|
[Sponsors] |
October 4, 2009, 09:02 |
serial run fine, but parallel run diverged
|
#1 |
Senior Member
Pei-Ying Hsieh
Join Date: Mar 2009
Posts: 334
Rep Power: 18 |
Dear OF experts,
I modified the interFoam solver for my purpose. The serial run ran fine, but, the parallel run diverged quickly after about 100 time steps (the serial run took several thousand time steps). I tried different decomposition, but, all failed. Any suggestion will be highly appreciated. Pei |
|
October 5, 2009, 17:15 |
|
#2 |
Member
Ola Widlund
Join Date: Mar 2009
Location: Sweden
Posts: 87
Rep Power: 17 |
Hi Pei,
It's propably not this simple, but you never know: I assume you have changed the flag "floatTransfer" in the main OpenFOAM controlDict? Otherwise, if set, all mpi communication would be in single precision. That can be a killer... /Ola |
|
October 6, 2009, 09:33 |
|
#3 |
Senior Member
Pei-Ying Hsieh
Join Date: Mar 2009
Posts: 334
Rep Power: 18 |
Hi, Ola,
I have floatTransfer set to 0, is this double precision? Pei |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Script to Run Parallel Jobs in Rocks Cluster | asaha | OpenFOAM Running, Solving & CFD | 12 | July 4, 2012 23:51 |
nonNewtonianIcoFoam - problem with parallel run | chris_sev | OpenFOAM Bugs | 4 | April 1, 2009 10:13 |
Help: Serial code to parallel but even slower | Zonexo | Main CFD Forum | 4 | May 14, 2008 11:26 |
Cant run in parallel on two nodes using OpenMPI | CHristofer | Main CFD Forum | 0 | October 26, 2007 10:54 |
How to run parallel in ICEM_CFD? | Kiddo | Main CFD Forum | 2 | January 24, 2005 09:53 |