Join is one of the most important operations in data analytics systems. Prior works focus mainly on join optimization using GPUs, but little is known about performance impact on the MICs. In order to investigate potential benefits of the use of MIC accelerators in improving performance of join operation, in this paper we design a join scheme with a CPU and MICs working collaboratively. This scheme includes task partitioning, a data transfer mode, join algorithm design. Experimental results show that our collective join scheme is effective for a heterogeneous platform with two Xeon Phi cards, and can improve performance by up to 30 % over the CPU-only platform.
CITATION STYLE
Zhou, K., Sun, H., Chen, H., Wu, T., & Li, C. (2016). A collaborative join scheme on a MIC-based heterogeneous platform. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9932 LNCS, pp. 454–458). Springer Verlag. https://doi.org/10.1007/978-3-319-45817-5_45
Mendeley helps you to discover research relevant for your work.