W.A. Babich and M. Jazayeri. The Method of Attributes for Data Flow Analysis, Part II: Demand Analysis. Acta Informatica, 10(3):265-272, 1978.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... [6] algorithms for use at the source level [45] complications due to variable aliasing [35] more powerful approaches to data ow analysis [28] general improvements to ow analysis algorithms [8] methods to deal with the huge amounts of information that exhaustive data ow analysis generates [9], application to other areas of computer science including secure information ow [16] and most important to this dissertation, the application of data ow techniques to software maintenance problems [18, 53] 2.2.4 1980 1989 The early 1980s saw increasing interest in the area of applying ....

....dependencies of all call sites. If a pfdom arc is removed during pfdom resolution, then the related ifa target inherits the control dependence of the procedure called by the ifa source. These two steps of the algorithm are simple to perform. 61 v 2 1 3 4 5 6 2 I A = 2,3] 4,5] 7,8] [9,10] B.PFDF type = T I B = C.PFDF type = F A B 2,9 4,7 Figure 4.2: Items Required for Computing CCDG. Figure 4.2 shows the structures for computing compositional control dependence for procedures A, B, and C that make up the example program. These are referred to in the following ....

[Article contains additional citation context not shown here]

W.A. Babich and M. Jazayeri. The Method of Attributes for Data Flow Analysis, Part II: Demand Analysis. Acta Informatica, 10(3):265-272, 1978.

.... [5, 46] improved eciency [20, 29] the de nition of data ow problems and their related solutions [6] algorithms for use at the source level [45] complications due to variable aliasing [35] more powerful approaches to data ow analysis [28] general improvements to ow analysis algorithms [8], methods to deal with the huge amounts of information that exhaustive data ow analysis generates [9] application to other areas of computer science including secure information ow [16] and most important to this dissertation, the application of data ow techniques to software maintenance ....

....dependencies of all call sites. If a pfdom arc is removed during pfdom resolution, then the related ifa target inherits the control dependence of the procedure called by the ifa source. These two steps of the algorithm are simple to perform. 61 v 2 1 3 4 5 6 2 I A = 2,3] 4,5] [7,8], 9,10] B.PFDF type = T I B = C.PFDF type = F A B 2,9 4,7 Figure 4.2: Items Required for Computing CCDG. Figure 4.2 shows the structures for computing compositional control dependence for procedures A, B, and C that make up the example program. These are referred to in the ....

[Article contains additional citation context not shown here]

W.A. Babich and M. Jazayeri. The Method of Attributes for Data Flow Analysis, Part I: Exhaustive Analysis. Acta Informatica, 10(3):265-272, 1978.

....may avoid multiple passes of falsification and reiteration of the common a#ected aliases. Demand Driven Analysis While the complete dataflow analysis computes the information at all program points, a demand data flow analysis determines whether a given data flow fact holds at a given point [BJ78, Rep94, DGS95] The Landi Ryder aliasing algorithm is a whole program analysis; we are not seeking to derive a demand driven version from it, but to handle changes and queries on a demand driven basis. After a source change is made, if we are interested in the possible impact of the change on the ....

W. A. Babich and M. Jazayeri. The method of attributes for data flow analysis, part II: Demand analysis. Acta Informatica, 10:265--272, 1978.

....may avoid multiple passes of falsification and reiteration of the common a#ected aliases. 4. 8 Demand Driven Analysis While the complete data flow analysis computes the information at all program points, a demand data flow analysis predicts whether a given data flow fact holds at a given point [BJ78, Rep94, HRS95, DGS95, DGS97] The Landi Ryder aliasing algorithm is a whole program analysis; we are not seeking to derive a demand driven version from it, but to handle changes and queries on a demand driven basis. That is, after a change is made at a node n, if we are interested in the possible ....

W. A. Babich and M. Jazayeri. The method of attributes for data flow analysis, part II: Demand analysis. Acta Informatica, 10:265--272, 1978.

....techniques similar to Rosen s. Farrow et al. FKZ76] propose an approach similar to Sharir s, although only for the CFGs of semi structured pro5 grams. Because it uses the AST directly, the approach most similar in spirit to the one taken in this dissertation is Babich s method of attributes [BJ78] which associates data flow values ( attributes ) with each node in an AST, and then iterates over the AST re computing these values in a manner depending on the node type until they stabilize. Except for that of Farrow et al. these approaches all handle arbitrary GOTOs. The approach we take ....

....program and d is its maximum loop nesting depth [KU76] thus, in the worst case it can require O(N 2 ) primitive operations. Babich and Jazayeri considered the corresponding AST based algorithm for a canonical programming language having only WHILE DO, IF THEN, and GOTO as control structures [BJ78] and showed that it required no more than two passes over programs containing no GOTOs (in other words, structured programs) They also showed that it required no more than W L 1 passes in general, where W is the number of WHILE DO statements in the program, and L the number of labels that are ....

W. A. Babich and M. Jazayeri. The method of attributes for data flow analysis, parts I and II. Acta Informatica, 10(3):245--272, 1978.

....CFG. 4. Visit the mirParallel statements from outside in (the reverse order of inside out) and compute in and out of the process bodies. Outside in is the top down traversal of the PCFG along the parallel edges. This kind of computation of the data flow information is a mixture of the structural [2], 1, page 611] for the effect of the mirParallel statement) and iterative (all other statements) method. A similar idea has been presented for DFA of sequential problems by [8] The computation of the in and out information must be done using an iterative algorithm [11, 1] At a first glance the ....

W.A. Babich, M. Jazayeri. The method of attributes for data flow analysis. Acta Informatica, 10:345--272, 1978.

No context found.

W.A. Babich and M. Jazayeri. The Method of Attributes for Data Flow Analysis, Part II: Demand Analysis. Acta Informatica, 10(3):265-272, 1978.

No context found.

W.A. Babich and M. Jazayeri. The Method of Attributes for Data Flow Analysis, Part I: Exhaustive Analysis. Acta Informatica, 10(3):265-272, 1978.

No context found.

W. A. Babich and M. Jazayeri. The method of attributes for data flow analysis, part II: Demand analysis. Acta Informatica, 10:265--272, 1978.

No context found.

Wayne A. Babich and Mehdi Jazayeri. The method of attributes for data flow analysis. Acta Informatica, 10:345--272, 1978. Part I Exhaustive Analysis, Part II Demand Analysis.

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