Daniel Lieuwen. Ph.D Thesis. Optimizing and Parallelizing Loops in Object-Oriented Database Programming Languages. University of Wisconsin (1992). Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Exploiting Reflection to Add Persistence and Query Optimization .. - Kutlu, Moss (1998) (1 citation) (Correct)
....new functionalities in programming languages, such as atomic types [Stroud Wu, 1995] and fault tolerance [Fabre et al. 1995] We take advantage of compiler optimization techniques to build a query optimizer. In particular, we use loop optimizations to transform join loops in a query. Lieuwen [Lieuwen, 1992] has used similar techniques to transform loops in an object oriented database programming language. The use of Java s dynamic, late binding nature as an opportunity for improving program performance is similar to the approach taken in Self [Ungar Hozle, 1987] In Self, methods that are used ....
....it for loop invariant code elimination, index identification, loop tiling, etc. In particular, we reorder nested loops in order to take advantage of relational join optimization algorithms available. This is possible, it turns out, when the loop statement(s) satisfy self commutativity constraints [Lieuwen, 1992]. Consider the following simple groupby loop (expressed in pseudo code) for (x 1 of Set 1 ) suchthat (P 1 (x 1 ) for (x m of Set m ) suchthat (P m (x 1 , x m ) Stmt; The statement Stmt is defined to be self commutative relative to x 1 , x 2 , x m if the code segment above ....
D. F. Lieuwen. Optimizing and Parallelizing Loops in Object-Oriented Database Programming Languages. PhD thesis, University of Wisconsin --- Madison, 1992.
Parallel Pointer-based Join Techniques for Object-Oriented .. - Lieuwen, DeWitt, Mehta (1993) (5 citations) Self-citation (Lieuwen) (Correct)
....attribute is a key. Thus, Probechildren and Hybrid hash node pointer can be used for parallel relational database systems that support range and or hash partitioning. We will ignore statement S21 in query (B) in the following discussion. S21 will be executed once for each result tuple produced. [LIEU92b] discusses the conditions under which the join algorithms can be modified to execute S21 at the join nodes. Otherwise, S21 must be executed centrally by the application program. Leaving out S21 will make it clearer that the algorithms are equally applicable to OODBSs and DBPLs. 3.1. Hash loops ....
....a large number of oids relative to n. For instance, for n=32, if each setvalued attribute contains 32 oids, a i r is expected to be 1.57 even if no clustering is attempted and references are randomly distributed across the n nodes. that employs our algorithms (and our program transformations [LIEU92a, LIEU92b]) 7. ....
Daniel Lieuwen. Ph.D Thesis. Optimizing and Parallelizing Loops in Object-Oriented Database Programming Languages. University of Wisconsin (1992).
Pointer-based Join Techniques for Object-Oriented Databases - Lieuwen, DeWitt, Mehta (1992) (12 citations) Self-citation (Lieuwen) (Correct)
....attribute is a key. Thus, Probe children and Hybrid hash node pointer can be used for parallel relational database systems that support range and or hash partitioning. We will ignore statement S21 in query (2) in the following discussion. S21 will be executed once for each result tuple produced. [LIEU92b] discusses the conditions under which the join algorithms can be modified to execute S21 at the join nodes. Otherwise, S21 must be executed centrally by the application program. Leaving out S21 will make it clearer that the algorithms are equally applicable to OODBSs and DBPLs. Analysis of the ....
....each set valued attribute contains 32 oids, a i r is expected to be 1.57 even if no clustering was attempted and references are randomly distributed across the n nodes. 6 When a join is expressed using a nested iterator and the enclosed statement S21 does not fit certain patterns described in [LIEU92b], partitioning is needed. The following is an example of a join expressed with a nested iterator. for (P of CompositePart) for (C of P subparts) suchthat (C cost 100) printf( s s d ,P name,C name,C cost) 25 offered only modest performance improvements, and only in the case ....
[Article contains additional citation context not shown here]
Daniel Lieuwen. Ph.D Thesis. Optimizing and Parallelizing Loops in Object-Oriented Database Programming Languages. University of Wisconsin (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