J. Knoop and B. Steffen. Efficient and optimal bit-vector data flow analyses: A uniform interprocedural framework. Technical Report Nr. 9309, Christian-Albrechts-Universitat zu Kiel, Kiel,Germany, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the procedures in reverse calling order. If the algorithm is applied in the presence of recursive procedures, approximations are made. The interprocedural extension is not a general interprocedural approach, but specifically developed for partial redundancy elimination. J. Knoop and B. Steffen [56, 57] present a generic interprocedural data flow framework which can be used to solve arbitrary data flow systems. It is a generalization of the well known intraprocedural Coincidence Theorem of Kam and Ullman [42] The generalization covers imperative programs with recursive procedures, global and ....

J. Knoop and B. Steffen. Efficient and optimal bit-vector data flow analyses: A uniform interprocedural framework. Technical Report Nr. 9309, Christian-Albrechts-Universitat zu Kiel, Kiel,Germany, 1993.

.... are described in [13, 48, 28, 49, 32] an alternative framework using graph reachability is presented in [47] A non exhaustive, demand driven framework is described in [17] Interprocedural data flow analysis algorithms for specific individual problems are examined in [3, 5, 24, 4, 36] and [33] developed an efficient interprocedural framework for locally separable problems. We have seen that this body of interprocedural data flow techniques can be utilized for analyses within each tuplespace process. Unfortunately, the differences between data flow between procedures and data flow ....

Jens Knoop and Bernhard Steffen. Efficient and optimal bit--vector data flow analyses: A uniform interprocedural framework. Technical Report Bericht Nr. 9309, Institut fur Informatik und Praktische Mathematik, Christian--Albrechts--Universitat Kiel, 1993.

.... switch n [17] case n Call p : 18] for each d 3 such that ( n, d 2 ) s calledProc(n) d 3 ) E # do [19] Propagate(PathEdge, s calledProc(n) d 3 ) s calledProc(n) d 3 ) WorkList) 20] od [21] for each d 3 such that ( n, d 2 ) succ 1 (n) d 3 ) E # SummaryEdge) do [22] Propagate(PathEdge, s p , d 1 ) succ 1 (n) d 3 ) WorkList) 23] od [24] end case [25] case n = e p : 26] for each c callers(p) do [27] for each d 4 , d 5 such that ( c, d 4 ) s p , d 1 ) E # and ( e p , d 2 ) succ 1 (c) d 5 ) E # do [28] if ( c, d 4 ) succ 1 (c) ....

.... from the invariant we knowthat (a) there is some c # C # such that there is a realizable path q # from c # to (n, d 2 ) and (b) there is a same levelrealizable path s # from (s p , d 1 )to(n, d 2 ) Propagate is called from four places in procedure ForwardTabulateSLRPs (lines [19] [22], 31] and [39] In each case, we can demonstrate that if the call on Propagate actually does insert a newedge into PathEdge, then there is a realizable path q ## and a same levelrealizable path s ## as required to re establish the invariant. For example, in the call in line [39] q ## is ....

[Article contains additional citation context not shown here]

Knoop, J. and Steffen, B., "Efficient and optimal bit-vector data flowanalyses: A uniform interprocedural framework,"Bericht Nr.9309, Institut fuer Informatik und Praktische Mathematik, Christian-Albrechts-Universitaet zu Kiel, Kiel, Germany(April 1993).

....algorithm is linear in the size of the system s representation and exponential in the size of the property. This is quite promising, as many practically relevant problems can be composed of very small properties: e.g. bitvector analyses, which are common in data flow analysis, have exponent one ([KnSt93a]) 6 Implementation and Performances A prototype of the fixpoint analysis machine has been implemented in C as part of the META Frame( StMC95, SFCM94, MaCS95] our environment for the development of heterogeneous analysis and verification tools, which currently runs on a SUN SparcStation 20 ....

J. Knoop, B. Steffen: "Efficient and Optimal Bit-vector Data Flow Analyses: A Uniform Interprocedural Framework", Inst. fur Informatik und Praktische Mathematik, Universitat Kiel (Germany), Bericht Nr. 9309 (1993).

....algorithm is linear in the size of the system s representation and exponential in the size of the property. This is quite promising, as many practically relevant problems can be composed of very small properties: e.g. bitvector analyses, which are common in data flow analysis have exponent one ([KnSt93a]) 6 Concluding Remarks We have reviewed finite and infinite state model checking from two pragmatic perspectives: the application to the static analysis of imperative programs or data flow analysis, and the implementational aspect. This application driven review gives some further evidence for ....

J. Knoop, B. Steffen: "Efficient and optimal bit-vector data flow analyses: A uniform interprocedural framework," Inst. fur Informatik und Praktische Mathematik, Universitat Kiel (Germany), Bericht Nr. 9309 (1993).

No context found.

Knoop, J., and Steffen, B. Efficient and optimal bit-vector data flow analyses: A uniform interprocedural framework. To appear.

....which are important e.g. for the automatic parallization of sequential programs. Subsequently, we present the interprocedural extension of the basic version of the algorithm for lazy code motion of [48] which optimally eliminates the interprocedural partial redundancies of a program (cf. [1, 49]) All analyses satisfy the four conditions guaranteeing their preciseness. Moreover, all analyses are as efficient as their corresponding intraprocedural versions, i.e. they have the same worst case time complexity. For the remainder of this section we consider an arbitrary, but fixed flow graph ....

.... in that procedure, of which it contains local variables (This assumes that the program is distinguished, i.e. different defining occurrences of identifiers use different identifiers) In fact, the corresponding optimization results in interprocedurally computationally optimal programs (cf. [1, 49]) Towards a tool kit for the automatic generation of interprocedural DFAs 21 1. Lattice: B 2 ; true ; false) false ; true) 2. Local Semantic Functional : 0 ea : N 0 ( B 2 B 2 ) defined by 8n 2 N 0 8 (b 1 ; b 2 ) 2 B 2 : n ] 0 ea (b 1 ; b 2 ) ....

J. Knoop and B. Steffen. Efficient and optimal bit-vector data flow analyses: A uniform interprocedural framework. Technical Report 9309, Institut fur Informatik und Praktische Mathematik, Christian-Albrechts-Universitat Kiel, Germany, 1993. 22 pages.

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