Advanced

High Performance I/O

Jackson, Adrian; Reid, Fiona; Hein, Joachim LU ; Soba, Alejandro and Saez, Xavier (2011) 19th International Euromicro Conference on Parallel, Distributed and Network-Based Processing (PDP 2011) In Parallel, Distributed and Network-Based Processing (PDP), 2011 19th Euromicro International Conference on p.349-356
Abstract
Parallelisation, serial optimisation, compiler tuning, and many more techniques are used to optimise and improve the performance scaling of parallel programs. One area which is frequently not optimised is file I/O. This is because it is often not considered to be key to the performance of a program and also because it is traditionally difficult to optimise and very machine specific. However, in the current era of Peta- and Exascale computing it is no longer possible to ignore I/O performance as it can significantly limit the scaling of many codes when executing on very large numbers of processors or cores. Furthermore, as producing data is the main purpose of most simulation codes any work that can be undertaken to provide improved... (More)
Parallelisation, serial optimisation, compiler tuning, and many more techniques are used to optimise and improve the performance scaling of parallel programs. One area which is frequently not optimised is file I/O. This is because it is often not considered to be key to the performance of a program and also because it is traditionally difficult to optimise and very machine specific. However, in the current era of Peta- and Exascale computing it is no longer possible to ignore I/O performance as it can significantly limit the scaling of many codes when executing on very large numbers of processors or cores. Furthermore, as producing data is the main purpose of most simulation codes any work that can be undertaken to provide improved performance of I/0 can be applicable to a very large range of simulation codes, and provide them with improved functionality (i.e. the ability to produce more data).This paper describes some of the issues surrounding I/O, the technology that is commonly deployed to provide I/O on HPC machines and the software libraries available to programmers to undertake I/O. The performance of all these aspects of I/O on a range of HPC systems were investigated by the authors and a represented in this paper to motivate the discussions in the paper. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
optimisation, input-output programs, Peta computing, high performance I/O, serial optimisation, compiler tuning, parallel programs, performance scaling, Exascale computing, File systems, Hardware, Program processors, Libraries, Optimization, Computational modeling, Servers, NetCDF, I/O, HPC, parallelisation, Lustre, GPFS, MPI-I/O, HDF5
in
Parallel, Distributed and Network-Based Processing (PDP), 2011 19th Euromicro International Conference on
editor
Cotronis, Yiannis; Danelutto, Marco and Papadopoulos, George Angelos
pages
8 pages
publisher
IEEE--Institute of Electrical and Electronics Engineers Inc.
conference name
19th International Euromicro Conference on Parallel, Distributed and Network-Based Processing (PDP 2011)
external identifiers
  • Scopus:79954985940
ISBN
9781424496822 (print)
DOI
10.1109/PDP.2011.16
language
English
LU publication?
yes
id
a7d4f03f-3e07-4aa3-8d49-e684ac73bec0 (old id 3809837)
alternative location
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5739034&tag=1
date added to LUP
2013-10-18 20:36:44
date last changed
2017-01-01 08:07:46
@inproceedings{a7d4f03f-3e07-4aa3-8d49-e684ac73bec0,
  abstract     = {Parallelisation, serial optimisation, compiler tuning, and many more techniques are used to optimise and improve the performance scaling of parallel programs. One area which is frequently not optimised is file I/O. This is because it is often not considered to be key to the performance of a program and also because it is traditionally difficult to optimise and very machine specific. However, in the current era of Peta- and Exascale computing it is no longer possible to ignore I/O performance as it can significantly limit the scaling of many codes when executing on very large numbers of processors or cores. Furthermore, as producing data is the main purpose of most simulation codes any work that can be undertaken to provide improved performance of I/0 can be applicable to a very large range of simulation codes, and provide them with improved functionality (i.e. the ability to produce more data).This paper describes some of the issues surrounding I/O, the technology that is commonly deployed to provide I/O on HPC machines and the software libraries available to programmers to undertake I/O. The performance of all these aspects of I/O on a range of HPC systems were investigated by the authors and a represented in this paper to motivate the discussions in the paper.},
  author       = {Jackson, Adrian and Reid, Fiona and Hein, Joachim and Soba, Alejandro and Saez, Xavier},
  booktitle    = {Parallel, Distributed and Network-Based Processing (PDP), 2011 19th Euromicro International Conference on},
  editor       = {Cotronis, Yiannis and Danelutto, Marco and Papadopoulos, George Angelos},
  isbn         = {9781424496822 (print)},
  keyword      = {optimisation,input-output programs,Peta computing,high performance I/O,serial optimisation,compiler tuning,parallel programs,performance scaling,Exascale computing,File systems,Hardware,Program processors,Libraries,Optimization,Computational modeling,Servers,NetCDF,I/O,HPC,parallelisation,Lustre,GPFS,MPI-I/O,HDF5},
  language     = {eng},
  pages        = {349--356},
  publisher    = {IEEE--Institute of Electrical and Electronics Engineers Inc.},
  title        = {High Performance I/O},
  url          = {http://dx.doi.org/10.1109/PDP.2011.16},
  year         = {2011},
}