CSS-Tree Implementation and Optimization on Graphics Processing Units

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


FYT Presentation & Demonstration

Title: "CSS-Tree Implementation and Optimization on Graphics Processing Units"

By

Mr. FANG Rui

Abstract

GPUs are commodity processors traditionally designed for
graphics applications. Re-cent research has shown that they can accelerate
simple database operations such as selections and projections orders of
magnitude over CPUs with its massively parallel computing power, yet
little has been done so far on how GPUs can be programmed for complicated
database constructs and operations, such as tree indexes and joins, and
how well a full-fledged GPU query co-processor performs in comparison with
their CPU counterparts. In this thesis, we explore the ar-chitectural
features of the GPUs for query processing and present our GPU
implementation of an efficient indexing structure - the CSS-tree. Our
performance results have shown improvement by more than an order of
magnitude over the optimized CPU implementation.

Keywords
Graphics Processing Units, Parallel Query Processing, Relational Databases

Note: This work is part of the on-going research carried out by the GPUQP
(Query Processing on Graphics Processors) project group at HKUST.

Relevant information can be found in the FYT report available online at
http://ihome.ust.hk/~eg_frx/FYT%20Report.doc




Date:			Wednesday, 9 May 2007

Time:			3:00p.m - 4:00p.m.

Venue:			Room 4480
			Lifts 25-26

Advisor:		Dr. Luo Qiong

2nd Reader:		Prof. Fred Lochovsky


**** ALL are Welcome ****