Harald Kosch. Exploiting serialized bushy trees for parallel relational query optimization. PhD thesis, Ecole Normale Superieure de Lyon, Lyon, France, June 1997. 34 Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
The Use of Randomized Search Strategies for Complex Parallel.. - Kosch Self-citation (Kosch) (Correct)
....chosen in a way that on each processor, no deteriorisation can occur for normal workload estimates. For computing the work done on non preemptable resources, like the memory, and for determining their allocation, special features must be introduced. A regulation factor is presented (see in detail [45, 46]) which guarantees a fair utilization of the memory. The higher the value of this factor, the more memory is granted to the first join, which consequently will leave less memory to the other joins running in parallel to it. This factor is also set a priori in dependency of the initial memory ....
Harald Kosch. Exploiting serialized bushy trees for parallel relational query optimization. PhD thesis, Ecole Normale Superieure de Lyon, Lyon, France, June 1997. 34
Intégration D'heuristiques D'ordonnancement Dans.. - Brunie, Kosch Self-citation (Kosch) (Correct)
....ont ete concu afin de penaliser par des surco uts l utilisation sur intensive des ressources partagees. Cinqmodules dynamiquescompletent cette architecture logicielle : le gestionnaire de transformations implemente une strategie de recherche aleatoire (differentes strat egies sont decrites dans [Kos97] Le role de ce module consiste a choisir une transformation optimisante a appliquer a l arbre de requete courant (ce qui implique, entre autres, de definir le degre de parallelisme inter operations) Pour ce faire, ce module communique avec le module d allocation de ressources. Celui ci a ....
Harald Kosch. Exploiting serialized bushy trees for parallel relational query optimization. PhD thesis, Ecole Normale Superieure de Lyon, Lyon, France, June 1997.
Optimizing Complex Decision Support Queries for Parallel.. - Brunie, Kosch (1997) Self-citation (Kosch) (Correct)
....3 Experimental results In this section we report an overview of the experiments we have done with ModParOpt on a decision support database schema with 100 relations and 432 queries. First, the parameter of the randomized search algorithms were tuned. We refer the reader for a deep analyze to [14]. Then we studied the quality of the execution scenario found by the different search algorithms and examined the optimization effort compared to the estimated processing time. 3.1 Experiment profile Queries were tested in conjunction with different types of relation cardinality, tuple size and ....
....chosen clearly lesser inter operation parallelism than for all the other cases. As for memory contentions, the trend to lesser intra operation parallelism, was not so obvious. For 64 processors the optimizer decided for more inter operation parallelism in comparison to 32 or 8 processors (see [14] for a detailed analysis) The average local minima cost was more than 10 times lower than the average randomly sample cost. Moreover the cost of the best local minima was more than 30 above the optimal cost found, stating that it is hard to generate a good scenario. Consequently we continued the ....
Harald Kosch. Exploiting serialized bushy trees for parallel relational query optimization. PhD thesis, Ecole Normale Superieure de Lyon, Lyon, France, June 1997.
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