Massively Parallel Relational Database Processing on the Connection Machine CM-2
نویسندگان
چکیده
Several research effort has been devoted to increase the performance of the relational database systems. Various types of database machines were so far proposed. Among them, shared nothing parallel database system such as Teradata, GAMMA, TANDEM are actually implemented and detail performance evaluation is undergoing. On the other hand, recently massively parallel processor like the Connection Machine, attracts strong attention for the future super parallel system. Such parallel system is called “Data Parallel Computer” exploiting not procedure level parallelism but data level parallelism. One of the most promising application for such Data Parallel Machine be a relational database processing. However so far there has been no research to examine the effectiveness of massively parallel machine for RDB processing. In this paper we propose the two relation join algorithms for data parallel machine, that is, Data Parallel Sort Merge Join and Data Parallel Hash Join. These two algorithms are implemented on the Connection Machine CM-Z. Its performance is reported.
منابع مشابه
An Optimization Technique of Discrimination Networks in Active Database Systems for Massively Parallel Processing
Because rule checking in active database systems is slow, its speed-up has been required. In this paper, we propose a dynamic query optimization technique for TREAT-like discrimination networks under a massively parallel processing environment. Although our technique requires extra-processors to optimize, it can derive an optimal query tree for the TREAT as the state of databases changes. We al...
متن کاملMethods of Large Grammar Representation in Massively Parallel Parsing Systems
This paper describes techniques for massively parallel parsing where sequences of lexical categories are assigned to single processors and compared in parallel to a given input string. Because ven small grammars result in full expansions that are much larger than the largest existing massively parallel computers, we need to develop techniques for "doubling up" sequences on processors so that th...
متن کاملComparison of Direct Particle Simulation on the MasPar MP - 2 and the Connection Machine CM - 2
Particle simulation as applied in the direct simulation Monte Carlo (DSMC) method is a technique for analyzing low density flows and is used extensively for engineering analysis of aerospace vehicles. This work compares two implementations of this method on outwardly similar massively parallel architectures. The MasPar MP-2 and the Connection Machine CM-2 are both massively parallel SIMD archit...
متن کاملComputing Data Cubes Using Massively Parallel Processors
To better support decision making, it was proposed to extend SQL to include data cube operations. Computation of data cube requires computing a number of interrelated group-bys, which is rather expensive operation when databases are large. In this paper, we propose to couple a relational database management system with massively parallel processors (MPP) to facilitate on-line analytic processin...
متن کاملThree-Dimensional Monte Carlo Device Simulation for Massively Parallel Architectures
The applicability of a massively parallel processing (MPP) paradigm to the simulation of charge transport in semiconductor devices through the Monte Carlo method is investigated. A unique mapping of Monte Carlo simulation to a data-parallel software model has been developed in which the problem is decoupled into multiple computational domains, thereby increasing the locality of computation. The...
متن کامل