RT Conference Proceedings
T1 CuDB : a Relational Database Engine Boosted by Graphics Processing Units
A1 Cremer, Samuel
A1 Bagein, Michel
A1 Mahmoudi, Saïd
A1 Manneback, Pierre
A2 Carretero Pérez, Jesús
A2 García Blas, Javier
A2 Petcu, Dana
AB GPUs benefit from much more computation power with the same order of energy consumption than CPUs. Thanksto their massive data parallel architecture, GPUs can outperform CPUs, especially on Single Program MultipleData (SPMD) programming paradigm on a large amount of data. Database engines are now everywhere, fromdifferent sizes and complexities, for multiple usages, embedded or distributed; in 2012, 500 million of SQLiteactive instances were estimated over the world. Our goal is to exploit the computation power of GPUs to improveperformance of SQLite, which is a key software component of many applications and systems. In this paper, weintroduce CuDB, a GPU-boosted in-memory database engine (IMDB) based on SQLite. The SQLite API remainsunchanged, allowing developers to easily upgrade database engine from SQlite to CuDB even on already existingapplications. Preliminary results show significant speedups of 70x with join queries on datasets of 1 million records.We also demonstrate the "memory bounded" character of GPU-databases and show the energy efficiency of ourapproach.
SN 978-84-608-6309-0
YR 2016
FD 2016-02
LK https://hdl.handle.net/10016/22891
UL https://hdl.handle.net/10016/22891
LA eng
NO Proceedings of the First PhD Symposium on Sustainable UltrascaleComputing Systems (NESUS PhD 2016) Timisoara, Romania. February 8-11, 2016.
NO European Cooperation in Science and Technology. COST
DS e-Archivo
RD 17 jul. 2024