A. Deutsch. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages, pages 2--13, Washington, DC, 1992. IEEE Press.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....(in other words: they are aliased) The most relevant alias analysis algorithms are [LR92, LRZ93] The length of access paths are 1: limited, using a relatively simple truncation mechanism to eliminate extra path elements. Deutsch presents an alias analysis for an imperative subset of ML [Deu92]. Access paths are defined in terms of monomial relations (a kind of multi variable polynomial expression with structure accessors as the variables) The analysis is therefore only relevant for strongly typed languages such as ML and strongly typable programs written in weakly typed languages such ....

Alain Deutsch. A storeless model of aliasing and its abstractions using finite representations of fight-regular equivalence relations. In International Conference on Computer Languages, pages 13. IEEE, April 1992.

....references to the objects in local, global or field reference variables. When a reference variable has a non null value, we say that it poits to the object for which it contains a reference. The goal of static points to analysis is to statically ap proximate relationships among concrete objects [13, 23, 18]. One approach, which forms the basis of our datarace analysis framework, is to statically create abstract objects during analysis, associate them with accesses (defs uses) of reference variables in the program. We say that the abstract objects associated with an access are being polluted to by ....

A. Deutsch. A storeless model of aliasing and its abstrac- tions using finite representations of right-regular equivalence relations. In IEEE'92 International Conference on Computer Languages, Apr. 1992.

.... for the assignment a = b, #a and #b become aliased (this is called a type 1 alias e#ect) and so do i a and i b for all i # 2, assuming those object names make sense (this is called a type 2 alias e#ect) The type 2 alias e#ect induces the right regular property 21 discussed in [Deu92, ZRL96] The FIAlias algorithm partitions the program s object names using a union find data structure [ASU86] An overview of the steps of FIAlias is: 1. Consider each object name to be in its own partition element. 2. Perform unions of partition elements to account for type 1 alias e#ects at ....

A. Deutsch. A storeless model of aliasing and its abstractions using finite representations of right-regular equivalence relations. In Proceedings of the IEEE 1992 Conference on Computer Languages, pages 2--13, April 1992.

....by a eld. De ne a partial function E from pointer chains to objects in H as follows: E(v) Ref(v) if v 2 Var E(pc f) Ref(E(pc) f) otherwise: We call the relation Ker E the sharing pattern of the memory (Var; H;Ref) this relation is actually the same as the alias relation of Deutsch [3]) We assume to have the following statements in our programming language: x = new(f1, fk) x = y, x = y.f (we call these statements eld loads ) x.f = y (we call these statements eld stores ) x.f : y (we call these statements deep copies ) and statements for controlling control ow. ....

A. Deutsch. A Storeless Model of Aliasing and its Abstractions using Finite Representations of Right-Regular Equivalence Relations. In IEEE 1992 Int. Conf. on Computer Languages, San Francisco, April 1992.

....1993] Previous to this article, it was an open question whether such precision could ever be obtained by any method that uses graphs to model storage usage. Furthermore, as far as we know, no other shape analysis typechecking method (whether based on graphs or other principles [Choi et al. 1993; Deutsch 1992; Deutsch 1994; Ghiya and Hendren 1996; Hendren 1990; Hendren and Nicolau 1990; Landi and Ryder 1991] has the ability to determine that circular listness is preserved by the list insert program. What does our method do that allows it to obtain such qualitatively better results on the ....

Deutsch, A. 1992. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In Proceedings of the IEEE International Conference on Computer Languages. IEEE Computer Society, Washington, D.C., 2--13.

....location (in other words: they are aliased) The most relevant alias analysis algorithms are [LR92, LRZ93] The length of access paths are k limited, using a relatively simple truncation mechanism to eliminate extra path elements. Deutsch presents an alias analysis for an imperative subset of ML [Deu92]. Access paths are defined in terms of monomial relations (a kind of multi variable polynomial expression with structure accessors as the variables) The analysis is therefore only relevant for strongly typed languages such as ML and strongly typable programs written in weakly typed languages such ....

Alain Deutsch. A storeless model of aliasing and its abstractions using finite representations of right-regular equivalence relations. In International Conference on Computer Languages, pages 2--13. IEEE, April 1992.

....alias information kept by compact and graph based representations. In this graph, alias relation h p; vi holding after B3 can be inferred from alias relations hp; qi and hq; vi holding at the same place. By contrast, a full representation would represent h p; vi explicitly [Landi and Ryder 1992; Deutsch 1992; 1994; Sagiv et al. 1990] Recall from Section 3 that identifying aliases of p after B3 with a compact representation is equivalent to finding all (alias) paths of length two starting at p over the alias graph. In this section we first use alias relations, instead of alias instances, to ....

Deutsch, A. 1992. A storeless model of aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE'92 International Conference on Computer Languages.

....relation of Fig. 1. Instead of a correctness proof of the analyser with respect to the axiomatics as suggested here, the soundness of the analysis would then be shown as a consequence of the soundness of the abstract interpretation of the basic rules with respect to the operational semantics (see [10] for an illustration of this approach) Again, the most difficult rule is the assignment. It is not clear whether the overall effort would be less important but the formulation in terms of abstract interpretation would make it easier to show the optimality of the analyser (in terms of precision) ....

A. Deutsch, A storeless model of aliasing and its abstraction using finite representations of right-regular equivalence relations, in Proc. of the IEEE 1992 Conf. on Computer Languages, Apr. 92, pp. 2-13. RR n2895 24 Pascal Fradet, Ronan Gaugne and Daniel Le M'etayer

....it is often possible to determine precise shape information for programs that manipulate several (possibly cyclic) data structures (see Section 6. 1) Other staticanalysis techniques (including ones that are not based on shape graphs [Landi and Ryder 1991; Hendren 1990; Hendren and Nicolau 1990; Deutsch 1992; 1994] yield very imprecise information on these programs. 1.1.6 Relating 2 Valued Logic to 3 Valued Logic. The fundamental logical principles on which our shape analysis framework relies appear to be new. To relate 2 valued logic to 3 valued logic, we introduce a general notion of ....

....based on shape graphs that are incomparable to our method. For example, the algorithm of [Hendren 1990; Parametric Shape Analysis via 3 Valued Logic 5 Hendren and Nicolau 1990] handles destructive updates in many cases, but does not handle cyclic or doubly linked lists. The algorithm of [Deutsch 1992; 1994] is able to represent cyclic lists but fails to handle many kinds of destructive update operations (due to the absence of must alias information) The framework creates intraprocedural shape analysis algorithms, not interprocedural ones. Methods for handling procedures are presented in ....

[Article contains additional citation context not shown here]

Deutsch, A. 1992. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages.

....had an inaccuracy we had to modify it. After this we show a larger interprocedural example, the results of the analysis, and some measurements. Lastly we discuss how the analyzed information can be exploited for dependency analysis and parallelization. 2 Heap analysis According to [Deu92] heap analyses may be classi ed into three approaches: those using k limited graphs, those using abstract storage locations and those in ferring type information. The central problem of heap analysis is how to abstract from the statically unknown number of heap objects, especially from ....

....coalesce heap objects according to other, specialized criteria. e.g. CWZ90] uses so called summary nodes to represent an unknown number of objects, but leaves those heap objects uncoalesced that are directly and undoubtedly reachable. Methods inferring type information, for instance [Deu92] Bur92] represent the unknown number of objects belonging to an abstract data type as type information for the pointers. While this approach is very elegant its validity has to be evaluated further. Apart from this we can classify the methods according to the memory model they use. ....

[Article contains additional citation context not shown here]

A. Deutsch. A storeless model of aliasing and its abstractions using nite representations of right-regular equivalence relations. In J. Cordy, editor, IEEE International Conference on Computer Languages, March 1992.

....2 distinct subproblems: 1) disambiguating pointers that point to objects on the stack, and (2) disambiguating pointers that point to objects on the heap. There has been a considerable amount of work in both of these areas [JM81, JM82, LH88, Lar91, Gua88, HPR89, Har89, CWZ90, HN90, LR92, CBC93, Deu92, Deu94, PCK93, EGH94, WL95, Ruf95] although more attention has been paid to actually implementing methods that work well for stack allocated objects [LR92, CBC93, EGH94, Ruf95, WL95] A complete discussion and comparison of these methods can be found in [Ghi95] Stack directed pointers exhibit ....

....features. We present some empirical data in Section 4, to demonstrate the cost and e ectiveness of this abstraction for its intended domain of applications. Section 5 gives our conclusions and brie y describes our future work. 2 Connection Analysis Connection analysis uses a simple, storeless [Deu92] abstraction designed to disambiguate heap accesses at a coarse level, but in an ecient and cost e ective manner. For each program point the analysis computes a connection matrix, which is a boolean matrix summarizing the connectivity of heap objects. A heap object is de ned as an object ....

Alain Deutsch. A storeless model of aliasing and its abstractions using nite representations of right-regular equivalence relations. In Proc. of the 1992 Intl. Conf. on Computer Languages, pages 2-13, Oakland, Calif., Apr. 1992.

....results. A simple example is that one can take advantage of type information, and the absence of type casts to infer that a variable of one type cannot be aliased to a variable of another type. A more advanced example is the use of recursive types for dynamically allocated pointer structures [5, 13]. Pervasive Dataflow Information: It should be possible to transmit important dataflow information collected at higher level intermediate representations to the lower level representations, and thus improve the effectiveness of the low level transformations. For example, alias analysis ....

....analyses of the variables of these types, we need to break them up in a systematic manner so that appropriate alias analysis techniques can be applied. We define a varname to be a simple variable name (e.g. myname ) a pure structure reference(e.g. a.b. c ) a pure array reference (e.g. a[5][7] or a pointer to a structure reference (e.g. a) b.c ) Any other reference is broken down into these simple references. By standardizing the variable references in this way we can reduce the number and complexity of advanced alias analysis rules that we must define. Thus, structure and ....

Alain Deutsch. A storeless model of aliasing and its abstractions using finite representations of rightregular equivalence relations. In 1992 International Conference on Computer Languages, pages 2--13, April 1992.

....in programs. Only the most standard Pascal library functions are predefined. Programs with pointers to heap allocated objects are accepted, but are not always handled safely with respect to aliasing; other works on the abstract interpretation of heap allocated data structures such as Deutsch [13, 14] could be used to handle pointer induced aliasing. Records are accepted, but no information is given on their fields. This decision was made to simplify the design of the debugger, but records can be handled without much trouble. Jumps to local and non local labels are fully supported. It should ....

Alain Deutsch: "A Storeless Model of Aliasing and its Abstractions using Finite Representations of Right-Regular Equivalence Relations", Proc. of the IEEE'92 International Conference on Computer Languages, IEEE Press (1992)

....large programs. Static cleanness checking, when being conservative, can detect all errors. 1.3 Motivation In this thesis I investigated ways to conservatively discover cleanness violations with minimal false alarms. Rather than developing new algorithms, we exploit pointer analysis techniques [31, 32, 37, 5, 24, 26, 34, 11, 52, 35, 6, 42, 12, 14, 20, 21, 51, 46, 49, 47] to build a conservative algorithm for checking pointer dereferencing. Being interested in static analysis, we do not restrict ourselves to linear time algorithms. We are aware of the scalability problem of these algorithms. One 6 way to achieve scalability is by using global procedure ....

....5; s 3 : Figure 1.5: C code that explains the need to analyze conditions. s 1 : p = NULL; s 2 : q = p ; s 3 : q = i s 4 : p = 5; Figure 1.6: C code that explains why may alias analysis is not precise enough. ffl Abstract representations that record only pointer may aliases such as [34, 11, 12] are not precise enough. In the code shown in Figure 1.6, pointer may aliases algorithms can not conclude that after s 3 , p is not NULL and is pointing to i. This is since at s 3 these algorithms do not know that q must point to p. They conclude that p is either NULL or pointing to i and report ....

[Article contains additional citation context not shown here]

A. Deutsch. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages, pages 2--13, Washington, DC, 1992. IEEE Press.

No context found.

A. Deutsch. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages, pages 2--13, Washington, DC, 1992. IEEE Press.

No context found.

A. Deutsch. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages, pages 2--13, Washington, DC, 1992. IEEE Press.

No context found.

A. Deutsch. A storeless model for aliasing and its abstractions using finite representations of right-regular equivalence relations. In IEEE International Conference on Computer Languages, pages 2--13, Washington, DC, 1992. IEEE Press.

No context found.

A. Deutsch, "A storeless model of aliasing and its abstractions using finite representations of right-regular equivalence relations," in Proceedings of the 1992.

No context found.

A. Deutsch. A storeless model of aliasing and its abstraction using finite representations of right-regular equivalence relations. In Proceedings of the 1992 International Conference on Computer Languages, pages 2--13. IEEE Computer Society Press, 1992.

No context found.

A. Deutsch, A storeless model of aliasing and its abstraction using #nite representations of right-regular equivalence relations, 4th IEEE Intern. Conf. on Computer Languages (ICCL'92), San Francisco, CA, April 1992.

No context found.

A. Deutsch. A storeless model of aliasing and its abstraction using finite representations of right-regular equivalence relations. In Proceedings of the 1992 International Conference on Computer Languages, pages 2--13. IEEE Computer Society Press, Los Alamitos, California, U.S.A., 1992.

No context found.

A. Deutsch. A storeless model of aliasing and its abstraction using nite representations of right-regular equivalence relations. In Proceedings of the 1992 International Conference on Computer Languages, pages 2-13. IEEE Computer Society Press, 1992.

No context found.

A. Deutsch. A storeless model of aliasing and its abstraction using finite representations of right-regular equivalence relations. In Proceedings of the 1992 International Conference on Computer Languages, pages 2--13. IEEE Computer Society Press, 1992.

No context found.

A. Deutsch. A storeless model for aliasing and its abstractions using finite representation of right-regular equivalence relations. In IEEE Int. Conf. on Computer Languages, pp. 2--13, 1992.

No context found.

A. Deutsch, A storeless model of aliasing and its abstraction using #nite representations of right-regular equivalencerelations, in Proc. of the IEEE 1992 Conf. on Computer Languages, Apr. 92, pp. 2-13.

First 50 documents  Next 50

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