Optimization for Multi-Join Queries on the GPU

doi:10.1109/ACCESS.2020.3002610

Return

Optimization for Multi-Join Queries on the GPU

2020-01-01

DOI：10.1109/ACCESS.2020.3002610

Original

Save

View PDF

Abstract

En 中文

Multi-join queries are important operations in data management systems and data integration systems, and their efficiency has attracted the attention of researchers. In recent years, graphics processing units (GPUs) have developed rapidly and become a powerful tool for parallel computing, providing a new idea for multi-join query optimization. This paper studies the use of GPU technology to optimize multi-join queries and focuses on two points: 1) a multi-phase optimization strategy and 2) optimization methods of each stage. For the first point, we discuss a two-phase optimization strategy on the GPU and prove the effectiveness of this strategy. For the second point, we provide an establishment method of a minimum cost join tree on the GPU, the parallel execution methods of intra-join and inter-join on the GPU, and a strategy of scheduling multiple joins to execute in parallel on the GPU. Experimental results show that the multi-join query optimization proposed in this paper improves the efficiency of multi-join queries, especially in the case of high load and complex join queries, achieving higher throughput than that of previous optimization algorithms.

Keywords:

Graphics processing units

Heuristic algorithms

Parallel algorithms

Optimization methods

Estimation

GPU

multi-join query

parallel optimization

two-phase optimization strategy

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.