J. C. KERGOMMEAUX and P. CODOGNET. Parallel logic programming systems. Computing Surveys, 26(3):295--336, September 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and other control information only need a constant number of machine words. And, if we initially broadcast the root problem to all PEs, subproblems can often be represented by sequences of O(h) bits indicating how to derive the subproblem from the root problem by subsequent split operations (e.g. [11]) Under the usual assumption that O(log n) bits fit into a machine word our assumption is strictly justified for the frequent case [14] h 2 O(log T seq ) and T seq polynomial in n (Larger problems are easy to load balance anyway) Still, in other cases it is better to treat the message length as ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

.... randomized load balancing algorithm 1 Introduction Many algorithms in operations research and artificial intelligence are based on the backtracking principle for traversing large irregularly shaped trees [8, 9, 15, 13, 19] Similar problems also play a role in parallel programming languages [1, 12] and even for loop scheduling and some numerical problems like adaptive numerical integration it can be useful to view the computations as an implicitly defined tree. Section 2 introduces the abstract model of tree shaped computations which makes the common properties of these applications visible ....

....5, 24] Random polling belongs to a family of receiver initiated load balancing algorithms which have the advantage to split subproblems only on demand by idle PEs. This adaptive approach has been used successfully for a variety of purposes such as parallel functional [1] and logic programming [12] or game tree search [8] Randomized partner selection goes at least back to [9] The partner selection strategy turns out to be crucial. The apparently economic option to poll neighbors in the interconnection network can be extremely inefficient since it leads to a buildup of clusters of busy ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

....easy problems which produce a very irregular search tree and only a moderate amount of work. 1 Introduction Many algorithms in operations research and artificial intelligence are based on the backtracking or depth first traversal principle for traversing large implicitly defined trees (e.g. [1, 11, 12, 14, 16, 15, 18]) In addition, some adaptive numerical algorithms for integration [21] for finding eigenvalues of tridiagonal matrices [7] or for nonlinear optimization [20] have a similar structure. Even modeling the seemingly unrelated problem of loop scheduling in this way can be advantageous [30] In ....

....in [28] for a more detailed discussion. Early work on random polling and an application independent library is described in [12] Random polling and other receiver initiated load balancing methods are also of central importance for parallel functional and logical programming languages (e.g. [1, 14]) Tree shaped computations can be considered a generalization of the ff splitting model used in [16] A related model based on multithreaded computations is used in the Cilk project [4, 3, 2] The ZRAM library [6] is another recent implementation effort. 2 The Abstract Model All the work to be ....

[Article contains additional citation context not shown here]

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

.... in operations research and artificial intelligence are based on the backtracking principle for traversing large irregularly shaped trees that are only defined implicitly by the computation [3, 4, 6, 9, 12 14, 19, 17, 21, 35] Similar problems also play a role in parallel programming languages [1, 16]. Even for loop scheduling and some numerical problems [7, 24] like adaptive numerical integration [25] it can be useful to view the computations as an implicitly defined tree (refer to [34] for a more detailed discussion of examples) For parallelizing tree shaped computations, a load balancing ....

....7, 27] Random polling belongs to a family of receiver initiated load balancing algorithms which have the advantage to split subproblems only on demand by idle PEs. This adaptive approach has been used successfully for a variety of purposes such as parallel functional [1] and logic programming [16] or game tree search [12] Randomized partner selection goes at least back to [13] The partner selection strategy turns out to be crucial. The apparently economic option to poll neighbors in the interconnection network can be extremely inefficient since it leads to a buildup of clusters of busy ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

....and other control information only need a constant number of machine words. And, if we initially broadcast the root problem to all PEs, subproblems can often be represented by sequences of O(h) bits indicating how to derive the subproblem from the root problem by subsequent split operations (e.g. [12]) Under the usual assumption that O(log n) bits fit into a machine word our assumption is strictly justified for the frequent case [15] h 2 O(log T seq ) and T seq polynomial in n (Larger problems are easy to load balance anyway) Still, in other cases it is better to treat the message length as ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

....remains useful in the presence of some frequently observed subproblem dependencies. 1 Introduction Many algorithms in operations research and artificial intelligence are based on the backtracking or depth first principle for traversing large implicitly defined trees (e.g. ABF93, FMM94, FM87, KC94, KGGK94, Kor85, MT90] In addition, some adaptive numerical algorithms for numerical integration by [PTVF92] for finding eigenvalues of tridiagonal matrices [CRY94] or for nonlinear optimization [NM96] have a similar structure. We will see, that even modeling the seemingly unrelated problem of ....

....for a more detailed discussion. Early work on random polling and an application independent library is described in [FM87] Random polling and other receiver initiated load balancing methods are also of central importance for parallel functional and logical programming languages (e.g. ABF93, KC94] Tree shaped computations can be considered a generalization of the a splitting model used in [KGGK94] A related model based on a subclass of multithreaded computations is used in the Cilk project [BL94, BFJ 96] The ZRAM library [BMFN98] is another recent implementation effort. We ....

[Article contains additional citation context not shown here]

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

....(ii) the same side effects are observed, in the same order, as in sequential execution. In this paper, we only consider parallel implementations that preserves Prolog semantics. 2 Problems in Parallel Logic Programming Various parallel implementations of Prolog have been proposed and realized [3, 10]. Most of the current systems suffer from severe efficiency problems. Many of these inefficiencies, especially in and parallel systems, arise from the need of guaranteeing Prolog semantics during parallel execution. Other inefficiencies, such as those arising in or parallelism, are present due to ....

J.C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3), 1994.

....of a logic program, the set of goals that remain to be executed, i.e. the continuation of the computation, is called the current resolvent. The execution of a logic program starts by taking the query as initial resolvent and proceeds by transforming the current resolvent into a new one as follows [57]. First select any goal G of the resolvent, and apply a resolution step, which consists in finding a clause of the program whose head unifies with G1, then replacing G by the body of the unifying clause to produce a new resolvent. The bindings of variables resulting from the unification apply to ....

....to the program using assert) In this text, Prolog semantics refers to the ordering of the exploration of the search tree. Extra logical predicates whose behaviour depends on the order of execution are said order sensitive predicates. 3 Parallelism in Logic Programming Parallel logic programming [57, 17, 46, 73, 100] is an important research field whose history is closely related to logic programming languages evolution. The main goal in having parallel logic programming environments is to allow programmers to develop faster applications that are suitable to be developed in a declarative language such as ....

[Article contains additional citation context not shown here]

J. C. KERGOMMEAUX and P. CODOGNET. Parallel logic programming systems. Computing Surveys, 26(3):295--336, September 1994.

....and other control information only need a constant number of machine words. And, if we initially broadcast the root problem to all PEs, subproblems can often be represented by sequences of O(h) bits indicating how to derive the subproblem from the root problem by subsequent split operations (e.g. [11]) Under the usual assumption that O(log n) bits fit into a machine word our assumption is strictly justified for the Peter Sanders: Better Algorithms for Parallel Backtracking 21 frequent case [14] h 2 O(log T seq ) and T seq polynomial in n (Larger problems are easy to load balance anyway) ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

....and other control information only need a constant number of machine words. And, if we initially broadcast the root problem to all PEs, subproblems can often be represented by sequences of O(h) bits indicating how to derive the subproblem from the root problem by subsequent split operations (e.g. [12]) Under the usual assumption that O(log n) bits fit into a machine word our assumption is strictly justified for the frequent case [15] h 2 O(log T seq ) and T seq polynomial in n (Larger problems are easy to load balance anyway) Still, in other cases it is better to treat the message length as ....

J. C. Kergommeaux and P. Codognet. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

No context found.

J. C. KERGOMMEAUX and P. CODOGNET. Parallel logic programming systems. Computing Surveys, 26(3):295--336, September 1994.

No context found.

) Kergommeaux, J.C. and Codognet, P. Parallel logic programming systems. ACM Computing Surveys, 26(3):295--336, 1994.

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