Hybrid static/dynamic scheduling for already optimized dense matrix factorization

被引:9
|
作者
Donfack, Simplice [1 ]
Grigori, Laura [1 ]
Gropp, William D. [2 ]
Kale, Vivek [2 ]
机构
[1] Univ Paris 11, INRIA Saclay Ile France, Bat 425, F-91405 Orsay, France
[2] Univ Illinois, Dept Comp Sci, Urbana, IL 61801 USA
关键词
dynamic scheduling; communication-avoiding; LU factorization; numerical linear algebra; LOCALITY;
D O I
10.1109/IPDPS.2012.53
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
We present the use of a hybrid static/dynamic scheduling strategy of the task dependency graph for direct methods used in dense numerical linear algebra. This strategy provides a balance of data locality, load balance, and low dequeue overhead. We show that the usage of this scheduling in communication avoiding dense factorization leads to significant performance gains. On a 48 core AMD Opteron NUMA machine, our experiments show that we can achieve up to 64% improvement over a version of CALU that uses fully dynamic scheduling, and up to 30% improvement over the version of CALU that uses fully static scheduling. On a 16-core Intel Xeon machine, our hybrid static/dynamic scheduling approach is up to 8% faster than the version of CALU that uses a fully static scheduling or fully dynamic scheduling. Our algorithm leads to speedups over the corresponding routines for computing LU factorization in well known libraries. On the 48 core AMD NUMA machine, our best implementation is up to 110% faster than MKL, while on the 16 core Intel Xeon machine, it is up to 82% faster than MKL. Our approach also shows significant speedups compared with PLASMA on both of these systems.
引用
收藏
页码:496 / 507
页数:12
相关论文
共 50 条
  • [41] An optimized dynamic resource scheduling strategies for CDMA networks
    Liu, JL
    Lu, XW
    Zhu, JK
    VTC2004-FALL: 2004 IEEE 60TH VEHICULAR TECHNOLOGY CONFERENCE, VOLS 1-7: WIRELESS TECHNOLOGIES FOR GLOBAL SECURITY, 2004, : 3934 - 3937
  • [42] A hybrid recommendation approach using LDA and probabilistic matrix factorization
    Yulin Cao
    Wenli Li
    Dongxia Zheng
    Cluster Computing, 2019, 22 : 8811 - 8821
  • [43] Diversified Recommendation Algorithm for Hybrid Label Based on Matrix Factorization
    Zhang, Shuoshuo
    Xia, Yang
    Li, Xinru
    2018 IEEE 3RD INTERNATIONAL CONFERENCE ON CLOUD COMPUTING AND BIG DATA ANALYSIS (ICCCBDA), 2018, : 39 - 44
  • [44] A hybrid recommendation approach using LDA and probabilistic matrix factorization
    Cao, Yulin
    Li, Wenli
    Zheng, Dongxia
    CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS, 2019, 22 (Suppl 4): : S8811 - S8821
  • [45] An Optimized Framework for Matrix Factorization on the New Sunway Many-core Platform
    Ma, Wenjing
    Liu, Fangfang
    Chen, Daokun
    Lu, Qinglin
    Hu, Yi
    Wang, Hongsen
    Yuan, Xinhui
    ACM TRANSACTIONS ON ARCHITECTURE AND CODE OPTIMIZATION, 2023, 20 (02)
  • [46] STATIC SCHEDULING FOR DYNAMIC DATA-FLOW MACHINES
    BECK, M
    PINGALI, KK
    NICOLAU, A
    JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1990, 10 (04) : 279 - 288
  • [47] Static & Dynamic Appointment Scheduling with Stochastic Gradient Descent
    Cheng, Gary
    Chandrasekher, Kabir
    Walrand, Jean
    2019 AMERICAN CONTROL CONFERENCE (ACC), 2019, : 2092 - 2099
  • [48] A Unified Probabilistic Matrix Factorization Recommendation Fusing Dynamic Tag
    Zheng Dongxia
    Huang Jinghao
    2019 INTERNATIONAL CONFERENCE ON ROBOTS & INTELLIGENT SYSTEM (ICRIS 2019), 2019, : 69 - 72
  • [49] A State-Space Approach to Dynamic Nonnegative Matrix Factorization
    Mohammadiha, Nasser
    Smaragdis, Paris
    Panahandeh, Ghazaleh
    Doclo, Simon
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2015, 63 (04) : 949 - 959
  • [50] Coupling static and dynamic compressions: First measurements in dense hydrogen
    Loubeyre, P
    Celliers, PM
    Hicks, DG
    Henry, E
    Dewaele, A
    Pasley, J
    Eggert, J
    Koenig, M
    Occelli, F
    Lee, KM
    Jeanloz, R
    Neely, D
    Benuzzi-Mounaix, A
    Bradley, D
    Bastea, M
    Moon, S
    Collins, GW
    HIGH PRESSURE RESEARCH, 2004, 24 (01) : 25 - 31