Adding magic to an optimising datalog compiler

Sereni, Damien; Avgustinov, Pavel; Moor, Oege de

doi:10.1145/1376616.1376673

Cited by 17 publications

(10 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, the performance of different tabling strategies vary drastically [21], and bottom-up evaluation after MST may be much slower than the bottom-up evaluation of the original rules [23]. Choosing the best evaluation method for a given set of rules requires precise characterization of the time and space complexities of each evaluation method.…”

Section: Introductionmentioning

confidence: 99%

More efficient datalog queries

Tekle

Liu

2011

Proceedings of the 2011 ACM SIGMOD International Conference on Management of Data

View full text Add to dashboard Cite

Given a set of Datalog rules, facts, and a query, answers to the query can be inferred bottom-up starting with the facts or top-down starting with the query. The dominant strategies to improve the performance of answering queries are reusing answers to subqueries for top-down methods, and transforming rules based on demand from the query, such as the well-known magic sets transformation, for bottomup methods. However, the performance of these strategies vary drastically, and the most effective method remained unknown.This paper describes precise time and space complexity analysis for efficient implementation of Datalog queries using subsumptive tabling, a top-down evaluation method with more reuse of answers than the dominant tabling strategy, and shows that subsumptive tabling beats bottom-up evaluation of rules after magic sets transformation in both time and space complexities. It also describes subsumptive demand transformation, a novel method for transforming the rules so that bottom-up evaluation of the transformed rules mimics subsumptive tabling; we show that the time complexity of bottom-up evaluation after this transformation is equal to the the time complexity of top-down evaluation with subsumptive tabling. The paper further describes subsumption optimization, an optimization to increase the use of subsumption in subsumptive methods, and shows its application in the derivation of a well-known demand-driven pointer analysis algorithm. We support our analyses and comparisons through experiments with applications in ontology queries and program analysis.

show abstract

Section: Introductionmentioning

confidence: 99%

More efficient datalog queries

Tekle

Liu

2011

Proceedings of the 2011 ACM SIGMOD International Conference on Management of Data

View full text Add to dashboard Cite

show abstract

“…For LogicBlox (v 3.7), in the Opt column, we additionally measure the impact of the system's optimizer [17] aimed at improving the performance of equalitybased joins by reordering the goals and applying a variant of magic-set rewrite.…”

Section: Discussionmentioning

confidence: 99%

Optimizing Inequality Joins in Datalog with Approximated Constraint Propagation

Campagna

Sarna-Starosta²,

Schrijvers

2012

Practical Aspects of Declarative Languages

View full text Add to dashboard Cite

Abstract. Datalog systems evaluate joins over arithmetic (in)equalities as a naive generate-and-test of Cartesian products. We exploit aggregates in a source-to-source transformation to reduce the size of Cartesian products and to improve performance. Our approach approximates the well-known propagation technique from Constraint Programming. Experimental evaluation shows good run time speed-ups on a range of non-recursive as well as recursive programs. Furthermore, our technique improves upon the previously reported in the literature constraint magic set transformation approach.

show abstract

“…2) the query language is object-oriented, making it easy to create libraries of concepts that hide much implementation detail [2]; 3) special optimisations enable fast execution [12]- [14]; 4) Semmle incorporates extremely robust and precise fact extractors that handle all aspects of C, C++, Java, C ♯ and XML. These extractors have been deployed at Semmle's customer base on hundreds of millions of lines of code.…”

Section: A Related Workmentioning

confidence: 99%

Automated architectural reviews with Semmle

Schutter¹

2012

2012 28th IEEE International Conference on Software Maintenance (ICSM)

View full text Add to dashboard Cite

Keeping code at a high level of quality and in sync with the defined architecture is not a trivial matter when you are participating in a competitive market in which you would rather provide solutions today than tomorrow. Code checkers help, but most are too generic to be of real value and so end up being ignored. Customisable code checkers can provide real benefit at reasonable cost. By encoding company specific (or even project specific) rules, actual results become much more relevant. With the additional integration of UML diagrams and specifications into the code checker, it even becomes possible to manage the gap between the actual code and the architecture from day to day at reasonable cost.

show abstract

Adding magic to an optimising datalog compiler

Cited by 17 publications

References 20 publications

More efficient datalog queries

More efficient datalog queries

Optimizing Inequality Joins in Datalog with Approximated Constraint Propagation

Automated architectural reviews with Semmle

Contact Info

Product

Resources

About