Latest revision as of 20:18, 28 January 2021

Abstract

Heterogeneous architectures that combine multi-core CPUs with many-core GPGPUs have the potential to improve the performance of data-intensive stream processing applications. Yet, a stream processing engine must execute streaming SQL queries with sufficient data-parallelism to fully utilise the available heterogeneous processors, and decide how to use each processor in the most effective way. Addressing these challenges, we demonstrate S aber , a hybrid high-performance relational stream processing engine for CPUs and GPGPUs. S aber executes window-based streaming SQL queries in a data-parallel fashion and employs an adaptive scheduling strategy to balance the load on the different types of processors. To hide data movement costs, S aber pipelines the transfer of stream data between CPU and GPGPU memory. In this paper, we review the design principles of S aber in terms of its hybrid stream processing model and its architecture for query execution. We also present a web front-end that monitors processing throughput.

Original document

The different versions of the original document can be found in:

• http://hdl.handle.net/10044/1/41256

• https://spiral.imperial.ac.uk/bitstream/10044/1/41256/2/saber-demo-debs16.pdf,

https://dblp.uni-trier.de/db/conf/debs/debs2016.html#KoliousisWFWCP16,

https://dl.acm.org/citation.cfm?doid=2933267.2933291,

https://spiral.imperial.ac.uk/handle/10044/1/41256,

https://academic.microsoft.com/#/detail/2425435860

• http://dl.acm.org/ft_gateway.cfm?id=2933291&ftid=1742584&dwn=1,

http://dx.doi.org/10.1145/2933267.2933291 under the license http://www.acm.org/publications/policies/copyright_policy#Background