@inproceedings{3c22ff357e0c44939b726f9f2f1939f9,
title = "FuseME: Distributed Matrix Computation Engine based on Cuboid-based Fused Operator and Plan Generation",
abstract = "Operator fusion is essentially and widely used in a large number of matrix computation systems in science and industry. The existing distributed operator fusion methods focus on only either low communication cost with the risk of out of memory or large-scale processing with high communication cost. We propose a distributed elastic fused operator called Cuboid-based Fused Operator (CFO) that achieves both low communication cost and large-scale processing. We also propose a novel fusion plan generator called Cuboid-based Fusion plan Generator (CFG) that finds a fusion plan to fuse more operators including large-scale matrix multiplication. We implement a fast distributed matrix computation engine called FuseME by integrating both CFO and CFG seamlessly. FuseME outperforms the state-of-the-art systems including SystemDS by orders of magnitude.",
keywords = "distributed data-parallel system, matrix operators, operator fusion",
author = "Donghyoung Han and Jongwuk Lee and Kim, \{Min Soo\}",
note = "Publisher Copyright: {\textcopyright} 2022 ACM.; 2022 ACM SIGMOD International Conference on the Management of Data, SIGMOD 2022 ; Conference date: 12-06-2022 Through 17-06-2022",
year = "2022",
month = jun,
doi = "10.1145/3514221.3517895",
language = "English",
series = "Proceedings of the ACM SIGMOD International Conference on Management of Data",
publisher = "Association for Computing Machinery",
pages = "1891--1904",
booktitle = "SIGMOD 2022 - Proceedings of the 2022 International Conference on Management of Data",
}