Efficient Breadth First Search on Multi-GPU Systems Using GPU-Centric OpenSHMEM

被引：2

作者：

Potluri, Sreeram ^{[1
]}

Goswami, Anshuman ^{[1
]}

Venkata, Manjunath Gorentla ^{[2
]}

Imam, Neena ^{[2
]}

机构：

[1] NVIDIA Corp, Santa Clara, CA 95051 USA

[2] Oak Ridge Natl Lab, Comp Sci & Math Div, Oak Ridge, TN USA

来源：

OPENSHMEM AND RELATED TECHNOLOGIES: BIG COMPUTE AND BIG DATA CONVERGENCE, OPENSHMEM 2017 | 2018年 / 10679卷

关键词：

D O I：

10.1007/978-3-319-73814-7_6

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

NVSHMEM is an implementation of OpenSHMEM for NVIDIA GPUs which allows communication to be issued from inside CUDA kernels. In this work, we present an implementation of Breadth First Search for multi-GPU systems using NVSHMEM. We analyze the benefits and bottlenecks of moving fine-grained communication into CUDA kernels. Using our implementation of BFS, we achieve up to 75% improvement in performance compared to a CUDA-aware MPI-based implementation, in the best case.

引用

页码：82 / 96

页数：15

共 50 条

[1] Efficient breadth first search on multi-GPU systems
Mastrostefano, Enrico
Bernaschi, Massimo
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 2013, 73 (09) : 1292 - 1305
[2] GPU-centric Communication on NVIDIA GPU Clusters with InfiniBand: A Case Study with OpenSHMEM
Potluri, S.
Goswami, A.
Rossetti, D.
Newburn, C. J.
Venkata, M. Gorentla
Imam, N.
2017 IEEE 24TH INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE COMPUTING (HIPC), 2017, : 253 - 262
[3] Efficient parallel A* search on multi-GPU system
He, Xin
Yao, Yapeng
Chen, Zhiwen
Sun, Jianhua
Chen, Hao
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE, 2021, 123 : 35 - 47
[4] Benchmarking multi-GPU applications on modern multi-GPU integrated systems
Bernaschi, Massimo
Agostini, Elena
Rossetti, Davide
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2021, 33 (14):
[5] Modelling Multi-GPU Systems
Spampinato, Daniele G.
Elster, Anne C.
Natvig, Thorvald
PARALLEL COMPUTING: FROM MULTICORES AND GPU'S TO PETASCALE, 2010, 19 : 562 - 569
[6] Efficient Solving of Scan Primitive on Multi-GPU Systems
Dieguez, Adrian P.
Amor, Margarita
Doallo, Ramon
Nukada, Akira
Matsuoka, Satoshi
2018 32ND IEEE INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM (IPDPS), 2018, : 794 - 803
[7] Gossip: Efficient Communication Primitives for Multi-GPU Systems
Kobus, Robin
Juenger, Daniel
Hundt, Christian
Schmidt, Bertil
PROCEEDINGS OF THE 48TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING (ICPP 2019), 2019,
[8] Storage access optimization for efficient GPU-centric information retrieval
Shrestha, Susav
Gautam, Aayush
Reddy, Narasimha
JOURNAL OF SUPERCOMPUTING, 2025, 81 (04):
[9] Efficient Implementation of MrBayes on Multi-GPU
Bao, Jie
Xia, Hongju
Zhou, Jianfu
Liu, Xiaoguang
Wang, Gang
MOLECULAR BIOLOGY AND EVOLUTION, 2013, 30 (06) : 1471 - 1479
[10] Parallel Breadth First Search on GPU Clusters
Fu, Zhisong
Dasari, Harish Kumar
Bebee, Bradley
Berzins, Martin
Thompson, Bryan
2014 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA), 2014, : 110 - 118

← 1 2 3 4 5 →