Gupta, G., Pontelli, E. et al. 1994. ACE: And/Or-parallel Copying-based Execution of Logic Programs. In Proc. ICLP'94, 93--109. MIT Press.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....q1 q2 q3 c3 p q p3 q1 q2 q3 D1 D2 (b) Equivalent OR tree p1 p2 p3 q q q q1 q1 q1 q2 q2 q2 q3 q3 q3 (a) Composition Tree Fig. 2. Composition Tree and its equivalent OR tree 3 Recomputation in Composition Tree After the OR forest [19] was proposed, other AND OR parallel execution models, e.g. ACE [7], were proposed based on Composition Tree. They are variants of solution combination based models, which were proposed in order to solve the side effects problem while executing Prolog in AND OR parallel. The Composition Tree essentially uses the idea of recomputing independent goals of a parallel ....

G. Gupta, M. Hermenegildo: "ACE: And/Or-parallel Copying-based Execution of Logic Programs," In Proc. ICLP91 Workshop on Parallel Execution of Logic Programs, Lecture Notes in Computer Science 569, Springer Verlag, 1991.

....existing Prolog based NLP applications can also be parallelized, by porting them with little or no effort to parallel systems. Fast parallel implementations of Prolog are currently available (e.g. Ali Karlsson 1990) or about to be released into the public domain, including the ACE system (Gupta et al. 1994) used in this work. In this work we present two case studies in which two large NLP applications, independently developed with no goal of parallelization in mind, have been studied and parallelized using the ACE parallel Prolog system. The parallelization effort was very limited in both cases and ....

....one goal is present in the query or in the body of a clause, and at runtime these goals to do not compete for any unbound variable. iii) Dependent and parallelism (DAP) arises when two or more goals of a clause access common variables and are executed in parallel. The ACE System The ACE model (Gupta et al. 1994; Pontelli 1997) uses stack copying (Ali Karlsson 1990) and recomputation (Gupta et al. 1994) to efficiently support combined or and independent and parallel execution of logic programs. ACE represents an efficient combination of or and independent and parallelism in the sense that penalties ....

[Article contains additional citation context not shown here]

Gupta, G., Pontelli, E. et al. 1994. ACE: And/Or-parallel Copying-based Execution of Logic Programs. In Proc. ICLP'94, 93--109. MIT Press.

....existing Prolog based NLP applications can also be parallelized, by porting them with little or no effort to parallel systems. Fast parallel implementations of Prolog are currently available (e.g. Ali Karlsson 1990) or about to be released into the public domain, including the ACE system (Gupta et al. 1994) used in this work. In this work we present two case studies in which two large NLP applications, independently developed with no goal of parallelization in mind, have been studied and parallelized using the ACE parallel Prolog system. The parallelization effort was very limited in both cases and ....

....body of a clause, and at runtime these goals to do not compete for any unbound variable. iii) Dependent and parallelism (DAP) arises when two or more goals of a clause access common variables and are executed in parallel. The ACE System The ACE (And Or Parallel Copying based Execution) model (Gupta et al. 1994; Pontelli 1997) uses stackcopying (Ali Karlsson 1990) and recomputation (Gupta et al. 1994) to efficiently support combined orand independent and parallel execution of logic programs. ACE represents an efficient combination of or and independent and parallelism in the sense that penalties for ....

[Article contains additional citation context not shown here]

Gupta, G., Pontelli, E. et al. 1994. ACE: And/Orparallel Copying-based Execution of Logic Programs. In Proc. ICLP'94, 93--109. MIT Press.

.... goal over their reuse; section 3 presents an abstract model called the Composition tree for representing and or parallel execution of Prolog with recomputation; Section 4 presents reasons for favouring binding arrays method over other environment representation techniques (notably stack copying [GH91]) Section 4 also presents an extension of the Binding arrays method for and or parallel execution of Prolog programs based on the composition tree. Section 5 describes an abstract machine for compiled execution of and or parallel Prolog based on the Composition tree and extended Binding Arrays; ....

....per processor are better. With this observation in mind we have extended both types of environment representation schemes to accommodate independent and parallelism with recomputation of goals. In a separate paper we report how Muse can be extended to incorporate independent and parallelism [GH91]. In this paper we show how binding arrays can be extended for independent and parallelism. 4.2 The Fragmentation Problem in Binding Arrays Recall that in the binding array method [W84, W87] an offset counter is maintained for each branch of the or parallel tree for assigning offsets to ....

G. Gupta and M. Hermenegildo, "ACE: And/Or-parallel Copying-based Execution of Logic Programs," Technical Report TR-91-25, Department of Computer Science, University of Bristol, Oct. 1991.

....is readily found in applications that involve producer consumer interactions. So far researchers have been focusing their attention on devising systems that exploit only a single form of parallelism [AK90, LWH90, HG90, S87] or combining or parallelism with independent and parallelism [GS92, GH91, RK89, WR87] or dependent and parallelism [SWY91c] However, recently proposals have been made for exploiting all three sources of parallelism in a single framework, such as the Extended Andorra Principle [W90, HJ90] which allows for exploitation of all three forms of parallelism in an integrated framework. ....

....in its entirety for every solution of an independent goal to the left. However, we have argued that although adopting the former approach leads to fewer inferences being performed, it requires complex machinery to implement it and, further, supporting Prolog as a user language becomes trickier [GH92, GS92, GH91]. We therefore advocate an approach where even in the presence of or parallelism independent goals to the right are recomputed, just as in sequential Prolog. Taking the approach of re computing independent goals to the right in a CGE, we have developed the notion of a composition tree to ....

G. Gupta, M. Hermenegildo, "ACE: And/Or-parallel Copying-based Execution of Logic Programs," Lecture Notes in Computer Science 569, Springer Verlag.

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