ePubs
The open archive for STFC research publications
Home
About ePubs
Content Policies
News
Help
Privacy/Cookies
Contact ePubs
Full Record Details
DOI
10.5286/dltr.2017001
Persistent URL
http://purl.org/net/epubs/work/33668695
Record Status
Checked
Record Id
33668695
Title
Weak and strong scaling of the CIAN2 code coupling mini-app
Contributors
AD Taylor (STFC Daresbury Lab.)
Abstract
Numerical simulation of scientific systems often involves modelling different spatial and temporal scales and different physical processes to produce a single set of results. In order to achieve this, scientific codes need to be efficiently coupled so that the transfer of data between different codes can be performed effectively on large, multi-node computing clusters. The following investigates the strong and weak scaling of the CIAN2 mini-app on an Intel Xeon IvyBridge cluster over a range of 3-4600 processes. The CIAN2 mini-app is a 'toy' application designed to demonstrate the principles of code coupling. Two 3D meshes are coupled together using MPI communication, with the user specifying the number of MPI processes and the size of the source and target mesh, among other things. Results show that the application demonstrates excellent strong scaling, especially for larger problems. The overhead of communication is less significant for large problems ran across many processes than for small processes ran across many processes. Weak scaling is also good for CIAN2, with results showing that time to solution remains in the same order of magnitude when problem size and MPI process number is increased by the same factor. The above results, while not scientifically significant, show a benchmark standard of parallel performance that other code coupling problems can compare their performance with on this system.
Organisation
STFC
,
HC
Keywords
CCP
,
strong scaling
,
weak scaling
,
Software Outlook
,
Intel Xeon (IvyBridge)
,
code coupling
,
CIAN2
Funding Information
EPSRC
Related Research Object(s):
Licence Information:
Creative Commons Attribution 3.0 Unported (CC BY 3.0)
Language
English (EN)
Type
Details
URI(s)
Local file(s)
Year
Report
DL Technical Reports
DL-TR-2017-001. STFC, 2017.
DL-TR-2017-001.pdf
2017
Showing record 1 of 1
Recent Additions
Browse Organisations
Browse Journals/Series
Login to add & manage publications and access information for OA publishing
Username:
Password:
Useful Links
Chadwick & RAL Libraries
SHERPA FACT
SHERPA RoMEO
SHERPA JULIET
Journal Checker Tool
Google Scholar