 8, the miniml pth selected y the up*/down* routing scheme is Four hops re necessry. Therefore, it is etter to deliver the pcket from node 5 to node 7 long the pths in the ST 1. If pcket from node 2 to node 6 is delivered in the ST 1, the miniml pth cn e The miniml pth to deliver the pcket is when it is delivered cross the ST 8. Tht is, the pth lengths in oth STs re equl. It is ok to deliver the pcket in oth STs. D. Constrined Turn Selection for Dedlock Avoidnce Fig. 2() presents the input port dependency grph of the irregulr network with input ports s shown in Fig. 2(). The vertices of the grph re input ports. There is dependency from u to v if pcket request uffer t v when it occupies uffer t u. There my exist some potentil cyclic dependencies. We propose new technique in [31] to constrin smll numer of turns insted of introducing some prohiited turns. A scheme like the ule flow control [20] is dopted in irregulr NOWs for dedlock voidnce. Tht is, the constrined turns re llowed when the uffer requirement t the input port is stisfied. The following metric is used to select constrined turns. m(v i ) = size(c), (3) v i C where the input port i of node v is contined in the cycle C. Our method ssigns enefit metric size(c), the size of the cycle C, for ech input port when the node nd its input port re contined in the cycle. The size of cycle is the numer of links to estlish the cycle. The reson why we ssign ech input port the size of the cycle is tht cycle of igger size cn e more influentil to the performnce. Our method selects the input port with the most metric. The enefit metric for ech node nd its input port is updted fter constrined turn node hs een selected. Tht is, the enefit metric of ny node nd its input port, tht re contined in the

3 1 d c c 1 d c c c c 2 c 3 5 2c c d 4c 1 c () : constrined turn in spnning tree with root 8 : the spnning tree with root () : constrined turn in spnning tree with root 1 : the spnning tree with root 1 Fig. 2. Cyclic chnnel dependencies in n irregulr network with STs: () the irregulr network with input ports, nd () the cyclic dependencies. flow-control-irregulr-network() Inputs: coordintes of the current node C, coordintes of the destintion D, free uffers: (f 1, f 2,..., f n), sfe pckets: (s 1, s 2,..., s n ) Output: selected output chnnel. 1) S := ; ch := null; 2) if C = D, ch := internl; 3) for ech next node i of C in the shortest pth to d ccording to the seline routing function R, do if (flow-control(i)) S S {i} ; 4) if S, ch := select(s). Fig. 3. The generl frmework of fully dptive routing lgorithm in irregulr networks sed on the new flow control scheme. cycle, is reduced y the size of the cycle. The ove process continues until ll cycles hve een removed. As shown in Fig. 2, the input port of 4 c, 3, 5, nd 7 re selected to remove ll six cyclic input port dependencies. III. THE NEW FLOW CONTROL SCHEME FOR IRREGULAR NETWORKS Recently, Luo nd Xing [15] proposed fully dptive routing lgorithm for tori without ny virtul chnnels sed on new flow control scheme. Tht work is the initil motivtion of this pper. In this section, we propose the generl frmework of new fully dptive routing lgorithm for irregulr networks sed on the flow control s shown in Fig. 3. The seline routing function R is dedlock-free, where it cn e the up*/down* routing scheme, -sed routing [31], nd ny other refined dedlock-free routing schemes. The input uffers of flow-control-irregulr-network() re orgnized s dynmiclly llocted multi-queues. Two queues re needed to void dedlocks. By this cse, the performnce would e even etter. Two clsses of pckets, sfe nd unsfe pckets, re defined in flow-control-irregulr-network(). Bsed on routing lgoflow-control(i) Inputs: the numer of free uffers: f i, nd the numer of sfe pckets: s i. Output: whether the pcket cn route to the next node. 1) If (f i > 1) return true; exit. 2) If (f i = 1 nd s > 0), return true; exit. 3) If (f i = 1 nd s i = 0), nd the next hop conforms to the seline routing scheme R, return true; exit. 4) Return flse. Fig. 4. The generl flow control function using R s the seline routing scheme. rithm R s the seline routing scheme in n irregulr network, pcket is sfe to the downstrem node if it is delivered long hop provided y the dedlock-free seline routing scheme R; otherwise, it is unsfe. In the rest of this pper, we sy pcket is sfe or unsfe mens tht it is sfe or unsfe to the current node if not specificlly defined. Fig. 3 presents the generl frmework of the fully dptive routing lgorithm for irregulr networks sed on the new flow control scheme nd seline routing scheme R. Here, f i, nd s i stnd for the numer of free uffers nd sfe pckets in the input port of the ith neighor directly connected to the current node C, respectively. The input of this lgorithm including coordintes of the current node nd the destintion, the numer of free uffer numers nd specil pcket numers of ll neighoring input ports. The ville chnnel set nd the selected output chnnel re initilized s nd null. If the current node is equl to the destintion, the internl chnnel is selected. Otherwise, for ech upstrem node v in the minimum pths from the current node to the destintion checks whether the pcket would e forwrded through the function flow-control() s presented in Fig. 4. The function flow-control(i) returns 1 when the pcket is sfe to the ith neighor nd free uffer is ville. Otherwise, it returns 0 when the input port is unsfe to the pcket. Function flow-control(i) voids filling ny input port with only unsfe

4 flow-control(i) Inputs: the numer of free uffers: f i, nd the numer of sfe pckets: s i. Output: whether the pcket cn route to the next node. 1) If (f i > 1) return true; exit. 2) If (f i = 1 nd s > 0), return true no mtter whether the next hop conforms to up*/down* routing; exit. 3) If (f i = 1 nd s i = 0), nd the next hop conforms to the up*/down* routing scheme, return true; exit. 4) Return flse. Fig. 5. Flow control function using up*/down* routing s the seline routing scheme. -sed-route(v,d) 1) Deliver the pcket from v to d if v is in the miniml pths from v to d in the ssigned ST in the following wy until it reches the destintion d, ) If the link is constrined turn, it cn e delivered if the input port of the next hop contins two empty uffers. ) Otherwise, it cn e delivered if the input port of the next hop contins t lest one empty uffer. c) Consume the pcket if the locl node hs een the destintion d. Fig. 6. -sed dedlock-free routing. pckets. It checks how mny free uffers (f) nd sfe pckets (s) in the input uffer of the ith input port.our method selects n output chnnel from S if it is not. Otherwise, the pcket is put in the witing list nd locked. The key point of the lgorithm flow-control-irregulrnetwork() is the flow-control(i) function. It voids filling ny input port with only unsfe pckets. The input uffers re orgnized s dynmiclly llocted multi-queues. There re two queues t ech input port: n sfe pcket queue nd unsfe pcket queue. All sfe pckets to node in n input port re linked s sfe pcket queue. The rest re linked s n unsfe pcket queue. So the sfe pckets would not e locked y unsfe pckets. In ll cses, no cyclic dependencies will e introduced y the dptive hops, therefore, the proposed fully dptive routing lgorithm is dedlock-free. We shll not provide more detiled proof on dedlock freedom of the proposed lgorithm in this pper. Fig. 5 presents the flow control function for the seline routing scheme up*/down* routing. Wht is ment y the next hop conforms to the up*/down* routing is tht one of the following three conditions must e met when the ith neighor of C is in the shortest pth from C to d: (1) the pcket reches the current node C vi n up link nd expects to rech the ith neighor vi n up link; (2) the pcket reches C y n up link nd the next node vi down link; (3) the pcket reches C y down link nd the next node vi down link. The sed routing scheme [31] cn e used s the seline routing scheme. The preliminry section presents more detils of the lgorithm. The -sed routing scheme provides dedlock-free routing ecuse ny possile cycle contins t lest one constrined turn, which introduces no dedlock configurtion. The following rules determine whether flow-control-(i) Inputs: the numer of free uffers: f i, nd the numer of sfe pckets: s i. Output: whether the pcket cn e delivered to the next node. 1) If (f i > 1) return true; exit. 2) If (f i = 1 nd s i > 0), no mtter whether the next hop conforms to the seline routing scheme R, return true; exit. 3) If (f i = 1 nd s i = 0), the next hop conforms to the seline routing scheme nd the input port of the next node i is not constrined turn, return true; exit. 4) Return flse. Fig. 7. The flow control function using sed routing s the seline routing scheme. pcket cn e delivered to the next node when one of the sed routing schemes is used s the seline routing scheme. Rememer tht pcket is ssigned to single ST. As shown in Fig. 7, ny pcket is delivered in the sme ST sed on up*/down* routing except the dptive hops. f > 1, the pcket could e delivered ecuse there is more thn one free uffers in the next node. f = 1 nd s > 0 no mtter whether input port of the next node is constrined turn, the pcket could e delivered ecuse there is t lest one sfe pcket in the next node which cn lwys free the occupied uffer. f = 1 nd s = 0 nd the next node is not constrined turn, the pcket could e delivered if the hop conforms to the seline routing scheme; otherwise, keep the pcket in the witing list. f = 0, keep the pcket in the witing list. The -sed dedlock-free dptive routing is presented in Fig. 6. It is cler tht the pcket cn e delivered to the neighor in shortest pth if the input port contins more thn one free uffer (f > 1). The pcket cn e delivered to the free uffer of the neighor if it contins one free uffer nd one sfe uffer no mtter whether the input port of the next node is not constrined turn ecuse the sfe uffer cn lwys e freed. The pcket cn e delivered to the next node if the hop conforms to the seline routing scheme, there exists free uffer nd no sfe uffer in the input port of the next node, nd the input port of the next node is not constrined turn. In ll other cses, the pcket cnnot e delivered to the next node. Wht is ment y conform to the sed routing scheme? Two consecutive hops, if none of them is n dptive hop provided y the sed routing, they must follow one of the following rules: (1) the pcket reches the current node C vi n up link nd expects to rech the next node vi n up link; (2) the pcket reches C y n up link nd the next node vi down link; (3) the pcket reches C y down link nd the next node vi down link. IV. DEADLOCK-FREE ADAPTIVE ROUTING BASED ON THE NEW FLOW CONTROL SCHEME We present new dptive routing scheme with unfixed STs. We then introduce the fully dptive routing lgorithms

5 dptive-route-up-down(c, d) Inputs: the current node C, nd the destintion d, directions for ll links in the ST, nd sttes of uffers t the neighors. Output: the next chnnel. 1) For ech neighor i of C in shortest pth to d, do steps 2), 3), 4), nd 5); 2) if the pcket reches C long up link nd the link to the next node i is still up link, nd oth hops re provided y the seline routing scheme, S S {i} when flowcontrol(i) is true; 3) if the pcket reches C long up link nd the link to the next node i is down link, nd oth hops re provided y the seline routing scheme, S S {i} when flowcontrol(i) is true; 4) if the pcket reches C long down link nd the link to the next node i is still down link, nd oth hops re provided y the seline routing scheme, S S {i} when flow-control(i) is true; 5) for ll other neighor i of C in shortest pth to d, i is not in S, if flow-control(i) is true nd i is not in S, S S {i}; 6) if S, ch := select(s). Fig. 8. Adptive routing using up*/down* routing s the seline routing scheme. with different seline routing schemes without ny virtul chnnel: (1) up*/down* routing, (2) sed routing, nd (3) unfixed sed routing. A. Fully Adptive Routing Using Up*/Down* Routing s the Bseline Routing Scheme We first propose the fully dptive routing lgorithm y using the up*/down* routing scheme s the seline routing scheme. Fig. 8 presents the detiled routing lgorithm. The inputs of the lgorithm includes the coordintes of the current node, nd the destintion, directions of the ll links of the ST, the root of the ST, nd the uffer sttes t the input ports of the current node C s neighors. Steps 2), 3), nd 4) present ville inputs port of the neighors s the next hop which conform to the seline routing scheme, up*/down* routing. The routing lgorithm in Fig. 8 shows three different cses when the next hop conforms to the seline routing scheme up*/down* routing. The next node must e in the shortest pths from C to d. The three cses include: (1) the pcket reches the current node long up link nd the link to the next node i is still up link, where oth hops re provided y the seline routing scheme; (2) the pcket reches C long up link nd the link to the next node i is down link, where oth hops re provided y the seline routing scheme; nd (3) the pcket reches C long down link nd the link to the next node i is still down link, where oth hops re provided y the seline routing scheme. After ll next hops long the shortest pths hve een identified, our method checks the next hops tht do not conform to the seline routing scheme. However, one of the uffer conditions presented y flow-control(i) in Fig. 4 must e stisfied. The selection function finlly selects one of the next hops in S to deliver the pcket. The possiility to estlish cyclic dependency y the proposed fully dptive routing lgorithm is tht one or more dptive hops. However, ny dependencies estlished y the dptive-route-(c,d) Inputs: the current node C, nd the destintion d; directions of ll links in ; ssignment of the pckets to ll STs. Output: the next chnnel. 1) For ech neighor i of C in shortest pth to d, do steps 2), 3), 4), nd 5); 2) if the pcket reches C long up link vi the seline routing function provided hop nd the link to the next node i is still up link in the jth ST, if flow-control-(i) is true, S S {i}; 3) if the pcket reches C long up link nd the link to the next node i is down link in the jth ST, nd oth links re provided y the seline routing function, if flow-control- (i) is true, S S {i}; 4) if the pcket reches C long down link nd the link to the next node i is still down link in the jth ST, nd oth links re provided y the seline routing function, if flow-control-(i) is true, S S {i}; 5) for ll other neighor i of C in shortest pth to d, if flowcontrol-(i) s presented in Fig. 7 returns true, nd i is not in S, S S {i}; 6) if S, ch := select(s). Fig. 9. Adptive routing using sed routing s the seline routing scheme. dptive hops cn e eliminted ecuse pcket cn dvnce y hop not provided y the seline routing scheme when the input port of the next node is sfe to the pcket. B. Fully Adptive Routing Using Multiple Spnning Tree Bsed Routing s the Bseline Routing Scheme Fig. 9 presents the fully dptive routing lgorithm sed on the new flow control scheme when the sed routing scheme is used s the seline scheme. The sed routing scheme is very simple, which is presented in Fig. 6. Initilly, the s re estlished nd the pckets re ssigned to the STs for lod-lncing considertion. A pcket is delivered long the fixed ST. However, some of the hops, tht do not conform to the seline routing scheme, do not follow this. Similr conditions s presented in Fig. 8 must e met when oth consecutive hops in the sme ST re provided y the seline routing scheme. The three cses include: (1) the pcket reches the current node long n up link nd the link to the next node i is still n up link; (2) the pcket reches C long n up link nd the link to the next node i is down link; nd (3) the pcket reches C long down link nd the link to the next node i is still down link. Our method then checks ll other next nodes in the shortest pths from the current node to the destintion whether one of the uffer conditions s presented in Fig. 7 cn e met. If so, the next node cn e provided y the fully dptive routing lgorithm. In ny cse, the pcket wits for one of the uffers provided y the seline routing scheme in the fixed ST when trffic congestion occurs. C. Dedlock Freedom Proof The fully dptive routing lgorithms re dedlock-free ecuse: (1) there re no cyclic dependencies in the sme ST ecuse the pckets follow the dedlock-free seline

6 Ltency(cycles) Trffic (flits/cycle/switch) ltency(cycles Trffic (flits/cycle/switch) Ltency(cycles) Trffic (flits/cycle/switch) A 0.3 A A 0.2 A Injection Rte(flits/cycle/switch) Injection Rte(flits/cycle/switch) Fig. 10. Performnce comprison with previous methods (,, A, ) in n irregulr network (32,64) A 0.12 A 100 A 0.08 A Injection Rte(flits/cycle/switch) Injection Rte(flits/cycle/switch) Fig. 11. Performnce comprison with previous methods (,, A, ) in n irregulr network (64,128) A 120 A A A Injection Rte(flits/cycle/switch) Injection Rte(flits/cycle/switch) Fig. 12. Performnce comprison with previous method (,, A, ) in n irregulr network (128,256). routing such s: up*/down* routing or sed routing, whose proofs cn e found in [25] nd [31], respectively; (2) oviously, the dptive hops provided y the new flow control scheme cnnot introduce ny cyclic dependencies [15]; (3) no cyclic dependencies cn e estlished cross different STs, ecuse ny cycle contins t lest one constrined turn. In ll, we cn conclude tht the ove fully dptive routing lgorithms re dedlock-free. V. SIMULATION RESULTS The proposed routing lgorithms nd flow control scheme for VCT-switched irregulr networks hve een implemented in cycle-ccurte simultor y C++ [31]. Different seline routing schemes re used including up*/down* routing (), multiple spnning tree sed routing (). The implemented fully dptive routing lgorithms sed on the proposed new flow control scheme re A nd A. The previous methods, the originl up*/down* routing scheme [25] (), the dptive routing scheme sed on Duto s protocol [26] (), nd the sed routing scheme () [31], re lso implemented. Links re rndomly dded to the network, while the degree of ech node is no more thn four [32]. Two metrics re used to evlute the performnce of method: the ltency to deliver pcket (cycle) nd the ccepted trffic (flit/cycle/switch). The size of pcket is set to 16 flits for ll simultion results. The topology of the irregulr network is generted rndomly, while we present the verge results of 10 irregulr networks. We consider three STs for ll simultion results relted to multiple spnning trees nd the uniform trffic pttern for ll simultion results. We do not consider routing tle storge nd look-up for previous methods ( nd ), nd just present the performnce comprison with them. Fig. 10 presents the performnce evlution of irregulr networks with 32 switches/nodes nd 64 links. The up*/down* routing scheme reches the sturtion point very quickly when the pplied lod reches 0.1. The extended Duto s

7 protocol reches the sturtion point when the pplied lod hs reched The dptive routing lgorithm A with the up*/down* routing scheme s the seline routing works etter thn nd, whose sturtion point is out 0.26, while the ccepted trffic is lso pprently etter thn the previous two methods. The most recent work (), sed on multiple spnning tree, works etter thn A, lthough it is not fully dptive. The sturtion point of is little lter, which is t The dptive routing lgorithm A is sed on the flow control scheme, which uses the multiple spnning tree sed routing s the seline routing scheme. The A routing lgorithm pprently works etter thn,, A, nd on ccepted trffic nd ltency to deliver pcket in ll cses. Figs. 11 nd 12, present the performnce evlution on irregulr networks with 64 switches nd 128 links, 128 switches nd 256 links. Fig. 11 presents the performnce comprison of the proposed lgorithms with previous methods. The lgorithm does not work etter thn nd A on ltency to deliver pcket when the injection rte is very low, just like tht in the networks with 32 switches. As shown in Fig. 11, the fully dptive routing lgorithm A still works etter thn when the injection rte reches The reson should e tht it is esy for nd A to reserve dptive chnnels when the injection rte is not high enough. The order for the lgorithms to rech the sturtion points is still,, A, nd A. Fig. 12 provides performnce comprison of the new lgorithms with previous ones in networks with 128 switches nd 256 links. The situtions re still the sme. The lgorithm works worse thn A when the injection rte is less thn 0.05, which consistently works etter thn,, nd A fter tht. The A lgorithms work even etter in ll cses. The proposed method voids routing tle look-up y implementing the routing tle look-up through some very simple extr logic. The detils for the router rchitecture re not presented in this pper. Therefore, the sclility is not prolem for the proposed method. In ll simultion results for nd, the extr ltency to look up the routing tles is not included. VI. CONCLUSIONS A new flow control scheme ws proposed for irregulr networks, which provides series of fully dptive routing lgorithms for VCT-switched (or pcket-switched) irregulr networks. Our methods provide dedlock-free dptive routing scheme using unfixed multiple spnning trees, tht is, pcket cn chnge its spnning tree t the intermedite nodes in order to reduce ltency to deliver it. The fully dptive routing frmework is pplied to up*/down* routing, multiple spnning tree, nd the unfixed multiple spnning tree routing schemes. Simultion results show tht the proposed flow control scheme cn provide very effective fully dptive routing lgorithms without ny virtul chnnels, nd ring 10% improvement when compred to the originl routing lgorithms. REFERENCES [1] N. J. Boden, D. Coen, R. E. Feldermn, A. E. Kulwik, C. L. Seitz, J. Seizovic nd W. Su, Myrinet - A gigit per second locl re network, IEEE Micro, pp , Fe [2] E. Bolotin, I. Cidon, R. Ginosr, nd A. Kolodny, Routing tle minimiztion for irregulr mesh NoCs, in Proc. of Design Automtion, nd Test in Europe, pp , [3] J. Cno, J. Flich, J. Duto, M. Coppol, nd R. Loctelli, Efficient routing implementtion in complex systems-on-chip, in Proc. of Int. Symp. on Networks-on-Chip, [4] J. Cong, C. Liu, nd G. Reinmn, ACES: Appliction-specific cycle elimintion nd splitting for dedlock-free routing on irregulr networkon-chip, in Proc. Design Automtion Conf., pp , June [5] J. Duto, A. Roles, F. Sill, nd R. Beivide, A comprison of router rchitectures for virtul cut-through nd wormhole switching in NOW environment, Journl of Prllel nd Distriuted Computing, vol. 61, no. 2, pp , [6] J. Duto, A necessry nd sufficient condition for dedlock-free routing in cut-through nd store-nd-forwrd networks, IEEE Trnsctions on Prllel nd Distriuted Systems, vol. 7, no. 8, pp , Aug [7] J. Flich, P. Lopez, M. P. Mlumres, nd J. Duto, Boosting the performnce of myrinet networks, IEEE Trns. Prllel nd Distriuted Systems, vol. 13, no. 12, pp , [8] J. Flich, S. Rodrigo, nd J. Duto, An efficient implementtion of distriuted routing lgorithms for NoCs, in Proc. of the Second Int. Symp. on Networks-on-Chip, pp , [9] R. Horst, TNet: A relile system re network, IEEE Micro, vol. 15, no. 1, pp , Fe [10] A. Jourku, M. Koiuchi, nd H. Amno, An effective design of dedlock-free routing lgorithms sed on 2D turn model for irregulr networks, IEEE Trns. on Prllel nd Distriuted Systems, vol. 18, no. 3, pp , Mrch [11] R. Kesvn, K. Bondlpti, nd D. K. Pnd, Multicst on irregulr switch-sed networks wormhole-routing, in Proc. of High- Performnce Computer Architecture, pp , [12] M. Koiuchi, A. Journku, nd H. Amno, Descending lyers routing: A dedlock-free deterministic routing using virtul chnnels in system re networks with irregulr topologies, in Proc. of Int. Conference on Prllel processing, pp , [13] L. Levitin, K. Krpovsky, nd M. Mustf, Miniml sets of turns for reking cycles in grphs modeling networks, IEEE Trns. Prllel nd Distriuted Systems, vol. 21, no. 9, pp , [14] S. Y. Lin, C. H. Hung, C. H. Cho, K. H. Hung, nd A. Y. Wu, Trffic-lnced routing lgorithm for irregulr mesh-sed on-chip networks, IEEE Trns. on Computers, vol. 57, no. 9, pp , Sept [15] W. Luo nd D. Xing, An efficient dptive dedlock-free routing lgorithm for torus networks, IEEE Trns. Prllel nd Distriuted Systems, vol. 23, no. 5, pp , My [16] O. Lysne, T. Skeie, S. Reiemo, nd I. Theiss, Lyered routing in irregulr networks, IEEE Trns. on Prllel nd Distriuted Systems, vol. 17, no. 1, pp , [17] A. Meji, M. Plesi, J. Flich, S. Kumr, P. Lopez, R. Holsmrk, nd J. Duto, Region-sed routing: A mechnism to support efficient routing lgorithm in NoCs, IEEE Trns. on VLSI Systems, vol. 17, no. 3, pp , Mrch [18] R. Morveiji, H. Srzi-Azd, nd A. Y. Zomy, Multispnning tree zone-ordered lel-sed routing lgorithms for irregulr networks, IEEE Trns. on Prllel nd Distriuted Systems, vol. 22, no. 5, pp , My [19] R. Perlmn, An lgorithm for distriuted computtion of spnning tree in n extended LAN, in Proc. of 9th Dt Communiction Symp., pp , [20] V. Puente, R. Beivide, J. A. Gregorio, J. M. Prellezo, J. Duto nd C. Izu, Adptive Bule Router: design to improve performnce in torus networks, in Proc. of Int. Conf. On Prllel Processing, Jpn, Sept [21] V. Puente, J. A. Gregorio, R. Beivide, F. Vllejo, nd A. Inez, A new routing mechnism for networks with irregulr topology, in Proc. of ACM/IEEE Int. Conf. on Supercomputing, pp , [22] V. Puente, J. A. Gregorio, F. Vllejo, R. Beivide, nd C. Izu, Highperformnce dptive routing for networks with ritrry topology, Journl of Systems Architecture, vol. 52, pp , 2006.

8 [23] W. Qio, L. Ni, nd T. Rokicki, Adptive-tril routing nd performnce evlution in irregulr networks using cut-through switches, IEEE Trns. on Prllel nd Distriuted Systems, vol. 10, no. 11, pp , Nov [24] S. Rodrigo, J. Flich, A. Roc, S. Medrdoni, D. Bertozzi, J. Cmcho, F. Sill, nd J. Duto, Addressing mnufcturing chllenges with costefficient fult-tolernt routing, in Proc. of 4th Int. Symp. on Networkson-Chip, pp , [25] M. D. Schroeder, A. Birrell, M. Burrows, H. Murry, R. Needhm, nd T. Rodeheffer, Autonet: A high-speed, self-configuring locl re network using point-to-point links, IEEE Journl Selected Ares in Communiction, vol. 9, no. 8, pp , Oct [26] F. Sill nd J. Duto, High-performnce routing in networks of worksttions with irregulr topology, IEEE Trns. Prllel nd Distriuted Systems, vol. 11, no. 7, pp , July [27] D. Stroinski, M. Krpovsky, nd L. A. Zrevski, Appliction of network clculus to generl topologies using turn prohiition, IEEE/ACM Trns. on Networking, vol. 11, no. 3, pp , June [28] I. Theiss nd O. Lysne, FRoots: A fult-tolernt nd topology-flexile routing technique, IEEE Trns. on Prllel nd Distriuted Systems, vol. 17, no. 10, pp , [29] J. Wu nd L. Sheng, Dedlock-free routing in irregulr networks using prefix routing, in Proc. of Int. Conf. Prllel nd Distriuted Systems, pp , [30] J. Wu, A fult-tolernt nd dedlock-free routing protocol in 2D meshes sed on odd-even turn model, IEEE Trns. Computers, vol. 52, no. 9, pp , Sept [31] D. Xing nd J. Hn, Multiple spnning tree construction for dedlockfree dptive routing in irregulr networks, in Proc. of IEEE Int. Symp. on Prllel nd Distriuted Processing with Appliction, pp. 9-16, July [32] M. Koiuchi, H. Mtsutni, H. Amno, D. Frnk Hsu, nd He. Csnov. A cse for rndom shortcut topologies for HPC interconnects, in Proc. of Int. Symp. on Computer Architecture, pp , [33] S. M, N. Enright Jerger, nd Z. Y. Wng. Whole pcket forwrding: Efficient design of fully dptive routing lgorithms for networks-onchip, in Proc of Int. Symp. on High-Performnce Computer Architecture,pp. 1-12, 2012.

Network Interconnection: Bridging CS 571 Fall Kenneth L. Calvert All rights reserved

Network Interconnection: Bridging CS 571 Fall Kenneth L. Calvert All rights reserved Network Interconnection: Bridging CS 57 Fll 6 6 Kenneth L. Clvert All rights reserved The Prolem We know how to uild (rodcst) LANs Wnt to connect severl LANs together to overcome scling limits Recll: speed

More information

Presentation Martin Randers

Presentation Martin Randers Presenttion Mrtin Rnders Outline Introduction Algorithms Implementtion nd experiments Memory consumption Summry Introduction Introduction Evolution of species cn e modelled in trees Trees consist of nodes

More information

COMP 423 lecture 11 Jan. 28, 2008

COMP 423 lecture 11 Jan. 28, 2008 COMP 423 lecture 11 Jn. 28, 2008 Up to now, we hve looked t how some symols in n lphet occur more frequently thn others nd how we cn sve its y using code such tht the codewords for more frequently occuring

More information

If you are at the university, either physically or via the VPN, you can download the chapters of this book as PDFs.

If you are at the university, either physically or via the VPN, you can download the chapters of this book as PDFs. Lecture 5 Wlks, Trils, Pths nd Connectedness Reding: Some of the mteril in this lecture comes from Section 1.2 of Dieter Jungnickel (2008), Grphs, Networks nd Algorithms, 3rd edition, which is ville online

More information

Fig.25: the Role of LEX

Fig.25: the Role of LEX The Lnguge for Specifying Lexicl Anlyzer We shll now study how to uild lexicl nlyzer from specifiction of tokens in the form of list of regulr expressions The discussion centers round the design of n existing

More information

P(r)dr = probability of generating a random number in the interval dr near r. For this probability idea to make sense we must have

P(r)dr = probability of generating a random number in the interval dr near r. For this probability idea to make sense we must have Rndom Numers nd Monte Crlo Methods Rndom Numer Methods The integrtion methods discussed so fr ll re sed upon mking polynomil pproximtions to the integrnd. Another clss of numericl methods relies upon using

More information

Performance Evaluation of Dynamic Reconfiguration in High-Speed Local Area Networks

Performance Evaluation of Dynamic Reconfiguration in High-Speed Local Area Networks Performnce Evlution of Dynmic Reconfigurtion in High-Speed Locl Are Networks Rfel Csdo, Aurelio Bermúdez, Frncisco J. Quiles, JoséL.Sánchez Depto. de Informátic Universidd de Cstill-L Mnch 271- Albcete,

More information

Ma/CS 6b Class 1: Graph Recap

Ma/CS 6b Class 1: Graph Recap M/CS 6 Clss 1: Grph Recp By Adm Sheffer Course Detils Adm Sheffer. Office hour: Tuesdys 4pm. dmsh@cltech.edu TA: Victor Kstkin. Office hour: Tuesdys 7pm. 1:00 Mondy, Wednesdy, nd Fridy. http://www.mth.cltech.edu/~2014-15/2term/m006/

More information

In the last lecture, we discussed how valid tokens may be specified by regular expressions.

In the last lecture, we discussed how valid tokens may be specified by regular expressions. LECTURE 5 Scnning SYNTAX ANALYSIS We know from our previous lectures tht the process of verifying the syntx of the progrm is performed in two stges: Scnning: Identifying nd verifying tokens in progrm.

More information

Dr. D.M. Akbar Hussain

Dr. D.M. Akbar Hussain Dr. D.M. Akr Hussin Lexicl Anlysis. Bsic Ide: Red the source code nd generte tokens, it is similr wht humns will do to red in; just tking on the input nd reking it down in pieces. Ech token is sequence

More information

IP: Network Layer. Goals and Tasks. Routing. Switching. Switching (cont.) Datagram v/s Virtual Circuit. Overview Addressing Routing

IP: Network Layer. Goals and Tasks. Routing. Switching. Switching (cont.) Datagram v/s Virtual Circuit. Overview Addressing Routing IP: Network Lyer Overview Addressing Routing Overview Gols nd Tsks Routing Switching Issues Bsic ides TOC IP TOC IP Overview Gols nd Tsks Gols of Network Lyer Guide pckets from source to destintion Use

More information

Slides for Data Mining by I. H. Witten and E. Frank

Slides for Data Mining by I. H. Witten and E. Frank Slides for Dt Mining y I. H. Witten nd E. Frnk Simplicity first Simple lgorithms often work very well! There re mny kinds of simple structure, eg: One ttriute does ll the work All ttriutes contriute eqully

More information

What are suffix trees?

What are suffix trees? Suffix Trees 1 Wht re suffix trees? Allow lgorithm designers to store very lrge mount of informtion out strings while still keeping within liner spce Allow users to serch for new strings in the originl

More information

CS321 Languages and Compiler Design I. Winter 2012 Lecture 5

CS321 Languages and Compiler Design I. Winter 2012 Lecture 5 CS321 Lnguges nd Compiler Design I Winter 2012 Lecture 5 1 FINITE AUTOMATA A non-deterministic finite utomton (NFA) consists of: An input lphet Σ, e.g. Σ =,. A set of sttes S, e.g. S = {1, 3, 5, 7, 11,

More information

2 Computing all Intersections of a Set of Segments Line Segment Intersection

2 Computing all Intersections of a Set of Segments Line Segment Intersection 15-451/651: Design & Anlysis of Algorithms Novemer 14, 2016 Lecture #21 Sweep-Line nd Segment Intersection lst chnged: Novemer 8, 2017 1 Preliminries The sweep-line prdigm is very powerful lgorithmic design

More information

Ma/CS 6b Class 1: Graph Recap

Ma/CS 6b Class 1: Graph Recap M/CS 6 Clss 1: Grph Recp By Adm Sheffer Course Detils Instructor: Adm Sheffer. TA: Cosmin Pohot. 1pm Mondys, Wednesdys, nd Fridys. http://mth.cltech.edu/~2015-16/2term/m006/ Min ook: Introduction to Grph

More information

Definition of Regular Expression

Definition of Regular Expression Definition of Regulr Expression After the definition of the string nd lnguges, we re redy to descrie regulr expressions, the nottion we shll use to define the clss of lnguges known s regulr sets. Recll

More information

Network Layer: Routing Classifications; Shortest Path Routing

Network Layer: Routing Classifications; Shortest Path Routing igitl ommuniction in the Modern World : Routing lssifictions; Shortest Pth Routing s min prolem: To get efficiently from one point to the other in dynmic environment http://.cs.huji.c.il/~com com@cs.huji.c.il

More information

MTH 146 Conics Supplement

MTH 146 Conics Supplement 105- Review of Conics MTH 146 Conics Supplement In this section we review conics If ou ne more detils thn re present in the notes, r through section 105 of the ook Definition: A prol is the set of points

More information

UT1553B BCRT True Dual-port Memory Interface

UT1553B BCRT True Dual-port Memory Interface UTMC APPICATION NOTE UT553B BCRT True Dul-port Memory Interfce INTRODUCTION The UTMC UT553B BCRT is monolithic CMOS integrted circuit tht provides comprehensive MI-STD- 553B Bus Controller nd Remote Terminl

More information

Distributed Systems Principles and Paradigms

Distributed Systems Principles and Paradigms Distriuted Systems Principles nd Prdigms Chpter 11 (version April 7, 2008) Mrten vn Steen Vrije Universiteit Amsterdm, Fculty of Science Dept. Mthemtics nd Computer Science Room R4.20. Tel: (020) 598 7784

More information

A Tautology Checker loosely related to Stålmarck s Algorithm by Martin Richards

A Tautology Checker loosely related to Stålmarck s Algorithm by Martin Richards A Tutology Checker loosely relted to Stålmrck s Algorithm y Mrtin Richrds mr@cl.cm.c.uk http://www.cl.cm.c.uk/users/mr/ University Computer Lortory New Museum Site Pemroke Street Cmridge, CB2 3QG Mrtin

More information

Tries. Yufei Tao KAIST. April 9, Y. Tao, April 9, 2013 Tries

Tries. Yufei Tao KAIST. April 9, Y. Tao, April 9, 2013 Tries Tries Yufei To KAIST April 9, 2013 Y. To, April 9, 2013 Tries In this lecture, we will discuss the following exct mtching prolem on strings. Prolem Let S e set of strings, ech of which hs unique integer

More information

Before We Begin. Introduction to Spatial Domain Filtering. Introduction to Digital Image Processing. Overview (1): Administrative Details (1):

Before We Begin. Introduction to Spatial Domain Filtering. Introduction to Digital Image Processing. Overview (1): Administrative Details (1): Overview (): Before We Begin Administrtive detils Review some questions to consider Winter 2006 Imge Enhncement in the Sptil Domin: Bsics of Sptil Filtering, Smoothing Sptil Filters, Order Sttistics Filters

More information

1.5 Extrema and the Mean Value Theorem

1.5 Extrema and the Mean Value Theorem .5 Extrem nd the Men Vlue Theorem.5. Mximum nd Minimum Vlues Definition.5. (Glol Mximum). Let f : D! R e function with domin D. Then f hs n glol mximum vlue t point c, iff(c) f(x) for ll x D. The vlue

More information

Complete Coverage Path Planning of Mobile Robot Based on Dynamic Programming Algorithm Peng Zhou, Zhong-min Wang, Zhen-nan Li, Yang Li

Complete Coverage Path Planning of Mobile Robot Based on Dynamic Programming Algorithm Peng Zhou, Zhong-min Wang, Zhen-nan Li, Yang Li 2nd Interntionl Conference on Electronic & Mechnicl Engineering nd Informtion Technology (EMEIT-212) Complete Coverge Pth Plnning of Mobile Robot Bsed on Dynmic Progrmming Algorithm Peng Zhou, Zhong-min

More information

this grammar generates the following language: Because this symbol will also be used in a later step, it receives the

this grammar generates the following language: Because this symbol will also be used in a later step, it receives the LR() nlysis Drwcks of LR(). Look-hed symols s eplined efore, concerning LR(), it is possile to consult the net set to determine, in the reduction sttes, for which symols it would e possile to perform reductions.

More information

Mobile IP route optimization method for a carrier-scale IP network

Mobile IP route optimization method for a carrier-scale IP network Moile IP route optimiztion method for crrier-scle IP network Tkeshi Ihr, Hiroyuki Ohnishi, nd Ysushi Tkgi NTT Network Service Systems Lortories 3-9-11 Midori-cho, Musshino-shi, Tokyo 180-8585, Jpn Phone:

More information

CS311H: Discrete Mathematics. Graph Theory IV. A Non-planar Graph. Regions of a Planar Graph. Euler s Formula. Instructor: Işıl Dillig

CS311H: Discrete Mathematics. Graph Theory IV. A Non-planar Graph. Regions of a Planar Graph. Euler s Formula. Instructor: Işıl Dillig CS311H: Discrete Mthemtics Grph Theory IV Instructor: Işıl Dillig Instructor: Işıl Dillig, CS311H: Discrete Mthemtics Grph Theory IV 1/25 A Non-plnr Grph Regions of Plnr Grph The plnr representtion of

More information

PARALLEL AND DISTRIBUTED COMPUTING

PARALLEL AND DISTRIBUTED COMPUTING PARALLEL AND DISTRIBUTED COMPUTING 2009/2010 1 st Semester Teste Jnury 9, 2010 Durtion: 2h00 - No extr mteril llowed. This includes notes, scrtch pper, clcultor, etc. - Give your nswers in the ville spce

More information

Lexical Analysis: Constructing a Scanner from Regular Expressions

Lexical Analysis: Constructing a Scanner from Regular Expressions Lexicl Anlysis: Constructing Scnner from Regulr Expressions Gol Show how to construct FA to recognize ny RE This Lecture Convert RE to n nondeterministic finite utomton (NFA) Use Thompson s construction

More information

Efficient Rerouting Algorithms for Congestion Mitigation

Efficient Rerouting Algorithms for Congestion Mitigation 2009 IEEE Computer Society Annul Symposium on VLSI Efficient Rerouting Algorithms for Congestion Mitigtion M. A. R. Chudhry*, Z. Asd, A. Sprintson, nd J. Hu Deprtment of Electricl nd Computer Engineering

More information

LECT-10, S-1 FP2P08, Javed I.

LECT-10, S-1 FP2P08, Javed I. A Course on Foundtions of Peer-to-Peer Systems & Applictions LECT-10, S-1 CS /799 Foundtion of Peer-to-Peer Applictions & Systems Kent Stte University Dept. of Computer Science www.cs.kent.edu/~jved/clss-p2p08

More information

From Dependencies to Evaluation Strategies

From Dependencies to Evaluation Strategies From Dependencies to Evlution Strtegies Possile strtegies: 1 let the user define the evlution order 2 utomtic strtegy sed on the dependencies: use locl dependencies to determine which ttriutes to compute

More information

Rethinking Virtual Network Embedding: Substrate Support for Path Splitting and Migration

Rethinking Virtual Network Embedding: Substrate Support for Path Splitting and Migration Rethinking Virtul Network Emedding: Sustrte Support for Pth Splitting nd Migrtion Minln Yu, Yung Yi, Jennifer Rexford, Mung Ching Computer Science, Princeton University, Emil: {minlnyu,jrex}@cs.princeton.edu

More information

Efficient Algorithms For Optimizing Policy-Constrained Routing

Efficient Algorithms For Optimizing Policy-Constrained Routing Efficient Algorithms For Optimizing Policy-Constrined Routing Andrew R. Curtis curtis@cs.colostte.edu Ross M. McConnell rmm@cs.colostte.edu Dn Mssey mssey@cs.colostte.edu Astrct Routing policies ply n

More information

CS 268: IP Multicast Routing

CS 268: IP Multicast Routing Motivtion CS 268: IP Multicst Routing Ion Stoic April 5, 2004 Mny pplictions requires one-to-mny communiction - E.g., video/udio conferencing, news dissemintion, file updtes, etc. Using unicst to replicte

More information

F. R. K. Chung y. University ofpennsylvania. Philadelphia, Pennsylvania R. L. Graham. AT&T Labs - Research. March 2,1997.

F. R. K. Chung y. University ofpennsylvania. Philadelphia, Pennsylvania R. L. Graham. AT&T Labs - Research. March 2,1997. Forced convex n-gons in the plne F. R. K. Chung y University ofpennsylvni Phildelphi, Pennsylvni 19104 R. L. Grhm AT&T Ls - Reserch Murry Hill, New Jersey 07974 Mrch 2,1997 Astrct In seminl pper from 1935,

More information

Lecture 10 Evolutionary Computation: Evolution strategies and genetic programming

Lecture 10 Evolutionary Computation: Evolution strategies and genetic programming Lecture 10 Evolutionry Computtion: Evolution strtegies nd genetic progrmming Evolution strtegies Genetic progrmming Summry Negnevitsky, Person Eduction, 2011 1 Evolution Strtegies Another pproch to simulting

More information

Solving Problems by Searching. CS 486/686: Introduction to Artificial Intelligence Winter 2016

Solving Problems by Searching. CS 486/686: Introduction to Artificial Intelligence Winter 2016 Solving Prolems y Serching CS 486/686: Introduction to Artificil Intelligence Winter 2016 1 Introduction Serch ws one of the first topics studied in AI - Newell nd Simon (1961) Generl Prolem Solver Centrl

More information

The Distributed Data Access Schemes in Lambda Grid Networks

The Distributed Data Access Schemes in Lambda Grid Networks The Distributed Dt Access Schemes in Lmbd Grid Networks Ryot Usui, Hiroyuki Miygi, Yutk Arkw, Storu Okmoto, nd Noki Ymnk Grdute School of Science for Open nd Environmentl Systems, Keio University, Jpn

More information

10.5 Graphing Quadratic Functions

10.5 Graphing Quadratic Functions 0.5 Grphing Qudrtic Functions Now tht we cn solve qudrtic equtions, we wnt to lern how to grph the function ssocited with the qudrtic eqution. We cll this the qudrtic function. Grphs of Qudrtic Functions

More information

Today. CS 188: Artificial Intelligence Fall Recap: Search. Example: Pancake Problem. Example: Pancake Problem. General Tree Search.

Today. CS 188: Artificial Intelligence Fall Recap: Search. Example: Pancake Problem. Example: Pancake Problem. General Tree Search. CS 88: Artificil Intelligence Fll 00 Lecture : A* Serch 9//00 A* Serch rph Serch Tody Heuristic Design Dn Klein UC Berkeley Multiple slides from Sturt Russell or Andrew Moore Recp: Serch Exmple: Pncke

More information

On the Detection of Step Edges in Algorithms Based on Gradient Vector Analysis

On the Detection of Step Edges in Algorithms Based on Gradient Vector Analysis On the Detection of Step Edges in Algorithms Bsed on Grdient Vector Anlysis A. Lrr6, E. Montseny Computer Engineering Dept. Universitt Rovir i Virgili Crreter de Slou sin 43006 Trrgon, Spin Emil: lrre@etse.urv.es

More information

Premaster Course Algorithms 1 Chapter 6: Shortest Paths. Christian Scheideler SS 2018

Premaster Course Algorithms 1 Chapter 6: Shortest Paths. Christian Scheideler SS 2018 Premster Course Algorithms Chpter 6: Shortest Pths Christin Scheieler SS 8 Bsic Grph Algorithms Overview: Shortest pths in DAGs Dijkstr s lgorithm Bellmn-For lgorithm Johnson s metho SS 8 Chpter 6 Shortest

More information

Reducing a DFA to a Minimal DFA

Reducing a DFA to a Minimal DFA Lexicl Anlysis - Prt 4 Reducing DFA to Miniml DFA Input: DFA IN Assume DFA IN never gets stuck (dd ded stte if necessry) Output: DFA MIN An equivlent DFA with the minimum numer of sttes. Hrry H. Porter,

More information

Approximation of Two-Dimensional Rectangle Packing

Approximation of Two-Dimensional Rectangle Packing pproximtion of Two-imensionl Rectngle Pcking Pinhong hen, Yn hen, Mudit Goel, Freddy Mng S70 Project Report, Spring 1999. My 18, 1999 1 Introduction 1-d in pcking nd -d in pcking re clssic NP-complete

More information

Revisiting the notion of Origin-Destination Traffic Matrix of the Hosts that are attached to a Switched Local Area Network

Revisiting the notion of Origin-Destination Traffic Matrix of the Hosts that are attached to a Switched Local Area Network Interntionl Journl of Distributed nd Prllel Systems (IJDPS) Vol., No.6, November 0 Revisiting the notion of Origin-Destintion Trffic Mtrix of the Hosts tht re ttched to Switched Locl Are Network Mondy

More information

Performance enhancement of IEEE DCF using novel backoff algorithm

Performance enhancement of IEEE DCF using novel backoff algorithm Kuo et l. EURASIP Journl on Wireless Communictions nd Networking 212, 212:274 http://jis.eursipjournls.com/content/212/1/274 RESEARCH Open Access Performnce enhncement of IEEE 82.11 using novel ckoff lgorithm

More information

ΕΠΛ323 - Θεωρία και Πρακτική Μεταγλωττιστών

ΕΠΛ323 - Θεωρία και Πρακτική Μεταγλωττιστών ΕΠΛ323 - Θωρία και Πρακτική Μταγλωττιστών Lecture 3 Lexicl Anlysis Elis Athnsopoulos elisthn@cs.ucy.c.cy Recognition of Tokens if expressions nd reltionl opertors if è if then è then else è else relop

More information

6.3 Volumes. Just as area is always positive, so is volume and our attitudes towards finding it.

6.3 Volumes. Just as area is always positive, so is volume and our attitudes towards finding it. 6.3 Volumes Just s re is lwys positive, so is volume nd our ttitudes towrds finding it. Let s review how to find the volume of regulr geometric prism, tht is, 3-dimensionl oject with two regulr fces seprted

More information

1. SEQUENCES INVOLVING EXPONENTIAL GROWTH (GEOMETRIC SEQUENCES)

1. SEQUENCES INVOLVING EXPONENTIAL GROWTH (GEOMETRIC SEQUENCES) Numbers nd Opertions, Algebr, nd Functions 45. SEQUENCES INVOLVING EXPONENTIAL GROWTH (GEOMETRIC SEQUENCES) In sequence of terms involving eponentil growth, which the testing service lso clls geometric

More information

Compression Outline :Algorithms in the Real World. Lempel-Ziv Algorithms. LZ77: Sliding Window Lempel-Ziv

Compression Outline :Algorithms in the Real World. Lempel-Ziv Algorithms. LZ77: Sliding Window Lempel-Ziv Compression Outline 15-853:Algorithms in the Rel World Dt Compression III Introduction: Lossy vs. Lossless, Benchmrks, Informtion Theory: Entropy, etc. Proility Coding: Huffmn + Arithmetic Coding Applictions

More information

Efficient Regular Expression Grouping Algorithm Based on Label Propagation Xi Chena, Shuqiao Chenb and Ming Maoc

Efficient Regular Expression Grouping Algorithm Based on Label Propagation Xi Chena, Shuqiao Chenb and Ming Maoc 4th Ntionl Conference on Electricl, Electronics nd Computer Engineering (NCEECE 2015) Efficient Regulr Expression Grouping Algorithm Bsed on Lbel Propgtion Xi Chen, Shuqio Chenb nd Ming Moc Ntionl Digitl

More information

A Priority-based Distributed Call Admission Protocol for Multi-hop Wireless Ad hoc Networks

A Priority-based Distributed Call Admission Protocol for Multi-hop Wireless Ad hoc Networks A Priority-bsed Distributed Cll Admission Protocol for Multi-hop Wireless Ad hoc Networks un Sun Elizbeth M. Belding-Royer Deprtment of Computer Science University of Cliforni, Snt Brbr suny, ebelding

More information

EECS150 - Digital Design Lecture 23 - High-level Design and Optimization 3, Parallelism and Pipelining

EECS150 - Digital Design Lecture 23 - High-level Design and Optimization 3, Parallelism and Pipelining EECS150 - Digitl Design Lecture 23 - High-level Design nd Optimiztion 3, Prllelism nd Pipelining Nov 12, 2002 John Wwrzynek Fll 2002 EECS150 - Lec23-HL3 Pge 1 Prllelism Prllelism is the ct of doing more

More information

Answering Label-Constraint Reachability in Large Graphs

Answering Label-Constraint Reachability in Large Graphs Answering Lel-Constrint Rechility in Lrge Grphs Kun Xu Peking University Beijing, Chin xukun@icst.pku.edu.cn Lei Chen Hong Kong Univ. of Sci. & Tech. Hong Kong, Chin leichen@cse.ust.hk Lei Zou Peking University

More information

A Sparse Grid Representation for Dynamic Three-Dimensional Worlds

A Sparse Grid Representation for Dynamic Three-Dimensional Worlds A Sprse Grid Representtion for Dynmic Three-Dimensionl Worlds Nthn R. Sturtevnt Deprtment of Computer Science University of Denver Denver, CO, 80208 sturtevnt@cs.du.edu Astrct Grid representtions offer

More information

On String Matching in Chunked Texts

On String Matching in Chunked Texts On String Mtching in Chunked Texts Hnnu Peltol nd Jorm Trhio {hpeltol, trhio}@cs.hut.fi Deprtment of Computer Science nd Engineering Helsinki University of Technology P.O. Box 5400, FI-02015 HUT, Finlnd

More information

A New Learning Algorithm for the MAXQ Hierarchical Reinforcement Learning Method

A New Learning Algorithm for the MAXQ Hierarchical Reinforcement Learning Method A New Lerning Algorithm for the MAXQ Hierrchicl Reinforcement Lerning Method Frzneh Mirzzdeh 1, Bbk Behsz 2, nd Hmid Beigy 1 1 Deprtment of Computer Engineering, Shrif University of Technology, Tehrn,

More information

GENERATING ORTHOIMAGES FOR CLOSE-RANGE OBJECTS BY AUTOMATICALLY DETECTING BREAKLINES

GENERATING ORTHOIMAGES FOR CLOSE-RANGE OBJECTS BY AUTOMATICALLY DETECTING BREAKLINES GENEATING OTHOIMAGES FO CLOSE-ANGE OBJECTS BY AUTOMATICALLY DETECTING BEAKLINES Efstrtios Stylinidis 1, Lzros Sechidis 1, Petros Ptis 1, Spiros Sptls 2 Aristotle University of Thessloniki 1 Deprtment of

More information

Solutions to Math 41 Final Exam December 12, 2011

Solutions to Math 41 Final Exam December 12, 2011 Solutions to Mth Finl Em December,. ( points) Find ech of the following its, with justifiction. If there is n infinite it, then eplin whether it is or. ( ) / ln() () (5 points) First we compute the it:

More information

Solving Problems by Searching. CS 486/686: Introduction to Artificial Intelligence

Solving Problems by Searching. CS 486/686: Introduction to Artificial Intelligence Solving Prolems y Serching CS 486/686: Introduction to Artificil Intelligence 1 Introduction Serch ws one of the first topics studied in AI - Newell nd Simon (1961) Generl Prolem Solver Centrl component

More information

The Greedy Method. The Greedy Method

The Greedy Method. The Greedy Method Lists nd Itertors /8/26 Presenttion for use with the textook, Algorithm Design nd Applictions, y M. T. Goodrich nd R. Tmssi, Wiley, 25 The Greedy Method The Greedy Method The greedy method is generl lgorithm

More information

Efficient K-NN Search in Polyphonic Music Databases Using a Lower Bounding Mechanism

Efficient K-NN Search in Polyphonic Music Databases Using a Lower Bounding Mechanism Efficient K-NN Serch in Polyphonic Music Dtses Using Lower Bounding Mechnism Ning-Hn Liu Deprtment of Computer Science Ntionl Tsing Hu University Hsinchu,Tiwn 300, R.O.C 886-3-575679 nhliou@yhoo.com.tw

More information

FastRoute: A Step to Integrate Global Routing into Placement

FastRoute: A Step to Integrate Global Routing into Placement FstRoute: A Step to Integrte Glol Routing into Plcement Min Pn nd Chris Chu Deprtment of Electricl nd Computer Engineering Iow Stte University, Ames, IA 00 Emil: {pnmin, cnchu}@istte.edu ABSTRACT Becuse

More information

cisc1110 fall 2010 lecture VI.2 call by value function parameters another call by value example:

cisc1110 fall 2010 lecture VI.2 call by value function parameters another call by value example: cisc1110 fll 2010 lecture VI.2 cll y vlue function prmeters more on functions more on cll y vlue nd cll y reference pssing strings to functions returning strings from functions vrile scope glol vriles

More information

Graphs with at most two trees in a forest building process

Graphs with at most two trees in a forest building process Grphs with t most two trees in forest uilding process rxiv:802.0533v [mth.co] 4 Fe 208 Steve Butler Mis Hmnk Mrie Hrdt Astrct Given grph, we cn form spnning forest y first sorting the edges in some order,

More information

Agilent Mass Hunter Software

Agilent Mass Hunter Software Agilent Mss Hunter Softwre Quick Strt Guide Use this guide to get strted with the Mss Hunter softwre. Wht is Mss Hunter Softwre? Mss Hunter is n integrl prt of Agilent TOF softwre (version A.02.00). Mss

More information

LR Parsing, Part 2. Constructing Parse Tables. Need to Automatically Construct LR Parse Tables: Action and GOTO Table

LR Parsing, Part 2. Constructing Parse Tables. Need to Automatically Construct LR Parse Tables: Action and GOTO Table TDDD55 Compilers nd Interpreters TDDB44 Compiler Construction LR Prsing, Prt 2 Constructing Prse Tles Prse tle construction Grmmr conflict hndling Ctegories of LR Grmmrs nd Prsers Peter Fritzson, Christoph

More information

Finite Automata. Lecture 4 Sections Robb T. Koether. Hampden-Sydney College. Wed, Jan 21, 2015

Finite Automata. Lecture 4 Sections Robb T. Koether. Hampden-Sydney College. Wed, Jan 21, 2015 Finite Automt Lecture 4 Sections 3.6-3.7 Ro T. Koether Hmpden-Sydney College Wed, Jn 21, 2015 Ro T. Koether (Hmpden-Sydney College) Finite Automt Wed, Jn 21, 2015 1 / 23 1 Nondeterministic Finite Automt

More information

Unit 5 Vocabulary. A function is a special relationship where each input has a single output.

Unit 5 Vocabulary. A function is a special relationship where each input has a single output. MODULE 3 Terms Definition Picture/Exmple/Nottion 1 Function Nottion Function nottion is n efficient nd effective wy to write functions of ll types. This nottion llows you to identify the input vlue with

More information

Digital Signal Processing: A Hardware-Based Approach

Digital Signal Processing: A Hardware-Based Approach Digitl Signl Processing: A Hrdwre-Bsed Approch Roert Esposito Electricl nd Computer Engineering Temple University troduction Teching Digitl Signl Processing (DSP) hs included the utilition of simultion

More information

Computer Arithmetic Logical, Integer Addition & Subtraction Chapter

Computer Arithmetic Logical, Integer Addition & Subtraction Chapter Computer Arithmetic Logicl, Integer Addition & Sutrction Chpter 3.-3.3 3.3 EEC7 FQ 25 MIPS Integer Representtion -it signed integers,, e.g., for numeric opertions 2 s s complement: one representtion for

More information

Fig.1. Let a source of monochromatic light be incident on a slit of finite width a, as shown in Fig. 1.

Fig.1. Let a source of monochromatic light be incident on a slit of finite width a, as shown in Fig. 1. Answer on Question #5692, Physics, Optics Stte slient fetures of single slit Frunhofer diffrction pttern. The slit is verticl nd illuminted by point source. Also, obtin n expression for intensity distribution

More information

Deterministic. Finite Automata. And Regular Languages. Fall 2018 Costas Busch - RPI 1

Deterministic. Finite Automata. And Regular Languages. Fall 2018 Costas Busch - RPI 1 Deterministic Finite Automt And Regulr Lnguges Fll 2018 Costs Busch - RPI 1 Deterministic Finite Automton (DFA) Input Tpe String Finite Automton Output Accept or Reject Fll 2018 Costs Busch - RPI 2 Trnsition

More information

COMBINATORIAL PATTERN MATCHING

COMBINATORIAL PATTERN MATCHING COMBINATORIAL PATTERN MATCHING Genomic Repets Exmple of repets: ATGGTCTAGGTCCTAGTGGTC Motivtion to find them: Genomic rerrngements re often ssocited with repets Trce evolutionry secrets Mny tumors re chrcterized

More information

Video-rate Image Segmentation by means of Region Splitting and Merging

Video-rate Image Segmentation by means of Region Splitting and Merging Video-rte Imge Segmenttion y mens of Region Splitting nd Merging Knur Anej, Florence Lguzet, Lionel Lcssgne, Alin Merigot Institute for Fundmentl Electronics, University of Pris South Orsy, Frnce knur.nej@gmil.com,

More information

On Achieving Optimal Throughput with Network Coding

On Achieving Optimal Throughput with Network Coding In IEEE INFOCOM On Achieving Optiml Throughput with Network Coding Zongpeng Li, Bochun Li, Dn Jing, Lp Chi Lu Astrct With the constrints of network topologies nd link cpcities, chieving the optiml end-to-end

More information

CSCI 446: Artificial Intelligence

CSCI 446: Artificial Intelligence CSCI 446: Artificil Intelligence Serch Instructor: Michele Vn Dyne [These slides were creted by Dn Klein nd Pieter Abbeel for CS188 Intro to AI t UC Berkeley. All CS188 mterils re vilble t http://i.berkeley.edu.]

More information

CS481: Bioinformatics Algorithms

CS481: Bioinformatics Algorithms CS481: Bioinformtics Algorithms Cn Alkn EA509 clkn@cs.ilkent.edu.tr http://www.cs.ilkent.edu.tr/~clkn/teching/cs481/ EXACT STRING MATCHING Fingerprint ide Assume: We cn compute fingerprint f(p) of P in

More information

MA1008. Calculus and Linear Algebra for Engineers. Course Notes for Section B. Stephen Wills. Department of Mathematics. University College Cork

MA1008. Calculus and Linear Algebra for Engineers. Course Notes for Section B. Stephen Wills. Department of Mathematics. University College Cork MA1008 Clculus nd Liner Algebr for Engineers Course Notes for Section B Stephen Wills Deprtment of Mthemtics University College Cork s.wills@ucc.ie http://euclid.ucc.ie/pges/stff/wills/teching/m1008/ma1008.html

More information

Algorithm Design (5) Text Search

Algorithm Design (5) Text Search Algorithm Design (5) Text Serch Tkshi Chikym School of Engineering The University of Tokyo Text Serch Find sustring tht mtches the given key string in text dt of lrge mount Key string: chr x[m] Text Dt:

More information

Constrained Optimization. February 29

Constrained Optimization. February 29 Constrined Optimiztion Februry 9 Generl Problem min f( ) ( NLP) s.. t g ( ) i E i g ( ) i I i Modeling nd Constrints Adding constrints let s us model fr more richer set of problems. For our purpose we

More information

A Scalable and Reliable Mobile Agent Computation Model

A Scalable and Reliable Mobile Agent Computation Model A Sclble nd Relible Mobile Agent Computtion Model Yong Liu, Congfu Xu, Zhohui Wu, nd Yunhe Pn College of Computer Science, Zhejing University Hngzhou 310027, Chin cckffe@yhoo.com.cn Abstrct. This pper

More information

Bleach: A Distributed Stream Data Cleaning System

Bleach: A Distributed Stream Data Cleaning System Blech: A Distriuted Strem Dt Clening System Yongcho Tin Eurecom Biot, Frnce Emil: yongcho.tin@eurecom.fr Pietro Michirdi Eurecom Biot, Frnce Emil: pietro.michirdi@eurecom.fr Mrko Vukolić IBM Reserch Zurich,

More information

Suffix trees, suffix arrays, BWT

Suffix trees, suffix arrays, BWT ALGORITHMES POUR LA BIO-INFORMATIQUE ET LA VISUALISATION COURS 3 Rluc Uricru Suffix trees, suffix rrys, BWT Bsed on: Suffix trees nd suffix rrys presenttion y Him Kpln Suffix trees course y Pco Gomez Liner-Time

More information

Parallel Square and Cube Computations

Parallel Square and Cube Computations Prllel Squre nd Cube Computtions Albert A. Liddicot nd Michel J. Flynn Computer Systems Lbortory, Deprtment of Electricl Engineering Stnford University Gtes Building 5 Serr Mll, Stnford, CA 945, USA liddicot@stnford.edu

More information

Queue-Management Architecture for Delay Tolerant Networking

Queue-Management Architecture for Delay Tolerant Networking Queue-Mngement Architecture for Dely Tolernt Networking Sotirios-Angelos Lens, Stylinos Dimitriou, Fni Tspeli, Vssilis Tsoussidis Spce Internetworking Center (SPICE), ECE Deprtment, Democritus University

More information

ON THE DEHN COMPLEX OF VIRTUAL LINKS

ON THE DEHN COMPLEX OF VIRTUAL LINKS ON THE DEHN COMPLEX OF VIRTUAL LINKS RACHEL BYRD, JENS HARLANDER Astrct. A virtul link comes with vriety of link complements. This rticle is concerned with the Dehn spce, pseudo mnifold with oundry, nd

More information

An Algorithm for Enumerating All Maximal Tree Patterns Without Duplication Using Succinct Data Structure

An Algorithm for Enumerating All Maximal Tree Patterns Without Duplication Using Succinct Data Structure , Mrch 12-14, 2014, Hong Kong An Algorithm for Enumerting All Mximl Tree Ptterns Without Dupliction Using Succinct Dt Structure Yuko ITOKAWA, Tomoyuki UCHIDA nd Motoki SANO Astrct In order to extrct structured

More information

Readings : Computer Networking. Outline. The Next Internet: More of the Same? Required: Relevant earlier meeting:

Readings : Computer Networking. Outline. The Next Internet: More of the Same? Required: Relevant earlier meeting: Redings 15-744: Computer Networking L-14 Future Internet Architecture Required: Servl pper Extr reding on Mobility First Relevnt erlier meeting: CCN -> Nmed Dt Network 2 Outline The Next Internet: More

More information

INTRODUCTION TO SIMPLICIAL COMPLEXES

INTRODUCTION TO SIMPLICIAL COMPLEXES INTRODUCTION TO SIMPLICIAL COMPLEXES CASEY KELLEHER AND ALESSANDRA PANTANO 0.1. Introduction. In this ctivity set we re going to introduce notion from Algebric Topology clled simplicil homology. The min

More information

MATH 25 CLASS 5 NOTES, SEP

MATH 25 CLASS 5 NOTES, SEP MATH 25 CLASS 5 NOTES, SEP 30 2011 Contents 1. A brief diversion: reltively prime numbers 1 2. Lest common multiples 3 3. Finding ll solutions to x + by = c 4 Quick links to definitions/theorems Euclid

More information

9 Graph Cutting Procedures

9 Graph Cutting Procedures 9 Grph Cutting Procedures Lst clss we begn looking t how to embed rbitrry metrics into distributions of trees, nd proved the following theorem due to Brtl (1996): Theorem 9.1 (Brtl (1996)) Given metric

More information

CS201 Discussion 10 DRAWTREE + TRIES

CS201 Discussion 10 DRAWTREE + TRIES CS201 Discussion 10 DRAWTREE + TRIES DrwTree First instinct: recursion As very generic structure, we could tckle this problem s follows: drw(): Find the root drw(root) drw(root): Write the line for the

More information

Today. Search Problems. Uninformed Search Methods. Depth-First Search Breadth-First Search Uniform-Cost Search

Today. Search Problems. Uninformed Search Methods. Depth-First Search Breadth-First Search Uniform-Cost Search Uninformed Serch [These slides were creted by Dn Klein nd Pieter Abbeel for CS188 Intro to AI t UC Berkeley. All CS188 mterils re vilble t http://i.berkeley.edu.] Tody Serch Problems Uninformed Serch Methods

More information

CSCI 3130: Formal Languages and Automata Theory Lecture 12 The Chinese University of Hong Kong, Fall 2011

CSCI 3130: Formal Languages and Automata Theory Lecture 12 The Chinese University of Hong Kong, Fall 2011 CSCI 3130: Forml Lnguges nd utomt Theory Lecture 12 The Chinese University of Hong Kong, Fll 2011 ndrej Bogdnov In progrmming lnguges, uilding prse trees is significnt tsk ecuse prse trees tell us the

More information

ΕΠΛ323 - Θεωρία και Πρακτική Μεταγλωττιστών. Lecture 3b Lexical Analysis Elias Athanasopoulos

ΕΠΛ323 - Θεωρία και Πρακτική Μεταγλωττιστών. Lecture 3b Lexical Analysis Elias Athanasopoulos ΕΠΛ323 - Θωρία και Πρακτική Μταγλωττιστών Lecture 3 Lexicl Anlysis Elis Athnsopoulos elisthn@cs.ucy.c.cy RecogniNon of Tokens if expressions nd relnonl opertors if è if then è then else è else relop è

More information

Announcements. CS 188: Artificial Intelligence Fall Recap: Search. Today. Example: Pancake Problem. Example: Pancake Problem

Announcements. CS 188: Artificial Intelligence Fall Recap: Search. Today. Example: Pancake Problem. Example: Pancake Problem Announcements Project : erch It s live! Due 9/. trt erly nd sk questions. It s longer thn most! Need prtner? Come up fter clss or try Pizz ections: cn go to ny, ut hve priority in your own C 88: Artificil

More information