An Adaptive, Scalable, and Portable Technique for Speeding Up MPI-Based Applications

被引：0

作者：

Filgueira, Rosa ^{[1
]}

Atkinson, Malcolm ^{[1
]}

Nunez, Alberto ^{[2
]}

Fernandez, Javier ^{[3
]}

机构：

[1] Univ Edinburgh, Sch Informat, Edinburgh EH8 9AB, Midlothian, Scotland

[2] Univ Complutense Madrid, Dept Sistemas Informat Computac, Madrid 28040, Spain

[3] Univ Carlos III Madrid, Dept Arquitectura Comp, Leganes 3028911, Spain

来源：

EURO-PAR 2012 PARALLEL PROCESSING | 2012年 / 7484卷

关键词：

MPI Library; Parallel techniques; High-Performance Computing; Compression algorithms; Adaptive systems; Portable optimizations; FLOATING-POINT DATA; COMPRESSION; PERFORMANCE; IMPLEMENTATION;

D O I：

暂无

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

This paper presents a portable optimization for MPI communications, called PRAcTICaL-MPI (Portable Adaptive Compression Library- MPI). PRAcTICaL-MPI reduces the data volume exchanged among processes by using lossless compression and offers two main advantages. Firstly, it is independent of the MPI implementation and the application used. Secondly, it allows for turning the compression on and off and selecting the most appropriate compression algorithm at runtime, depending on the characteristics of each message and on network performance. We have validated PRAcTICaL-MPI in different MPI implementations and HPC clusters. The evaluation shows that compressing MPI messages with the best algorithm and only when it is worthwhile, we obtain a great reduction in the overall execution time for many of the scenarios considered.

引用

页码：729 / 740

页数：12

共 50 条

[1] Adaptive-Compi: Enhancing Mpi-Based Applications' Performance and Scalability by using Adaptive Compression
Filgueira, Rosa
Singh, David E.
Carretero, Jesus
Calderon, Alejandro
Garcia, Felix
[J]. INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS, 2011, 25 (01): : 93 - 114
[2] On Network Locality in MPI-Based HPC Applications
Zahn, Felix
Froening, Holger
[J]. PROCEEDINGS OF THE 49TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING, ICPP 2020, 2020,
[3] ScalScheduling: A Scalable Scheduling Architecture for MPI-based Interactive Analysis Programs
Yin, Jiangling
Foran, Andrew
Zhang, Xuhong
Wang, Jun
[J]. 2014 23RD INTERNATIONAL CONFERENCE ON COMPUTER COMMUNICATION AND NETWORKS (ICCCN), 2014,
[4] Runtime Support for Irregular Computation in MPI-Based Applications
Zhao, Xin
Balaji, Pavan
Gropp, William
[J]. 2015 15TH IEEE/ACM INTERNATIONAL SYMPOSIUM ON CLUSTER, CLOUD AND GRID COMPUTING, 2015, : 701 - 704
[5] Experimental and Theoretical Speedup Prediction of MPI-Based Applications
Elnashar, Alaa
Aljahdali, Sultan
[J]. COMPUTER SCIENCE AND INFORMATION SYSTEMS, 2013, 10 (03) : 1247 - 1267
[6] Performance Modeling of MPI-based Applications on Cloud Multicore Servers
Saad, Abdallah
El-Mandy, Ahmed
El-Shishiny, Hisham
[J]. PROCEEDINGS OF THE RAPID SIMULATION AND PERFORMANCE EVALUATION: METHODS AND TOOLS (RAPIDO '19) / HIPEAC'19 CONFERENCE, 2019,
[7] Web Services for MPI-based Parallel Applications on a Rocks Cluster
Sajjipanon, Pitch
Ngamsuriyaroj, Sudsanguan
[J]. 2008 IEEE ASIA-PACIFIC SERVICES COMPUTING CONFERENCE, VOLS 1-3, PROCEEDINGS, 2008, : 265 - 270
[8] Exploit Approximation to Support Fault Resiliency in MPI-based Applications
Rocco, Roberto
Palermo, Gianluca
[J]. 2023 53RD ANNUAL IEEE/IFIP INTERNATIONAL CONFERENCE ON DEPENDABLE SYSTEMS AND NETWORKS WORKSHOPS, DSN-W, 2023, : 187 - 190
[9] Scaling Betweenness Approximation to Billions of Edges by MPI-based Adaptive Sampling
van der Grinten, Alexander
Meyerhenke, Henning
[J]. 2020 IEEE 34TH INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM IPDPS 2020, 2020, : 527 - 535
[10] Efficient MPI-based Communication for GPU-Accelerated Dask Applications
Shafi, Aamir
Hashmi, Jahanzeb Maqbool
Subramoni, Hari
Panda, Dhabaleswar K.
[J]. 21ST IEEE/ACM INTERNATIONAL SYMPOSIUM ON CLUSTER, CLOUD AND INTERNET COMPUTING (CCGRID 2021), 2021, : 277 - 286

← 1 2 3 4 5 →