Bridging the Gap between Deep Learning and Sparse Matrix Format Selection

被引：75

作者：

Zhao, Yue ^{[1
]}

Li, Jiajia ^{[2
]}

Liao, Chunhua ^{[3
]}

Shen, Xipeng ^{[1
]}

机构：

[1] North Carolina State Univ, Comp Sci, Raleigh, NC 27695 USA

[2] Georgia Inst Technol, CSE, Atlanta, GA 30332 USA

[3] Lawrence Livermore Natl Lab, CASC, Livermore, CA USA

来源：

ACM SIGPLAN NOTICES | 2018年 / 53卷 / 01期

基金：

美国国家科学基金会;

关键词：

SpMV; Sparse matrix; Format selection; Convolutional neural network; Deep learning;

D O I：

10.1145/3200691.3178495

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

This work presents a systematic exploration on the promise and special challenges of deep learning for sparse matrix format selection-a problem of determining the best storage format for a matrix to maximize the performance of Sparse Matrix Vector Multiplication (SpMV). It describes how to effectively bridge the gap between deep learning and the special needs of the pillar HPC problem through a set of techniques on matrix representations, deep learning structure, and cross-architecture model migrations. The new solution cuts format selection errors by two thirds, and improves SpMV performance by 1.73x on average over the state of the art.

引用

页码：94 / 108

页数：15

共 50 条

[1] POSTER: Bridging the Gap between Deep Learning and Sparse Matrix Format Selection
Zhao, Yue
Li, Jiajia
Liao, Chunhua
Shen, Xipeng
[J]. 2017 26TH INTERNATIONAL CONFERENCE ON PARALLEL ARCHITECTURES AND COMPILATION TECHNIQUES (PACT), 2017, : 152 - 153
[2] Runtime sparse matrix format selection
Armstrong, Warren
Rendell, Alistair P.
[J]. ICCS 2010 - INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE, PROCEEDINGS, 2010, 1 (01): : 135 - 144
[3] SpTFS: Sparse Tensor Format Selection for MTTKRP via Deep Learning
Sun, Qingxiao
Liu, Yi
Dun, Ming
Yang, Hailong
Luan, Zhongzhi
Gan, Lin
Yang, Guangwen
Qian, Depei
[J]. PROCEEDINGS OF SC20: THE INTERNATIONAL CONFERENCE FOR HIGH PERFORMANCE COMPUTING, NETWORKING, STORAGE AND ANALYSIS (SC20), 2020,
[4] Bridging the gap between debiasing and privacy for deep learning
Barbano, Carlo Alberto
Tartaglione, Enzo
Grangetto, Marco
[J]. 2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW 2021), 2021, : 3799 - 3808
[5] “Transfer Learning” for Bridging the Gap Between Data Sciences and the Deep Learning
Sohail A.
[J]. Annals of Data Science, 2024, 11 (01) : 337 - 345
[6] Reinforcement Learning for Automated Performance Tuning: Initial Evaluation for Sparse Matrix Format Selection
Armstrong, Warren
Rendell, Alistair P.
[J]. 2008 IEEE INTERNATIONAL CONFERENCE ON CLUSTER COMPUTING, 2008, : 411 - 420
[7] MANet: An Architecture Adaptive Method for Sparse Matrix Format Selection
Sun, Zhenglun
Qiao, Peng
Dou, Yong
[J]. ALGORITHMS AND ARCHITECTURES FOR PARALLEL PROCESSING, ICA3PP 2023, PT II, 2024, 14488 : 311 - 326
[8] Sparse Matrix Format Selection with Multiclass SVM for SpMV on GPU
Benatia, Akrem
Ji, Weixing
Wang, Yizhuo
Shi, Feng
[J]. PROCEEDINGS 45TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING - ICPP 2016, 2016, : 496 - 505
[9] Bridging the Gap Between Memory and Communication Efficiency on Distributed Deep Learning Systems
Zhao, Shaofeng
Liu, Bo
Wang, Fang
Feng, Dan
[J]. IEEE ACCESS, 2021, 9 : 57075 - 57088
[10] CrabNet for Explainable Deep Learning in Materials Science: Bridging the Gap Between Academia and Industry
Anthony Yu-Tung Wang
Mahamad Salah Mahmoud
Mathias Czasny
Aleksander Gurlo
[J]. Integrating Materials and Manufacturing Innovation, 2022, 11 : 41 - 56

← 1 2 3 4 5 →