ITM Web Conf.
Volume 47, 20222022 2nd International Conference on Computer, Communication, Control, Automation and Robotics (CCCAR2022)
|Number of page(s)||6|
|Section||Algorithm Optimization and Application|
|Published online||23 June 2022|
FPGA-based heterogeneous acceleration study for multidimensional cubing
School of Information Science and Engineering, Yunnan University, Kunming, China
* Corresponding author: email@example.com
Today’s information processing not only faces an explosion of data volume and data dimensions, but also has to meet the growing user requirements for timeliness. The increase in the dimensionality of Kylin brings about the problem of dimensional explosion of multidimensional cubes, which puts great pressure on disk and network transmission. To solve the problem of dimensional explosion when building multidimensional cubes in Kylin, this paper proposes a bottom-up three-layer architecture that links from the hardware compression acceleration kernel layer to the query engine layer through the software driver layer. The software-driven layer is implemented through JNI, dynamic link libraries, and global shared resource pools to link the hardware-accelerated kernel layer to the query engine layer. Finally, comparative experiments on the performance of multidimensional cube building are conducted for heterogeneous clusters and normal clusters. The experimental results show that the hardware- accelerated cluster obtains a 3.7 times speedup ratio in build time and a 2 times speedup ratio in average query time, which can alleviate the IO pressure brought by Kylin during the dimensional explosion.
Key words: Deflate / FPGA / OpenCL / Kylin / OLAP
© The Authors, published by EDP Sciences, 2022
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.