Home arrow Blog arrow Default Category arrow OpenFOAM 1.7 motorBike - Run Times

OpenFOAM 1.7 motorBike - Run Times PDF Print E-mail


The motorBike case is part of the tutorials that come with OpenFOAM 1.6 and above. The version of the case used for these run times is from the OpenFOAM 1.6.x git version. These timing results are to be compared with the previous timings made here.

The set-up:

  • The tests that were performed were a single run with each available version of OpenFOAM, all executed in computers with 4GiB of RAM and have Intel Core2Duo E8400 @3.00GHz.

  • The Linux versions are still the same versions used in the previous timings, namely OpenFOAM 1.6 used are the ones available here (32bit) and here (64bit); both versions were executed in OpenSUSE 10.3 x86_64.

  • The Windows versions of OpenFOAM 1.6.0 and 1.7.0 were cross-compiled with mingw-w64, available with blueCFD.

  • The MPI cases were run with dual-core single-machine, with the mesh split in two on the Y axis, thus splitting the bike symmetrically.

  • The mesh and results are compressed with gzip internally (default of the case).

  • The MinGW versions are executed in Windows 7 x64.

  • The mingw-w32 Double Precision version from blueCFD 1.6-2 is also executed in Windows 7 x64 for comparison.

Run times:

The times are in seconds and the results are in the following table:

Single Core

mingw-w32 DP 1.6

mingw-w32 SP

mingw-w32 DP

mingw-w64 DP

Linux 32 (OF 1.6)

Linux 64 (OF 1.6)

snappyHexMesh (write) 161 132 149 106 104 83
snappyHexMesh (total) 284 252 271 231 203 158
simpleFoam 1957 1289 1960 1864 1845 1492
Dual Core            
snappyHexMesh (write) 161 132 149 106 105 83
snappyHexMesh (total) 284 251 271 231 204 158
simpleFoam 1534 947 1539 1471 1375 1280
simpleFoam Performance Ratio Single/Double 1.28 1.36 1.27 1.27 1.34 1.17


  • the version mingw-w32 DP 1.6 is cross-compiled with gcc-4.4.2 mingw-w64;
  • the versions mingw-w32 and mingw-w64 are cross-compiled with gcc-4.4.4 mingw-w64;
  • SP means Single Precision;
  • DP means Double Precision.


  • snappyHexMesh returns two timings: write mesh time and total execution time. These timings reflect the weight that writing the meshes has in the whole process of generating the mesh.

  • The meshes for each run with simpleFoam were generated by the respective version of snappyHexMesh, therefore the results are subjective to the generated mesh. In other words, the mesh for each run isn't the same.

  • These timings are of a single run for each version. Proper testing should have at least 5 runs to do an average.

  • The mingw-w32 SP 1.7-1 version of snappyHexMesh no longer seems to have problems.


As referenced in the previous timings post, the cross-compiled versions suffer from the conversion layer of using the mingw run-time.

The version mingw-w32 DP 1.6 as seen a major improvement, when compared with the previous timings. The only difference between both run times is that the previous one was made with Windows XP x64 and this timing was made with Windows 7 x64! This is due to Windows Seven's improved buffering data mechanism, when compared to XP x64.

The improvements between mingw versions of blueCFD 1.6-2 and 1.7-1 are most likelly thanks to OpenFOAM's improvements from 1.6 to 1.7.0.

The mingw-w64 DP version has only seen improvements when running simpleFoam, when compared to the previous timings. This improvement makes the mingw-w64 64bit version be closer to the 32bit Linux version of OpenFOAM, but keep in mind that the timings for the Linux version are still the ones made using OpenFOAM 1.6.

When dividing the timings between the mingw versions and the Linux versions, we can infer that the mingw versions running in Windows 7 x64 are 5 to 50% slower than the Linux versions. (NOTE: These values haven't been posted here but can easily be calculated.) This means that the mingw versions are useful for setting-up and development of cases to be executed with OpenFOAM, but for long run executions (weeks/months), it's best to use the Linux versions!

Add as favourites (1805) | Quote this article on your site | Views: 37397

Be first to comment this article
RSS comments

Write Comment
  • Please keep the topic of messages relevant to the subject of the article.
  • Personal verbal attacks will be deleted.
  • Please don't use comments to plug your web site. Such material will be removed.
  • If you can't read the security code, click on the image to generate a new one!
BBCode:Web AddressEmail AddressBold TextItalic TextUnderlined TextQuoteCodeOpen ListList ItemClose List

Code:* CodeIs code valid
I wish to be contacted by email regarding additional comments

Powered by AkoComment Tweaked Special Edition v.1.4.6
Captchas Security update made by blueCAPE, based on www.thecaptchas.com
AkoComment © Copyright 2004 by Arthur Konze - www.mamboportal.com
All right reserved

< Prev   Next >
(C) 2007-2024 blueCAPE´s Official Website - Privacy Policy
Joomla! is Free Software released under the GNU/GPL License. Using JoomlaWatch JoomlaWatch Stats 1.2.9 by Matej Koval