Task scheduling using a block dependency DAG for block-oriented sparse Cholesky factorization

Heejo Lee, Jong Kim, Sung Je Hong, Sunggu Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Block-oriented sparse Cholesky factorization decomposes a sparse matrix into rectangular sub-blocks; each block can then be handled as a computational unit in order to increase data reuse in a hierarchical memory system. Also, the factorization method increases the degree of concurrency with the reduction of communication volumes so that it performs more efficiently on a distributed-memory multiprocessor system than the customary column-oriented factorization method. But until now, mapping of blocks to processors has been designed for load balance with restricted communication patterns. In this paper, we represent tasks using a block dependency DAG that shows the execution behavior of block sparse Cholesky factorization in a distributed-memory system. Since the characteristics of tasks for the block Cholesky factorization are different from those of the conventional parallel task model, we propose a new task scheduling algorithm using a block dependency DAG. The proposed algorithm consists of two stages: early-start clustering, and affined cluster mapping. The early-start clustering stage is used to cluster tasks with preserving the earliest start time of a task without limiting parallelism. After task clustering, the affined cluster mapping stage allocates clusters to processors considering both communication cost and load balance. Experimental results on the Fujitsu parallel system show that the proposed task scheduling approach outperforms other processor mapping methods.

Original languageEnglish
Title of host publicationProceedings of the 2000 ACM Symposium on Applied Computing, SAC 2000
Pages641-648
Number of pages8
DOIs
Publication statusPublished - 2000
Externally publishedYes
Event2000 ACM Symposium on Applied Computing, SAC 2000 - Como, Italy
Duration: 2000 Mar 192000 Mar 21

Publication series

NameProceedings of the ACM Symposium on Applied Computing
Volume2

Other

Other2000 ACM Symposium on Applied Computing, SAC 2000
Country/TerritoryItaly
CityComo
Period00/3/1900/3/21

Keywords

  • Block-oriented Cholesky factorization
  • Directed acyclic graph
  • Parallel sparse matrix factorization
  • Task scheduling

ASJC Scopus subject areas

  • Software

Fingerprint

Dive into the research topics of 'Task scheduling using a block dependency DAG for block-oriented sparse Cholesky factorization'. Together they form a unique fingerprint.

Cite this