Full Text Views
46
Probabilistic frequent itemset mining, which discovers frequent itemsets from uncertain data, has attracted much attention due to inherent uncertainty in the real world. Many algorithms have been proposed to tackle this problem, but their performance is not satisfactory because handling uncertainty incurs high processing cost. To accelerate such computation, we utilize GPUs (Graphics Processing Units). Our previous work accelerated an existing algorithm with a single GPU. In this paper, we extend the work to employ multiple GPUs. Proposed methods minimize the amount of data that need to be communicated among GPUs, and achieve load balancing as well. Based on the methods, we also present algorithms on a GPU cluster. Experiments show that the single-node methods realize near-linear speedups, and the methods on a GPU cluster of eight nodes achieve up to a 7.1 times speedup.
Yusuke KOZAWA
University of Tsukuba
Toshiyuki AMAGASA
Information and Systems
Hiroyuki KITAGAWA
Information and Systems
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yusuke KOZAWA, Toshiyuki AMAGASA, Hiroyuki KITAGAWA, "Probabilistic Frequent Itemset Mining on a GPU Cluster" in IEICE TRANSACTIONS on Information,
vol. E97-D, no. 4, pp. 779-789, April 2014, doi: 10.1587/transinf.E97.D.779.
Abstract: Probabilistic frequent itemset mining, which discovers frequent itemsets from uncertain data, has attracted much attention due to inherent uncertainty in the real world. Many algorithms have been proposed to tackle this problem, but their performance is not satisfactory because handling uncertainty incurs high processing cost. To accelerate such computation, we utilize GPUs (Graphics Processing Units). Our previous work accelerated an existing algorithm with a single GPU. In this paper, we extend the work to employ multiple GPUs. Proposed methods minimize the amount of data that need to be communicated among GPUs, and achieve load balancing as well. Based on the methods, we also present algorithms on a GPU cluster. Experiments show that the single-node methods realize near-linear speedups, and the methods on a GPU cluster of eight nodes achieve up to a 7.1 times speedup.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E97.D.779/_p
Copy
@ARTICLE{e97-d_4_779,
author={Yusuke KOZAWA, Toshiyuki AMAGASA, Hiroyuki KITAGAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Probabilistic Frequent Itemset Mining on a GPU Cluster},
year={2014},
volume={E97-D},
number={4},
pages={779-789},
abstract={Probabilistic frequent itemset mining, which discovers frequent itemsets from uncertain data, has attracted much attention due to inherent uncertainty in the real world. Many algorithms have been proposed to tackle this problem, but their performance is not satisfactory because handling uncertainty incurs high processing cost. To accelerate such computation, we utilize GPUs (Graphics Processing Units). Our previous work accelerated an existing algorithm with a single GPU. In this paper, we extend the work to employ multiple GPUs. Proposed methods minimize the amount of data that need to be communicated among GPUs, and achieve load balancing as well. Based on the methods, we also present algorithms on a GPU cluster. Experiments show that the single-node methods realize near-linear speedups, and the methods on a GPU cluster of eight nodes achieve up to a 7.1 times speedup.},
keywords={},
doi={10.1587/transinf.E97.D.779},
ISSN={1745-1361},
month={April},}
Copy
TY - JOUR
TI - Probabilistic Frequent Itemset Mining on a GPU Cluster
T2 - IEICE TRANSACTIONS on Information
SP - 779
EP - 789
AU - Yusuke KOZAWA
AU - Toshiyuki AMAGASA
AU - Hiroyuki KITAGAWA
PY - 2014
DO - 10.1587/transinf.E97.D.779
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E97-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2014
AB - Probabilistic frequent itemset mining, which discovers frequent itemsets from uncertain data, has attracted much attention due to inherent uncertainty in the real world. Many algorithms have been proposed to tackle this problem, but their performance is not satisfactory because handling uncertainty incurs high processing cost. To accelerate such computation, we utilize GPUs (Graphics Processing Units). Our previous work accelerated an existing algorithm with a single GPU. In this paper, we extend the work to employ multiple GPUs. Proposed methods minimize the amount of data that need to be communicated among GPUs, and achieve load balancing as well. Based on the methods, we also present algorithms on a GPU cluster. Experiments show that the single-node methods realize near-linear speedups, and the methods on a GPU cluster of eight nodes achieve up to a 7.1 times speedup.
ER -