| M. Narasimhan and J. Ramanujam, "A Fast Approach to Computing Exact Solutions to the Resource-Constrained Scheduling Problem". ACM Trans. on Design Automation of Electronic Systems, v.6 n.4, p.490-500, Oct. 2001. |
....#t i ;t j #; ASAP , ASAP#t i #; ALAP#t j ;UB# , ALAP#UB#;UB#. 3 H1 ASAPG t i i th operation in topologically sorted order H1 #t i # # LowerBound Gpred t i ; H1 G pred #t i # ; ALAPjG pred #t i # ;UB # enddo Figure 3: Code fragment to calculate H1 (and H2 ) Proof. See [8] for a proof. Theorem 3.6. Let S 2 # be a valid schedule for the DFG G = #V;E#.LetG = #V;E # be another DFG, where hx i ;x j i 2 E , hx j ;x i i 2 E. Then, S : V Z where #x j # = Latency#S# , S#x j # 1is a valid schedule for G and Latency#S# # Latency#S #. PROOF: In order ....
M. Narasimhan and J. Ramanujam. A fast approach to computing exact solutions to the resource-constrained scheduling problem. ACM Trans. Design Automation of Electronic Systems, to appear (2001).
....UB) Gamma ALAP(UB) UB) 3 H1 ASAPG for i 1 to jT j do t i i th operation in topologically sorted order H1 (t i ) LowerBound i Gpred Gamma t i Delta ; H1 fi fi G pred (t i ) ALAPjG pred (t i ) UB j enddo Figure 3: Code fragment to calculate H1 (and H2 ) Proof. See [8] for a proof. Theorem 3.6. Let S 2 OE be a valid schedule for the DFG G = V; E) Let G r = V; E r ) be another DFG, where hx i ; x j i 2 E r , hx j ; x i i 2 E. Then, S r : V Z where S r (x j ) Latency(S) Gamma S(x j ) 1 is a valid schedule for G r and Latency(S) Latency(S ....
M. Narasimhan and J. Ramanujam. A fast approach to computing exact solutions to the resource-constrained scheduling problem. ACM Trans. Design Automation of Electronic Systems, to appear (2001).
....shown how and ILP formulation can solve the scheduling problem, and how the efficiency of such an approach can be improved. In this section, we will be exploring another branch and bound scheme. As we have shown in Chapter 4, there are a number of schemes to produce lower bounds for this problem [NR97, LC96, RJ94] We have also demonstrated that some of these techniques can actually generate better (tighter) bounds than linear relaxation can. However, if another bounding scheme is to be used, then the criteria of optimality (i.e. if the lower bound is achievable) can be checked by also ....
M. Narasimhan and J. Ramanujam. A fast approach to computing exact solutions to the resource-constrained scheduling problem, Jul. 1997. Submitted for publication. 72
....than is ASAP . To compute the value of H 2 we make use of the following Lemma. Lemma 5 A lower bound on the ASAP function for a graph G is also an upper bound on the ALAP function for the graph G r obtained from G by reversing all the edge directions (dependencies) and vice versa. Proof. See [14] for a proof. Theorem 6 Let S 2 OE be a valid schedule for the DFG G = V; E) Let G r = V; E r ) be another DFG, where hx i ; x j i 2 E r , hx j ; x i i 2 E. Then, S r : V Z where S r (x j ) Latency(S) GammaS(x j ) 1 is a valid schedule for G r and Latency(S) Latency(S r ) ....
M. Narasimhan and J. Ramanujam. A fast approach to computing exact solutions to the resource-constrained scheduling problem, Jul. 1997. Submitted for publication. Tighter lower bounds for scheduling problems in high-level synthesis 22
No context found.
M. Narasimhan and J. Ramanujam, "A Fast Approach to Computing Exact Solutions to the Resource-Constrained Scheduling Problem". ACM Trans. on Design Automation of Electronic Systems, v.6 n.4, p.490-500, Oct. 2001.
No context found.
M. Narasimhan, J. Ramanujam. A fast approach to computing exact solutions to the resource-constrained scheduling problem. ACM Trans. Design Automation of Electronic Systems, 6(4), pp490-500, 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC