|
[Sponsors] |
parallel get wrong results for parallel_computation.py inv_ONERAM6.cfg |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
May 21, 2017, 00:58 |
parallel get wrong results for parallel_computation.py inv_ONERAM6.cfg
|
#1 |
New Member
Join Date: May 2017
Posts: 5
Rep Power: 9 |
hello guys,
Im new for SU2. When learning the tutorial 3 for Inviscid ONERA M6 in the website, I faced problems. here is my question: I get SU2 from SU2-4.1.100-Windows-x64-Single-Threaded.exe using python 2.7 MPICH2 os: windows 8.1 In Serial: entering "SU2_CFD inv_ONERAM6.cfg" at the command line and entering "parallel_computation.py -f inv_ONERAM6.cfg -n 1" at the command line I get the right results as the toturial; In paralell: entering "parallel_computation.py -f inv_ONERAM6.cfg -n 6" at the command line, it prints 6 times every iter I get right CL,CD and some other coefficients, while the wrong Surface_flow.dat(wrong mesh even). It seems that it didnt compute in 6 cores or cut the wing into 6 parts, but compute it 6 times. what is paralell tools, how to check them work well, how I get SU2_DDC(if it work for my question)? someone help me please;thanks a lot. |
|
May 22, 2017, 07:19 |
|
#2 | |
New Member
WANG,Hao
Join Date: Sep 2016
Location: Xi'an China
Posts: 19
Rep Power: 10 |
Quote:
It's obvious that you didn't set the prallel environment properly. Please refer to this thread: Parallel execution of SU2 and:https://github.com/su2code/SU2/wiki/...ne-for-example good luck! |
||
May 23, 2017, 01:35 |
|
#3 |
New Member
Join Date: May 2017
Posts: 5
Rep Power: 9 |
thanks for your reply!
I re-compiled my su2_cfd with mpi and parmetis support, but here came new question,I ran with the command line in parallel, and it printed: job aborted: [ranks] message [0] terminated [1] process exited without calling finalize -----error analysis-------- [1] on admin D:\SU2\SU2_CFD ended prematurely and may have crashed. exit code 0xc0000005 --------error analysis------- SU2 process was terminated by signal '3' (ps: see my attachment for more info) I checked source code, and found words, same with the error tips came out, "poor initial vertex distrabution. processor .. has no vertices as signed to it!" But I dont know why this. Maybe may parmetis has not cut the mesh? How to check parmetis work well? any advice will be appreciated! thanks! Last edited by nuaa_bird; May 30, 2017 at 23:45. Reason: more details about my question. |
|
June 15, 2017, 09:55 |
|
#4 | |
New Member
Join Date: May 2017
Posts: 5
Rep Power: 9 |
Quote:
It appears that my metis was compiled in a wrong way(there ara something wrong with the sln get from cmake(VS2010 compiler) do not work), so I recompiled metis from source using VS 2010. Fortunately, the parallel_compute.py work well after that! hope this can help someone else! |
||
Tags |
inv_oneram6, paralle tools, parallel calculation, su2 |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
simpleFoam parallel | AndrewMortimer | OpenFOAM Running, Solving & CFD | 12 | August 7, 2015 19:45 |
Parallel processing results using FLUENT | haihek | FLUENT | 3 | August 5, 2012 08:55 |
interFoam 2.1.x gives wrong results on low Froude numbers | lt.quibbler | OpenFOAM Running, Solving & CFD | 12 | June 14, 2012 09:06 |
why dynamic mesh method give wrong results? | weiyang1980 | Main CFD Forum | 0 | September 22, 2009 22:06 |
view the results from parallel v3.2 | matej | Siemens | 3 | July 30, 2004 03:13 |