N. Kobayashi. Type-based useless variable elimination. In ACM Workshop on Partial Evaluation and Semantics-Based Program Manipulation, pages 84--93, Boston, Massachusett, January 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....f h y = if (y=0) then (h 1) else f h (h y) in Source program let fun f h y = if (y=0) then (g 1) else f g (g y) in Transformed program CFA predicts that the only function that can flow to h is g, and thus h can be replaced by g. A separate transformation, useless variable elimination [22, 38, 44], could be used to remove h from f. One might expect that b reduction or b equality could be used to prove that the source and transformed programs are the same. However, the source and transformed programs are not equivalent via b reduction or b equality, because of the recursive definition of ....

....ways to ensure that the linking can be done statically. We are working on a system based on existential and universal types to capture this recursive linking process. Another potential example for using defunctionalization is useless variable elimination [38, 44] although in this case Kobayashi [22] has already given an elegant operational proof of correctness. We also conjecture that our earlier work on dependency analyses [1] can be extended to handle useless variable elimination. One might apply our techniques to study access control security, as in the SLam calculus [16] whose ....

N. Kobayashi. Type-based useless variable elimination. In ACM SIGPLAN Workshop on Partial Evaluation and Semantics-based Program Manipulation, pages 84--93, ACM, 2000.

.... procedures cannot perform this optimization as they only improve programs where program calls contain dead functors or multiple occurrences of the same variable, or the functions are de ned by rules whose rhs s are normalizable [3, 11, 12] Known procedures for removing dead code such as [7, 19, 26] do not apply to this example either. It seems interesting to formalize program analysis techniques for detecting these kinds of redundancies as well as to formalize transformations for eliminating dead code which appears in the form of redundant function arguments or useless rules and which, in ....

....CS. Let f 2 F and i 2 f1; ar(f)g. If, for all l r 2 R f , lj i is a variable which is (f; i) redundant in r, then i 2 rarg eval R (f) 6 Example 4. A standard example in the literature on useless variable elimination (UVE) a popular technique for removing dead variables, see [36, 19] is the following program : loop(a,bogus,0) loop(f(a,0) s(bogus) s(0) loop(a,bogus,s(j) a Here it is clear that the second argument does not contribute to the value of the computation. By Theorem 3, the second argument of loop is redundant w.r.t. eval R . The following example ....

[Article contains additional citation context not shown here]

N. Kobayashi. Type-based useless variable elimination. In roc. of PEPM-00, pages 84-93, ACM Press, 2000.

....The application of automatic transformation processes during the formal development and optimization of programs can introduce encumbrances in the generated code that programmers usually (or presumably) do not write. Examples are redundant arguments in the functions de ned in the program [1, 2, 7, 13, 15, 18, 20, 21, 24, 27]. Example 1. Consider the following program, which can be used for adding and substracting natural numbers in Peano s notation: minus(x,0) x plus(0,y) y minus(0,s(y) 0 plus(s(x) y) s(plus(x,y) minus(s(x) s(y) minus(x,y) Work partially supported by CICYT TIC2001 2705 C03 01, ....

....) plus0(s(x) s(plus0(x) r 8 ) minus1(s(y) minus1(y) Note that the second argument of the function minusplus is redundant for the semantics of computed values and would not typically be written by a programmer who writes this program by hand. Known procedures for removing dead code such as [7, 18, 21] as well as standard (post specialization) renaming compression procedures (see e.g. 3] cannot remove the redundant argument either. Moreover, redundant argument ltering procedures for logic programs such as the one included in the partial deduction system ECCE [19] do not recognize the ....

N. Kobayashi. Type-based useless variable elimination. In Proceedings of PEPM00, pages 84-93. ACM Press, 2000.

....non linear recursive functions, and conditional test with recursive calls. Finally, type based analyses have traditionally been used to support both program safety and optimisation. More recently, they have also been used to support program transformations, such as useless variable elimination [12, 1]. However, these type systems are still based on the evaluation rules of the underlying language. In contrast to the usual type system based on the underlying language, our PType system is constructed and proven correct from a set of metarules that are used for transforming programs into skeletal ....

Naoki Kobayashi. Type-based useless variable elimination. In ACM Workshop on Partial Evaluation and Semantics-Based Program Manipulation, pages 84--93, Boston, Massachusett, January 2000.

.... 68] ow analysis [43, 42, 7, 30, 75, 73, 15, 56, 53] trust analysis [51] secure information ow analysis [60] closure conversion [25] resource allocation in compilers [69] continuation allocation [58] dependency analysis [1] communication analysis [48] and elimination of useless variables [34, 19]. Many of them have been proved correct, most have not yet been implemented for a full edged programming language, although some have been implemented for a toy language, and some still need an algorithm for performing the analysis. Nielson and Nielson [47] present the overall methodology behind ....

Naoki Kobayashi. Type-based useless variable elimination. In Proceedings of PEPM'00, ACM Symposium on Partial Evaluation and Semantics-Based Program Manipulation, pages 84-93, 2000.

....functional programs. However, usually, the underlying language misses features such as algebraic types and polymorphism. The assumption is usually that an extension, for example to ML style polymorphism, is straightforward. Recent work, however, suggests that this is not necessarily so [15, 27]. One aim of this paper is to show that Boolean constraints give a better handle on such extensions. In this work we take binding time analysis as our example. This analysis is used for program specialisation. The purpose of binding time analysis is to identify program expressions that can be ....

....presented here have many immediate applications. Binding time analysis is essentially identifying dependencies amongst components of expressions, and this is also the heart of several other analyses. For example, this is the case for security analysis [12, 20, 25] and for useless code analysis [5, 15, 26], program slice analysis etc. Several researchers have addressed the problem of how to systematically extend a type system with dependency information [1, 9, 18, 19] The goal of these researchers is similar to ours, but generally more limited with respect to scope. F ahndrich and Rehof have ....

N. Kobayashi. Type-based useless variable elimination. In Proc. 2000 ACM SIGPLAN Workshop on Partial Evaluation and Semantics-Based Program Manipulation (SIGPLAN Notices 34(11)), pages 84-93. ACM Press, 2000.

.... similar to the one we had for the first analysis for the more direct approach of using the analysis specification as the justification of the transformation. The soundness of the analysis and the correctness of the transformation become the same. Fischbach and Hannan [FH99] and Kobayashi [Kob99] recently extended our work by presenting it as a type based inference system. This approach takes advantage of a property mentioned above: all functions that flow to the same call site agree on their dependency information. Equality constraints can be used instead of subset relations without ....

....the precision of the analysis (though this is not proved) thus giving the chance to use type based algorithms with better performance results. Their results include an algorithm with better complexity and a more succinct specification of the analysis. FH99] integrates an effect analysis and [Kob99] includes structured data i.e. pairs. Neither [FH99] nor [Kob99] provide an exact definition of useless expressions and variables. 5 Correctness of Analysis Based Transformations 5.1 Proof Architecture We are concerned with the problem of proving the correctness of analysis based ....

[Article contains additional citation context not shown here]

Naoki Kobayashi. Type-based useless variable elimination. Technical Report TR99-02, Department of Information Science, University of Tokyo, 1999. Submitted for publication.

No context found.

N. Kobayashi. Type-based useless variable elimination. Higher-Order and Symbolic Computation, 14(2-3):221--260, 2001.

No context found.

N. Kobayashi. Type-based useless variable elimination. Higher-Order and Symbolic Computation, 14(2-3):221--260, 2001.

No context found.

N. Kobayashi. Type-based useless variable elimination. In Proc. of PEPM, pages 84--93, 2000.

....more detail and also discusses extensions to deal with side e ects, equality types, etc. Section 7 concludes this paper. For the space restriction, we omit some de nitions, detailed descriptions and analyses of the algorithm, and most of the proofs. They are found in the full version of this paper [13]. 2. SYNTAX AND OPERATIONAL SEMANTICS OF THE LANGUAGE In this section, we introduce the syntax and typing of an MLstyle polymorphic, call by value functional language, which is used as the source and target language of UVE. 2.1 Terms De nition 2.1 [terms] The syntax of terms is de ned as ....

.... proj i (M) i (T Proj) M1 : 1 M2 : 2 (M1 ; M2) 1 2 (T Pair) Figure 2: Typing Rules let fun f1(x) x fun f2(x) x x val g = if p(a,b) then f1 else f2 in g(x) end (which is the same result as given in [28] by applying our method and then removing ( See the full version [13] for intermediate steps of this translation. Example 3.3: We can eliminate not only top level parameters of functions, but also a part of structured data. Consider the following program: let fun f(x, y,z) x z in f(1, 2,3) end By assigning type int (unit int) int to f, we have let fun ....

[Article contains additional citation context not shown here]

N. Kobayashi. Type-based useless variable elimination. Technical Report 99-02, Department of Information Science, University of Tokyo, July 1999. Available from http://www.yl.is.s.u-tokyo.ac.jp/~koba /publications.html.

No context found.

N. Kobayashi. Type-based useless variable elimination. In ACM Workshop on Partial Evaluation and Semantics-Based Program Manipulation, pages 84--93, Boston, Massachusett, January 2000.

No context found.

N. Kobayashi. Type-based useless variable elimination. In Proceeedings of the ACM Symposium on Partial Evaluation and Semantics-Based Program Manipulation, 2000.

No context found.

Kobayashi, N. 2000. Type-based useless variable elimination. In Proceedings of PEPM-00. ACM Press, New York, 84--93.

No context found.

N. Kobayashi. Type-based useless variable elimination. In Proceeedings of the ACM Symposium on Partial Evaluation and Semantics-Based Program Manipulation, 2000.

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