Authors: Yumeng Guo; Wenke Wang; Sikun Li
Addresses: College of Computer, National University of Defense Technology, No. 109 Deya Road, Changsha, 410073, China ' Institute of Ocean Science and Engineering, National University of Defense Technology, No. 109 Deya Road, Changsha, 410073, China ' College of Computer, National University of Defense Technology, No. 109 Deya Road, Changsha, 410073, China
Abstract: As huge amounts of flow data come into being every day, it is challengeable for most flow field visualisation applications on a single PC to handle the large-scale data, because of the memory size restriction. To address the problem, out-of-core strategy with data prefetching is frequently applied to load the data on demand and fill in the speed gap between I/O and computation. In this paper, we focus on improving the efficiency of data-prefetching large-scale streamline visualisation by elevating the hit rate of data block prediction. Our key idea is to extract feature information of the field and then adopt a partitioning strategy that slices important regions into smaller blocks. Experiments show that the major measurement of our partitioning strategy for data prefetching is much better than conventional uniform-partitioned methods, and the total execution time of visualisation system decreases by 10% on average.
Keywords: streamline visualisation; out-of-core technique; data prefetching; block partition.
International Journal of Computational Science and Engineering, 2019 Vol.20 No.2, pp.200 - 208
Received: 06 Apr 2018
Accepted: 21 Oct 2018
Published online: 27 Nov 2019 *