L. V. Kale, Parallel Architectures for Problem Solving, PhD Thesis, SUNY, Stony Brook, 1985. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
The Design of an Optimistic AND-Parallel Prolog - Olthof, Cleary (1993) (Correct)
....goals that share variables are run sequentially, usually using some form of backtracking to produce all solutions. Parallelism is thus restricted to goals known to be independent. Such knowledge may be gained statically, dynamically, or by a combination of both. Kale s REDUCE OR Process Model [Kale 1985] uses static analysis to produce annotations (specifically, data join graphs) to control parallel execution. Such analysis must be conservative to avoid pitfalls like aliased variables that could result in dependent goals running in parallel. Conery s AND OR process model [Conery 1987] uses ....
L.V. Kale. Parallel Architectures for Problem Solving. PhD thesis, Department of Computer Science, SUNY Stony Brook, 1985.
Paged Binding Array: Environment Representation for And-Or.. - Gupta (1991) (Correct)
....brevity)y. Thus the extended tree for the above query would appear as shown in figure 2 for each alternative of and parallel goal a, goal b is computed in its entirety. This extended tree, which we call the Composition tree (C tree for brevity) is somewhat similar to the Reduce Or tree of Kal e [K85]. Note that each composition node in the tree corresponds to a different solution for the CGE. To represent the fact that a CGE can have multiple solutions we add a branch point (choice point) before the different composition nodes. The C tree can represent or and independent and parallelism ....
L. V. Kale, "Parallel Architectures for Problem Solving", Ph.D. Thesis, Dept. of Computer Science, SUNY-Stony Brook, 1985.
A New Implementation Scheme for Combining And/Or Parallelism - Shen (Correct)
....been a lot of interest in the implicit parallel execution of Logic Programming languages, especially of Prolog, and many schemes which exploit various forms of and and or parallelism have been proposed and implemented. Although many of the early proposals exploited both forms of parallelism (e.g.[9, 20, 36, 11, 15]) many of the more recent schemes, especially those that have been successfully implemented, have tended to concentrate on either and or or parallelism (e.g. 21, 1, 18, 5, 28] probably because of the complexities and difficulties in efficiently implementing both forms of parallelism ....
.... (e.g. 9, 11] did not have reusage, due to the severe restrictions on how and and or parallelism can be combined (and thus what would be available simultaneously) However, for all early schemes which did have sufficient freedom, reusage seemed so attractive that it was always included (e.g. [36, 20, 35, 15]) 3 As part of a study of IAP and or parallelism [29] I studied goal reusage by comparing it with an alternative scheme which did not implement reusage. This involved the development of an execution scheme that allowed great freedom in how and and or parallelism can be combined, but which ....
L. V. Kale. Parallel Architectures for Problem Solving. PhD thesis, State University of New York at Stony Brook, 1985.
Fully Distributed, AND/OR Parallel Execution of Logic Programs - Raman, Stark (1988) (Correct)
....in Sigma and Sigma n is collected, and subsumption is applied to reduce its size. The resulting set is the new set of alternatives for the AND node. 1.2 Related Work We classify previous work in parallel interpreters for logic programs on the basis of the ANDparallelism each supports. In [CK85, LM86, Kal86] AND parallelism is restricted to independent goals (those that do not share variables) In [Wis86, KL87] AND goals are evaluated bottom up without any exchange of information between the AND parallel goals. The committed choice languages of [CG86, Sha83] evaluate AND goals in parallel with ....
....data driven model in which solutions to literals in the body of a clause are piped between AND parallel processes in accordance with the dependency graph. Kale [Kal87] has pointed out that both methods [CK85] and [LM86] are incomplete, the authors claims to the contrary notwithstanding. In [Kal86], Kale has given a modified version, which is apparently complete. In [CDD85, Her86] it is shown how the expense of maintaining the dependency graphs can be reduced by compile time preprocessing. In [Wis86] and [KL87] AND parallelism is supported by computing answers in a bottom up fashion, and ....
L.V. Kale. Parallel Architectures for Problem Solving. PhD thesis, SUNY Stony Brook, 1986.
A Parallel Prolog Compiler and its Implementation - Wenfeng Li (1993) (Correct)
....which is responsible for determining the next literal to evaluate based on the evaluation outcome of the previous one, and perhaps on other information such as control heuristics and resource availability in the system. This approach had been taken by many systems including the reduced Or Model [9], and PRISM [8] The runtime driver approach offers flexibility in literal selection (which is not necessary for Prolog, but is desirable for logic programming in general) However, there appears to be a performance penalty associated with the runtime driver approach. Our design adopts the runtime ....
L. V. Kale. Parallel architectures for problem solving. Technical Report UIUCDCS-R-85-1237, Department of Computer Science, University of Illinois at Urbana-Champaign, 1985.
Task Granularity Analysis in Logic Programs - Saumya Debray Nai-Wei (1990) (20 citations) (Correct)
No context found.
L. V. Kale, Parallel Architectures for Problem Solving, PhD Thesis, SUNY, Stony Brook, 1985.
Mapping Search Graphs Onto Arbitrary Processor Networks (or.. - Douglas Eadline (Correct)
No context found.
Kale, L. V., Parallel Architectures for Problem Solving, Ph.D Dissertation, Computer Science, SUNY, Stony Brook, December 1985.
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