Jyh-Herng Chow and Williams Ludwell III Harrison. Compile time analysis of parallel programs that share memory. In Proceedings of the Nineteenth Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL '92), pages 130-141. ACMSIGPLAN, January 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....with the reduction of communication volume. 5 Related Work Research on analyses or optimizations for OpenMP programs is still rare but there are several articles on other explicit parallel programs. General frameworks for analyzing explicit parallel programs are presented by Chow and Harrison[2], Grunwald and Srinivasan[5] and Ferrante et al. 4] In their work, ####### ##### parallelism[2] event synchronizations[5] and ##### loops[4] are considered. In addition, Grunwald and Srinivasan[5] point out the advantages of relaxed memory consistency for compiler optimizations. Static single ....

....for OpenMP programs is still rare but there are several articles on other explicit parallel programs. General frameworks for analyzing explicit parallel programs are presented by Chow and Harrison[2] Grunwald and Srinivasan[5] and Ferrante et al. 4] In their work, ####### ##### parallelism[2], event synchronizations[5] and ##### loops[4] are considered. In addition, Grunwald and Srinivasan[5] point out the advantages of relaxed memory consistency for compiler optimizations. Static single assignment(SSA) form for parallel programs are proposed by Srinivasan et al. 25] Novillo et al. ....

J-H.Chow and W.L.Harrison. Compile-time Analysis of Parallel Programs that Share Memory. POPL'92, pp. 130-141, 1992.

....than reducing the communication volume. 5 Related Work Research on analyses or optimizations of OpenMP programs is still rare, but there are several articles on other explicit parallel programs. General frameworks for analyzing explicit parallel programs were presented by Chow and Harrison[2], Grunwald and Srinivasan[5] and Ferrante et al. 4] In their work, cobegin coend parallelism[2] event synchronizations[5] and doall loops[4] are considered. In addition, Grunwald and Srinivasan[5] point out the advantages of relaxed memory consistency for compiler optimizations. Static single ....

....of OpenMP programs is still rare, but there are several articles on other explicit parallel programs. General frameworks for analyzing explicit parallel programs were presented by Chow and Harrison[2] Grunwald and Srinivasan[5] and Ferrante et al. 4] In their work, cobegin coend parallelism[2], event synchronizations[5] and doall loops[4] are considered. In addition, Grunwald and Srinivasan[5] point out the advantages of relaxed memory consistency for compiler optimizations. Static single assignment (SSA) form for parallel programs was proposed by Srinivasan et al. 25] Novillo et al. ....

J-H.Chow and W.L.Harrison. Compile-time Analysis of Parallel Programs that Share Memory. POPL'92, pp. 130-141, 1992.

....and Pnueli [1981] use call strings to encode the interprocedural call context and discriminate different instances of the same procedure. Jones and Muchnick [1982] develop a more general abstraction of the call stack, which is used to name activation records. Other methods [Harrison III 1989; Chow and Harrison 1992] present a naming technique for dynamically allocated objects based on procedure strings, a generalization of call strings that includes the call stack and its evolution. Deutsch [1990] names dynamically allocated objects based on an arbitrary abstraction of the machine state at allocation time. ....

Chow, J. H. and Harrison, W. L. 1992. Compile-time analysis of parallel programs that share memory. In 19th Annual ACM Symposium on the Principles of Programming Languages. 130-- 141.

....7.2 Analysis of Explicitly Parallel Programs The transformations presented in this article operate on explicitly parallel programs. Other researchers have investigated the issues associated with performing standard serial compiler analyses and optimizations in the presence of explicit concurrency [Chow and Harrison III 1992; Midkiff and Padua 1990] Our research is orthogonal to this research in the sense that it focuses on optimization opportunities that appear only in explicitly parallel programs rather than on the significant challenges associated with applying standard optimizations to parallel programs. 7.3 ....

CHOW, J. AND HARRISON III, W. 1992. Compile-time analysis of parallel programs that share memory. In Proceedings of the 19th Annual ACM Symposium on the Principles of Programming Languages. 130--141.

....compilers are designed to work with, compilers can apply the same analysis techniques and transformations to Jade programs that they apply to serial programs. Explicit parallelism, on the other hand, can dramatically reduce the compiler s ability to analyse the program and apply transformations (Chow and Harrison 1992). 2 The Jade programming language Jade programmers start with a serial program and use Jade constructs to specify how parts of the program access data. The Jade implementation dynamically analyses this information to automatically extract the concurrency present in the application. For ....

Chow, J. and Harrison, W. III (1992) Compile-time analysis of parallel programs that share memory. In Proceedings of the Nineteenth Annual ACM Symposium on the Principles of Programming Languages, January, pp. 130--41.

.... Recent work discusses how to transform these multisource data flow problems into a k tuple framework formulation [MMR95] This model is especially useful for data flow analysis of explicitly parallel programs, where there may be control, synchronization, and call and return edges [MR93, GS93, CH92] Bidirectionality Bidirectional data flow problems were introduced by Morel and Renvoise for Partial Redundancy [MR79, Dha91] a generalization of Available Expressions) the approach has been extended to other problems [DP93] In [KRS92] it is shown how to solve Partial Redundancy as a ....

J-H Chow and W. L. Harrison III. Compile-time analysis of parallel programs that share memory. In Conference Record of the Nineteenth Annual ACM Symposium on Principles of Programming Languages, pages 130--141, January 1992.

....of a parallel program. Thus, either heuristics are proposed to avoid the consideration of all the interleavings [McD] or restricted situations are considered, which do not require to consider the interleavings at all. To our knowledge, the most relevant publications in this direction are [CH, GS]. In [CH] a situation without synchronization is considered, and in [GS] data independence of parallel components is required. Thus the result of a parallel execution does not depend on the particular choice of the interleaving. In [GS] this is exploited for the construction of an optimal and ....

....parallel program. Thus, either heuristics are proposed to avoid the consideration of all the interleavings [McD] or restricted situations are considered, which do not require to consider the interleavings at all. To our knowledge, the most relevant publications in this direction are [CH, GS] In [CH] a situation without synchronization is considered, and in [GS] data independence of parallel components is required. Thus the result of a parallel execution does not depend on the particular choice of the interleaving. In [GS] this is exploited for the construction of an optimal and efficient ....

Chow, J.-H., and Harrison, W. L. Compile time analysis of parallel programs that share memory. In Conference Record of the 19 th International Symposium on Principles of Programming Languages (POPL'92), Albuquerque, New Mexico, 1992, 130 - 141.

....like to thank R. Cousot also for her continued impulse and support. 8 Related work Seminal papers about static analysis of parallel programs include [BDER79] and [Tay93] The later focuses on the ADA programming language. Recent work by Chow and Harrison uses an intermediate Lisp like language [CI92]. CI94] stresses upon the reduction of the generated state space using Valmari s method of stubborn sets [Val89] ....

J.-H. Chow and W. L. Harrison III. Compile-time Analysis of Parallel Programs that Share Memory. In LISP and Functional Programming, June 1992.

.... heuristics are proposed to avoid the consideration of all the interleavings [McD] or restricted situations are considered, which do not require to consider the interleavings at all [GS] 1 Completely different is the approach of abstract interpretation based state space reduction proposed in [CH1, CH2]. This, however, requires the construction of an appropriately reduced version of the global state space, which is often still unmanageable. In [KSV1] we have recently demonstrated that for the large class of bitvector analyses , which are most relevant in practice, there is an elegant way to ....

Chow, J.-H., and Harrison, W. L. Compile time analysis of parallel programs that share memory. In Conference Record of the 19 th International Symposium on Principles of Programming Languages (POPL'92), Albuquerque, New Mexico, 1992, 130 - 141.

....and last but not least C. Colby for his very helpful proofreading. 9 Related work Seminal papers about static analysis of parallel programs include [BDER79] and [Tay93] The latter focuses on the ADA programming language. Recent work by Chow and Harrison uses an intermediate Lisp like language [CH92]. CH94] stresses upon the reduction of the generated state space using Valmari s method of stubborn sets [Val89] There has been much interest recently about abstractions preserving fragments of modal logic: DGG94] presents abstractions preserving 8CTL , 9CTL and a combination for full ....

J.-H. Chow and W. L. Harrison III. Compile-time Analysis of Parallel Programs that Share Memory. In LISP and Functional Programming, June 1992.

....Using the intensional representation of parallel algorithms it should be possible to obtain analyses of higher order parallel programs. I would like to consider some of the following properties: ffl Data dependence. This is one of the analyses that are usually attempted for parallel programs [13, 31], although with an imperative flavor. The parallel algorithm representation should be ideally suited for this. ffl Extra parallelism. We could find out if certain arguments are needed (strictness analysis) and if they can be evaluated in parallel. ffl Selective communication. In a language with ....

J.-H. Chow and W.L. Harrison III, Compile-time analysis of parallel programs that share memory, in: Proc. Principles of Programming Languages, 1992 , 130--141.

....does not depend on the particular choice of the interleaving, which is exploited for the construction of an optimal and efficient algorithm determining the reaching definition information. Completely different is the approach of abstract interpretation based state space reduction proposed in [CH1, CH2] which allows general synchronization mechanisms but still requires the construction of an appropriately reduced version of the global state space which is often still unmanageable. In this paper we show how to construct arbitrary bitvector analysis algorithms for parallel programs with ....

....consider a parallel imperative programming language with an interleaving semantics. Formally, this means that we view parallel programs semantically as abbreviations of usually much larger nondeterministic programs, which result from a product construction between parallel components (cf. CC2, CH1, CH2] In fact, in the worst case, the size of the nondeterministic product program grows exponentially in the number of parallel components of the corresponding parallel program. This immediately clarifies the dilemma of data flow analysis for parallel programs: even though it can be reduced ....

Chow, J.-H., and Harrison, W. L. Compile time analysis of parallel programs that share memory. In Conference Record of the 19 th International Symposium on Principles of Programming Languages (POPL'92), Albuquerque, New Mexico, 1992, 130 - 141.

....7.2 Analysis of Explicitly Parallel Programs The transformations presented in this paper operate on explicitly parallel programs. Other researchers have investigated the issues associated with performing standard serial compiler analyses and optimizations in the presence of explicit concurrency [3, 14]. Our research is orthogonal to this research in the sense that it focuses on optimization opportunities that appear only in explicitly parallel programs rather than on the significant challenges associated with applying standard optimizations to parallel programs. 7.3 Concurrent Constraint ....

J Chow and W. Harrison III. Compile-time analysis of parallel programs that share memory. In Proceedings of the Nineteenth Annual ACM Symposium on the Principles of Programming Languages, pages 130--141, January 1992.

.... analysis, testing, and debugging of Ada like concurrent programs [29, 31, 25, 20, 27, 14] There has also been considerable work in analyzing parallel programs for the detection of deadlock and race conditions [1, 6, 5] debugging [16, 10, 18, 21, 19] and optimizing analysis and transformation [8, 4, 24, 12, 22, 9]. However, little work has focused on the automatic generation of test cases for the testing of parallel programs, particularly with a program based approach. Multiprocessors are being used increasingly as desktop workstations, server platforms, and parallel computing resources. Simultaneously, ....

....both with and without synchronization operations, such as post, wait, or advance. In the paper [12] Srinivasan and Wolfe describe semantics for parallel constructs, and introduce new intermediate forms, called the Parallel Control Flow Graph, and the Parallel Precedence Graph. Chow and Harrison[4] present a data flow framework for analyzing parallel programs with cobegin coend parallelism without explicit synchronization. Long and Clarke[17] develop a data flow framework for Ada programs with task parallelism and rendezvous synchronization, but they do not handle shared variables. Dwyer ....

J.-H. Chow and W. Harrison. Compile time analysis of parallel programs that share memory. In Proceedings of the Nineteenth Annual ACM Symposium on Principles of Programming Languages, 1992.

....semantics is assumed. An intuitive but not formal derivation of the data flow equations is given, which solve only the reaching definition problem. 4] extends abstract interpretation to cope with communicating sequential processes. The problem there is that the resulting state space explodes. [3] applies abstract interpretation to analysis of parallel programs, but bases the semantics on a labeled transition system. 5] attacks this problem using the stubborn set theory [21] which decreases the state set using some heuristics. Hence the analysis is accurate for some examples, and less ....

J.H. Chow, W.L. Harrison III. Compile-time analysis of parallel programs that share memory. In 19. ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pages 130--141, 1992.

No context found.

Jyh-Herng Chow and Williams Ludwell III Harrison. Compile time analysis of parallel programs that share memory. In Proceedings of the Nineteenth Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL '92), pages 130-141. ACMSIGPLAN, January 1992.

No context found.

J.-H. Chow and W. L. Harrison. Compile time analysis of parallel programs that share memory. In Conference Record of the 19th Annual Symposium on Principles of Programming Languages (POPL'92), pages 130 -- 142. ACM, New York, 1992.

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