R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In POPL, 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....contains a pointer to another. This case is similar to the one optimized by traditional object inlining (except that our optimization is far more aggressive) To find these cases, the first step is to transform the code to SSA [12] form and create a heap approximation using techniques from [20]. SSA simplifies program analysis while the heap approximation provides the compiler with the sharing patterns between objects, allowing for more informed object combining decisions. The next step is to locate all occurrences of object allocations: X = allocate object(type) X = allocate ....

....each function of the program. If so, the two unrelated objects are combined. 8.3.3 Recursive Objects The algorithms described thus far combine a fixed number of objects. Many programs, however, use recursive data structures like lists, which are inherently dynamic. In a heap approximation graph [20], all elements of such a recursive data structure are typically represented by a single node, so the compiler cannot distinguish them. Still, in many cases it may be beneficial to apply object combining to such data structures. Our optimization combines the data structures into multiple slabs ....

[Article contains additional citation context not shown here]

R. Ghiya and L.J. Hendren. Putting Pointer Analysis To Work. In Proceedings of the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pages 121--133, San Diego, California, USA, January 1998. ACM Press.

....of dependence based queries [31] This metric will be particularly useful in evaluating analyses for strongly typed languages such as Java, where worst case assumptions are not as bad as in C. The second alternative static metric is to implement a pointer analysis client and report its precision [90, 63, 3, 30, 20, 86, 95, 55, 57, 40, 60, 31, 77]. The advantage of this metric is that it more clearly ascertains the e ectiveness of a pointer analysis for a particular client analysis. The limitation is that it only measures one client, and again, the importance of the static measure of the client may be questioned. Type inference attempts ....

....of granularity and tries to determine the (named) objects held in a pointer. The type inference literature, although certainly worth exploring, is beyond the scope of this paper. The third approach dynamically measures how pointer information a ects a runtime property, such as program performance [14, 105, 20, 30, 12, 31], dynamic points to relations [20, 62] or a dynamic characterization of the client analysis [20, 30, 12, 31] Such metrics are limited to a single execution of the program, and thus, represent a lower bound on their static counterparts. Because all metrics have their strengths and weaknesses a ....

[Article contains additional citation context not shown here]

R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In 25th Annual ACM SIGACT-SIGPLAN Symposium on the Principles of Programming Languages, pages 121-133, Jan. 1998.

....a reference to an object allocated at another object allocation site. Heap analysis also gives us the set of object allocation sites to which global variables, local variables, arguments, or parameters may refer. The implementation described here is a variation of the heap analysis described in [8] but extended to handle the parameter semantics of RMI. As described in the introduction, parameters and return values are cloned during the (de)serialization process. To arrive at the most precise representation of the runtime heap, the heap approximation needs to reflect the cloning process by ....

R. Ghiya and L.J. Hendren. Putting Pointer Analysis To Work. In Proceedings of the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pages 121--133, San Diego, CA., January 1998.

....at compile time. Jackal s compiler performs two optimizations: object graph aggregation and automatic computation migration. These two optimizations are similar to our connectivity based object pushing and synchronized method migration. Object graph aggregation uses a heap approximation algorithm [34] to identify those connected objects. However, the heap approximation algorithm cannot distinguish between di#erent objects that are created in the same allocation site. Thus this approach is e#ective only for the situation where the related objects are from di#erent allocation sites. In contrast, ....

R. Ghiya, L. J. Hendren, Putting Pointer Analysis to Work, in: 25th Annual ACM SIGACT-SIGPLAN Symposium on the Principles of Programming Languages, 1998, pp. 121--133.

....performance is done at compile time. Jackal s compiler performs two optimizations: object graph aggregation and automatic computation migration, whose aims are similar to those of our object pushing and synchronized method migration. Objectgraph aggregation uses a heap approximation algorithm [15] to identify those related objects. However, the heap approximation algorithm cannot distinguish between di#erent objects that are created at the same allocation site. Thus this approach is e#ective only for the situation when the related objects are from di#erent allocation sites. In contrast, ....

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In 25th Annual ACM SIGACT-SIGPLAN Symposium on the Principles of Programming Languages, pages 121--133, January 1998.

....relationship between any two variables. Although a knowledge of variable scope can cut down the number of dependency pairs, the assumption is still too strong for dependency analysis to derive an accurate relationship. Several methods of alias analysis and point to analysis have been proposed [12, 6, 19, 2, 20]. By using these methods, one can compute a points to set for each variable. The points to analysis is beyond the scope of this paper, and we assume that a points to set for each variables could be obtained via this analysis. The main concern of this paper is how to use the points to sets to build ....

R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In POPL, San Diego, CA USA, 1998.

....efforts to improve performance are done at compile time. Jackal s compiler enables two optimizations: object graph aggregation and automatic computation migration, which are similar to our object pushing and synchronized method migration. Object graph aggregation uses heap approximation algorithm [6] to identify those related objects. However, heap approximation algorithm cannot distinguish between different runtime objects that are created at the same allocation site. Hence this approach is effective only at the situation when the related objects are from different allocation sites. ....

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In 25th Annual ACM SIGACT-SIGPLAN Symposium on the Principles of Programming Languages, pages 121--133, January 1998.

....We provide a few examples below. 8.1 Preliminary Global Analysis and Simplication We have already indicated that a preliminary rough global program analysis can always be performed to improve the information available before performing a local analysis. A classical example is pointer analysis [35, 41, 5053, 56, 58, 59, 64, 67, 72], see an overview in [69] A preliminary pointer analysis is often mandatory since making conservative assumptions regarding pointer accesses can adversely aect the precision and eciency of the analysis of the program parts requiring this information. Such pointer alias analysis attempts to ....

R. Ghiya and L.J. Hendren. Putting pointer analysis to work. In 25 121133, San Diego, Jan. 1998. ACM Press.

....point. Without Extended SSA Numbers, one must create a new anchor handle for each indirect assignment in the program. However, using Extended SSA Numbers, anchor handles can be reused if the primary SSA number does not change. This optimization reduces the number of anchor handles by over 50 [Ghi97] 5.7 Summary Overall, we have found that SSA numbers are useful for many of our subsequent analyses and transformations, and SSA numbers are now an integral part of our compiler. 6 Related Work Most previous work has concentrated on extending SSA form to accommodate pointers[CG93, CCL ....

Rakesh Ghiya. Putting Pointer Analysis to Work. PhD thesis, McGill U., Montr'eal, Qu'e., Nov. 1997.

....p can be removed if a has already been checked on all paths that reach p, but only if no path contains a synchronization statement. Access checks to array elements that are accessed in a loop may be lifted into one aggragate array check before the loop. The compiler also performs heap analysis [11] to discover when subobjects referenced by an object are always accessed through that outer object. If this is the case, an aggregate access check is generated to fault in the outer object and all its referenced subobjects. This may greatly increase granularity, and may save a number of network ....

R. Ghiya and L. Hendren. Putting Pointer Analysis to Work. In Symp. on Principles of Programming Languages, pages 121--133, San Diego, CA, January 1998.

No context found.

R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In POPL, 1998.

No context found.

R. Ghiya and L. Hendren. Putting pointer analysis to work. In ACM Symposium on Principles of Programming Languages, New York, NY, 1998. ACM Press.

No context found.

R. Ghiya and L. Hendren. Putting pointer analysis to work. In Proceedings of the Symposium on Principles of Programming Languages, pp. 121--133, 1998.

No context found.

R. Ghiya and L. Hendren, "Putting pointer analysis to work," in Proc. of POPL, pp. 121--133, 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In Proc. 25th ACM POPL, 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In Proc. 25th ACM POPL, 1998.

No context found.

R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In 25th ACM Symp. on Principles of Programming Languages, January 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In 25th ACM Symp. on Principles of Programming Languages, January 1998. 13

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In ACM Symposium on Principles of Programming Languages (POPL), pages 121--133, San Diego, California, January 1998.

No context found.

R. GHIYA AND L. J. HENDREN, Putting pointer analysis to work, in Proc. 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, POPL '98, January 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In Proc. 25th ACM POPL, 1998.

No context found.

R. Ghiya and L. J. Hendren. Putting pointer analysis to work. In 25th ACM Symp. on Principles of Programming Languages, January 1998.

No context found.

R. Ghiya and L. Hendren. Putting Pointer Analysis to Work. In Symp. on Principles of Programming Languages, pages 121--133, San Diego, CA, Jan. 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In 25th ACM Symp. on Principles of Programming Languages, January 1998.

No context found.

Rakesh Ghiya and Laurie J. Hendren. Putting pointer analysis to work. In 25th ACM Symp. 13 on Principles of Programming Languages, January 1998.

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