Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with Combined Analysis for Pointer Aliasing. In Proceedings of the ACM SIGPLAN/SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11--18, Montreal, Canada, June 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....23 (dst ) src ) 24 dst=0; 25 return dst; 26 Figure 1: Example program. ments, these algorithms must iteratively propagate information throughout the program using the control flow graph. Studies indicate that these algorithms may be too expensive to be applied to large programs [9, 12]. Flow insensitive alias analysis algorithms (e.g. 1, 11] compute less precise alias information for a program. These algorithms avoid iteratively propagating information throughout the program by ignoring the execution order of the statements. Studies indicate that these algorithms are more ....

....information for a program. These algorithms avoid iteratively propagating information throughout the program by ignoring the execution order of the statements. Studies indicate that these algorithms are more efficient and thus, scale better to large programs, than the flow sensitive algorithms [9, 12]. A program analysis might compute imprecise program information when the analysis uses alias information provided by a flow insensitive alias analysis algorithm. One way that this imprecision occurs is that, to resolve pointer dereferences in a program, the program analysis may use the large ....

S. Zhang, B. G. Ryder, and W. Landi. Experiments with combined analysis for pointer aliasing. In Program Analysis for Software Tools and Engineering '98, pages 11--18, 1998. 9

.... Researchers have studied the effects of the alias information computed by different alias analysis algorithms, with varying levels of precision, on the precision and efficiency of data flow analyses when the pointer variables are handled using the method discussed in the previous paragraph [12, 14, 19]. The empirical results of these studies show that, when the size of the subject program is large, the side effects of a statement or a procedure can be greatly increased through dereferences of pointer variables. For example, Zhang, Ryder, and Landi [19] report that, when Landi and Ryder s ....

....discussed in the previous paragraph [12, 14, 19] The empirical results of these studies show that, when the size of the subject program is large, the side effects of a statement or a procedure can be greatly increased through dereferences of pointer variables. For example, Zhang, Ryder, and Landi [19] report that, when Landi and Ryder s algorithm [11] a flow sensitive alias analysis algorithm, is used to resolve the dereference of pointers, a statement in a C program that contains approximately 13,000 lines of code can modify, on average, 20 memory locations through dereference of pointer ....

[Article contains additional citation context not shown here]

S. Zhang, B. G. Ryder, and W. Landi. Experiments with combined analysis for pointer aliasing. In Program Analysis for Software Tools and Engineering '98, pages 11--18, 1998. 8

....of this analysis is unknown. Landi and Ryder [LR92] study a very precise ow sensitive, context sensitive analysis. Their ow sensitive system has diculty scaling to large programs; recent work has focused on combined analyses that apply di erent alias analyses to di erent parts of a program [ZRL98]. Chatterjee, Ryder, and Landi [CRL99] propose an analysis for Java and C that uses a ow sensitive analysis with conditional points to relations whose validity depends on the aliasing and type information provided by the context. While the style of polymorphism used in [CRL99] appears related ....

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with Combined Analysis for Pointer Aliasing. In Proceedings of the ACM SIGPLAN/SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11-18, Montreal, Canada, June 1998. 30

....all pointers simultaneously, as most of the existing approaches do, is the primary obstacle to efficient pointer analysis. To address the efficiency of pointer analysis, researchers have recently explored techniques to partition the pointer analysis into a group of subproblems [ZRL96, Ruf97, ZRL98] int a, p, q, pp; main ( s1: pp = p; s2: foo( s3: p = s4: pp = q; s5: foo( s6: q = void foo( s7: pp = a; call foo call foo analyzing main analyzing foo (a) An example program. b) Interleaving the intraprocedural analysis with the ....

....approach, we also separate the analysis of recursive data structures and non recursive data structures. We focus on the efficiency in analyzing nonrecursive data structures. Zhang et al. put the pointers that may contain the same value in the same group for the purpose of partitioning [ZRL96, ZRL98] Ruf groups variables by their types (declared types or inferred types) and analyzes one variable group at a time [Ruf97] A partitioned approach, in general, can improve efficiency due to several reasons: Separating subproblems with different characteristics allows the flexibility of ....

[Article contains additional citation context not shown here]

Sean Zhang, Barbara G. Ryder and William Landi. Experiments with Combined Analysis for Pointer Aliasing. In Proceedings of Workshop on Program Analysis for Software Tools and Engineering, Jun. 1998.

....large programs. Shapiro et. al [18] propose a hybrid method with con gurable complexity between Steensgaard s and Anderson s algorithms. Hasti et al. 34] propose the idea of iteratively improving the accuracy of pointer analysis in SSA form [3] but empirical result is not available. Zhang et al. [35] and Stocks et al. 36] balance the complexity by switching between various algorithms. The largest benchmark handled has around 60,000 lines of code. Our algorithm is the rst to demonstrate that a ow insensitive analysis accommodating complete C features can handle programs of more than 200,000 ....

S. Zhang, B. G. Ryder, and W. A. Landi, \Experiments with combined analysis for pointer aliasing," in Proceedings of the 1998 Workshop on Program Analysis for Software Tools and Engineering, June 1998.

....of this analysis is unknown. Landi and Ryder [20] study a very precise ow sensitive, context sensitive analysis. Their ow sensitive system has diculty scaling to large programs; recent work has focused on combined analyses that apply di erent alias analyses to di erent parts of a program [35]. Chatterjee, Ryder, and Landi [6] propose an analysis for Java and C that uses a ow sensitive analysis with conditional points to relations whose validity depends on the aliasing and type information provided by the context. While the style of polymorphism used in [6] appears related to ....

S. Zhang, B. G. Ryder, and W. A. Landi. Experiments with Combined Analysis for Pointer Aliasing. In Proceedings of the ACM SIGPLAN/SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11-18, Montreal, Canada, June 1998.

....containing a pointer dereference and, using the computed alias information, reports how many named objects are aliased to the pointer expression. We report the average number of such dereferences for both reads and writes. This form of counting, also used by Emami et al. 1994] Ruf [1995] and Zhang et al. 1998], provides a precision metric based on the use of alias information, and therefore can be more meaningful than recording the number of aliases at all program points. Further precision information is provided in Hind and Pioli [1998b] and Pioli [1999] A pointer expression containing multiple ....

....1993] to represent alias information. This representation is similar to the points to representation [Emami et al. 1994; Ghiya 1992] which has been used by a number of researchers [Andersen 1994; Hasti and Horwitz 1998; Ruf 1995; Shapiro and Horwitz 1997b; Steensgaard 1996; Wilson and Lam 1995; Zhang et al. 1998]. Landi and Ryder [1992] use a representation that explicitly describes all alias relations at the cost of increased storage. The Appendix illustrates how the compact and explicit representations have incomparable precision. Similar discussions can be found in Marlowe et al. 1993] Landi et al. ....

[Article contains additional citation context not shown here]

Zhang, S., Ryder, B. G., and Landi, W. 1998. Experiments with combined analysis for pointer aliasing. In ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering. 11--18.

....representation visiting each expression containing a pointer dereference and, using the computed alias information, reports how many named objects are aliased to the pointer expression. We report the average number of such dereferences for both reads and writes. This form of counting, also used in [13, 35, 53], provides a precision metric based on the use of alias information, and therefore can be more meaningful than recording the average alias set size, i.e. the average number of aliases at all program points. A pointer expression with multiple dereferences, such as p, is counted as multiple ....

....1.3 to 7.2 over existing methods. Diwan et al. 12] provide static and dynamic measurements of the effectiveness of three flow insensitive analyses for a type safe language (Modula 3) All three algorithms are less precise than the versions we have studied, except for the AT analysis. Zhang et al. [53] report the effectiveness of applying different pointer aliasing algorithms to different parts of a program. Chatterjee et al. 8] describe a technique for incorporating relevant context information into a data flow analysis and illustrate their approach for points to analysis. Empirical results ....

Sean Zhang, Barbara G. Ryder, and William Landi. Experiments with combined analysis for pointer aliasing. In ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11--18, June 1998. 24

....can control the context sensitive level of the algorithm. Our approach has the following features: ffl It enables the use of a syntax directed flow sensitive analysis technique which is generally more efficient than its iterative counterpart [1] ffl It facilitates a combined analysis technique [25, 20, 26] which applies different algorithms to analyze different data structure groups in order to improve the precision and the efficiency of the algorithm. ffl It can provide a spectrum of context sensitive algorithms, ranging from a context insensitive algorithm to a full context sensitive algorithm, ....

....information. 4) Their algorithm is based on a sparse evaluation graph [5, 3] with only the pointer related assignments and function calls in a program. Ignoring the non pointer related assignments this way can simplify the control flow graph, thereby speeding up the analysis. Both Zhang et al. [25, 26] and Ruf [20] presented a program decomposition alias analysis algorithm. They first divided object names into equivalent classes using pointer related assignments. They then used the prefix relation 2 between the object names to draw dependency edges between equivalent classes. Finally, they ....

Sean Zhang, Barbara G. Ryder, and William Landi. Experiments with combined analysis for pointer aliasing. In Proceedings of Workshop on Program Analysis for Software Tools and Engineering, June 1998.

....of Steensgaard s analysis [49] in that it groups all objects pointed to by a variable into an equivalence class. Although this can lead to spurious alias relations not present in the FI analysis, it does allow for an almost linear time algorithm, which has been shown to be fast in practice [49, 46, 55]. Emami et al. 17] report precision results for a flow sensitive context sensitive algorithm. Their results range from 1.0 to 1.77 for all indirect accesses using a heap naming scheme that represents all heap objects with one name. Because we were unable to obtain the benchmarks from the suite ....

....of 1.3 7.2 over existing methods. Diwan et al. 16] provide static and dynamic measurements of the effectiveness of three flow insensitive analyses for a type safe language (Modula 3) With the exception of AT, all three algorithms are less precise than the versions we have studied. Zhang et al. [55] report the effectiveness of applying different pointer aliasing algorithms to different parts of a program. Hasti and Horwitz [22] present a pessimistic algorithm that attempts to increase the precision of a flow insensitive analysis by iterating over a flow insensitive analysis and an SSA [12] ....

Sean Zhang, Barbara G. Ryder, and William Landi. Experiments with combined analysis for pointer aliasing. In ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11--18, June 1998. A Appendix

....of Steensgaard s analysis [45] in that it groups all objects pointed to by a variable into an equivalence class. Although this can lead to spurious alias relations not present in the FI analysis, it does allow for an almost linear time algorithm, which has been shown to be fast in practice [45, 42, 50]. Emami et al. 14] report precision results for a flow sensitive context sensitive algorithm. Their results range from 1.0 to 1.77 for all indirect accesses using a heap naming scheme that represents all heap objects with one name. Because we were unable to obtain the benchmarks from their suite, ....

....of 1.3 7.2 over existing methods. Diwan et al. 13] provide static and dynamic measurements of the effectiveness of three flow insensitive analyses for a type safe language (Modula3) With the exception of AT, all three algorithms are less precise than the versions we have studied. Zhang et al. [50] report the effectiveness of applying different pointer aliasing algorithms to different parts of a program. Hasti and Horwitz [18] present a pessimistic algorithm that attempts to increase the precision of a flow insensitive analysis by iterating over a flowinsensitive analysis and an SSA [10] ....

Sean Zhang, Barbara G. Ryder, and William Landi. Experiments with combined analysis for pointer aliasing. In ACM SIGPLAN-SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11--18, June 1998.

....at calls and for procedures are presented and discussed in Section 4. The empirical results show the utility of both analyses for specific applications and demonstrate the precision gains from sensitivity for certain data flow information. Recent work in partitioning programs for analyses [ZRL98, ZRL96, Ruf97] yields hope that analyses of varying cost and precision can be applied to di#erent parts of a program to obtain desired data flow information at practical cost. The experiments reported here can be viewed as the initial investigation into the cost precision tradeo#s involved when ....

.... Another approach to side e#ect analysis is to perform an interprocedural pointer aliasing algorithm and then identify all variables experiencing side e#ects at indirect stores through a pointer (i.e. at through dereference statements) using the aliases found [EGH94, GH98, Ruf95, HP98, SH97a, ZRL98, ZRL96] This is often used as an empirical test of the precision of the alias solution obtained. Related Analyses Related interprocedural analyses include compile time interprocedural program slicing [GL91, HRB90, OO84, RR95, GS96, LH96, HC98, SHR99, TCFR96, Tip96, Ven91, Wei84, TAFM97, AG96, ....

[Article contains additional citation context not shown here]

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with combined analysis for pointer aliasing. In Proceedings of ACM SIGPLAN Workshop on Program Analysis and Software Tools for Engineering, pages 11--18, June 1998. 80 A Comparison with the MOD Decomposition for FORTRAN The decomposition of the MOD problem for C in

....execution when two or more names refer to the same location. The precision of pointer alias information greatly a#ects its utility in applications such as dataflow based testers, semantic program browsers, debuggers, and compilers. For example, modification side e#ect analysis [LRZ93, YRLS97, ZRL98] requires pointer aliasing information in the presence of assignments through dereferenced pointers, to find the set of variables that may be modified. Data flow based testers utilize definition use information, needing aliases at uses of values obtained through dereferenced pointers. ....

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with combined analysis for pointer aliasing. In Proceedings of ACM SIGPLAN Workshop on Program Analysis and Software Tools for Engineering, pages 11--18, June 1998. 23

....at calls and for procedures are presented and discussed in Section 4. The empirical results show the utility of both analyses for specific applications and demonstrate the precision gains from sensitivity for certain data flow information. Recent work in partitioning programs for analyses [ZRL98, ZRL96, Ruf97] yields hope that analyses of varying cost and precision can be applied to different parts of a program to obtain desired data flow information at practical cost. The experiments reported here can be viewed as the initial investigation into the cost precision tradeoffs involved when ....

....no implementation results. Another approach to side effect analysis is to perform an interprocedural pointer aliasing algorithm and then identify all variables experiencing side effects at indirect stores through a pointer (i.e. at through dereference statements) using the aliases found [SH97a, ZRL98, ZRL96] This is often used as an empirical test of the precision of the alias solution obtained. Related Analyses Related interprocedural analyses include compile time interprocedural program slicing [GL91, HRB90, OO84, RR95, GS96, LH96, TCFR96, Tip96, Ven91, Wei84, TAFM97, AG96] ....

[Article contains additional citation context not shown here]

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with combined analysis for pointer aliasing. In Proceedings of ACM SIGPLAN Workshop on Program Analysis and Software Tools for Engineering, June 1998. A Comparison with the MOD Decomposition for FORTRAN The decomposition of the MOD problem for C in

No context found.

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with Combined Analysis for Pointer Aliasing. In Proceedings of the ACM SIGPLAN/SIGSOFT Workshop on Program Analysis for Software Tools and Engineering, pages 11--18, Montreal, Canada, June 1998.

No context found.

S. Zhang, B. G. Ryder, and W. A. Landi. Experiments with combined analysis for pointer aliasing. ACM SIGPLAN Notices, 33(7):11--18, July 1998.

No context found.

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with combined analysis for pointer aliasing. ACM SIGPLAN Notices, 33(7):11--18, July 1998.

No context found.

Sean Zhang, Barbara G. Ryder, and William A. Landi. Experiments with combined analysis for pointer aliasing. In PASTE [PASTE1998].

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