|
[Sponsors] |
July 22, 2016, 17:29 |
Parallel installation incorrectly compiled
|
#1 |
New Member
Alasdair Christison Gray
Join Date: Jul 2016
Location: Edinburgh, Scotland
Posts: 3
Blog Entries: 1
Rep Power: 10 |
I've recently installed SU2 4.2 in parallel in Kubuntu and I'm using MPICH. Everything in the installation seemed to go fine but I'm having a couple of problems:
1 - When I run models with parallel_computation.py the code gets output multiple times, according to the FAQ's on the SU2 site this means that my code hasn't compiled properly 2 - When I try to run shape_optimization.py in parallel it fails just before attempting the second design iteration, in the command line I get this: Traceback (most recent call last): File "/home/ali/SU2-install/bin/shape_optimization.py", line 155, in <module> main() File "/home/ali/SU2-install/bin/shape_optimization.py", line 99, in main options.quiet ) File "/home/ali/SU2-install/bin/shape_optimization.py", line 138, in shape_optimization SU2.opt.SLSQP(project,x0,xb,its,accu) File "/home/ali/SU2-install/bin/SU2/opt/scipy_tools.py", line 133, in scipy_slsqp epsilon = eps ) File "/usr/lib/python2.7/dist-packages/scipy/optimize/slsqp.py", line 206, in fmin_slsqp constraints=cons, **opts) File "/usr/lib/python2.7/dist-packages/scipy/optimize/slsqp.py", line 363, in _minimize_slsqp fx = func(x) File "/usr/lib/python2.7/dist-packages/scipy/optimize/optimize.py", line 289, in function_wrapper return function(*(wrapper_args + args)) File "/home/ali/SU2-install/bin/SU2/opt/scipy_tools.py", line 149, in obj_f obj = project.obj_f(x) File "/home/ali/SU2-install/bin/SU2/opt/project.py", line 213, in obj_f return self._eval(konfig, func,dvs) File "/home/ali/SU2-install/bin/SU2/opt/project.py", line 182, in _eval vals = design._eval(func,*args) File "/home/ali/SU2-install/bin/SU2/eval/design.py", line 142, in _eval vals = eval_func(*inputs) File "/home/ali/SU2-install/bin/SU2/eval/design.py", line 240, in obj_f func = su2func(this_obj,config,state) File "/home/ali/SU2-install/bin/SU2/eval/functions.py", line 85, in function aerodynamics( config, state ) File "/home/ali/SU2-install/bin/SU2/eval/functions.py", line 224, in aerodynamics info = su2run.direct(config) File "/home/ali/SU2-install/bin/SU2/run/direct.py", line 81, in direct SU2_CFD(konfig) File "/home/ali/SU2-install/bin/SU2/run/interface.py", line 115, in CFD run_command( the_Command ) File "/home/ali/SU2-install/bin/SU2/run/interface.py", line 295, in run_command raise exception , message RuntimeError: Path = /home/ali/SU2/TestCases/optimization_euler/steady_naca0012/DESIGNS/DSN_002/DIRECT/, Command = mpirun -n 4 /home/ali/SU2-install/bin/SU2_CFD config_CFD.cfg SU2 process returned error '139' and in the log_direct.out file, this is the final line: mpirun noticed that process rank 2 with PID 11088 on node Steve exited on signal 11 (Segmentation fault). I'm guessing these 2 issues are linked but I'm a complete newbie to Linux and coding in general (bar Matlab and a tiny bit of Python). Can anyone explain what the problem is and how I can go about fixing it in layman's terms Thanks in advance Ali |
|
August 16, 2016, 18:48 |
|
#2 |
Senior Member
Heather Kline
Join Date: Jun 2013
Posts: 309
Rep Power: 14 |
I recommend both trying the shape optimization in serial (1 processor) and recompiling (see instructions: https://github.com/su2code/SU2/wiki/Parallel-Build), making sure that you have all the prerequisites installed and that you are not accidentally calling a previously built executable.
For the shape optimization output, if you look in the subfolders created (DESIGNS/DSN_001/DIRECT, .../DEFORM, etc), there are several log files created. These can be helpful in determining where, and why, the simulation failed. For example, if the first flow solution diverged, you would find that information near the end of the DESIGNS/DSN_001/DIRECT/log_Direct.out file. This could be happening in your log file, just further back in the file before the extra errors printed out by the parallel system - if there are no other errors prior to the one you mentioned, running in serial instead could help narrow down the issue although it would be slower. |
|
August 16, 2016, 19:04 |
|
#3 |
New Member
Alasdair Christison Gray
Join Date: Jul 2016
Location: Edinburgh, Scotland
Posts: 3
Blog Entries: 1
Rep Power: 10 |
Hey Heather, thanks for the reply!
Forgot to post an update but I actually solved this a few days after posting. I checked the log files and sure enough one of the MPI calls for part of the optimisation process was failing and aborting the whole thing. Turns out I had both openMPI and MPICH installed and had referenced the wrong one during the SU2 installation. Uninstalling openMPI and reinstalling MPICH and SU2 solved all my problems. Cheers! Ali |
|
Tags |
compilation, parallel, su2 |
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Error running simpleFoam in parallel | Yuby | OpenFOAM Running, Solving & CFD | 14 | October 7, 2021 04:38 |
Case running in serial, but Parallel run gives error | atmcfd | OpenFOAM Running, Solving & CFD | 18 | March 26, 2016 12:40 |
Parallel FUN3D Installation | jgrisham | Main CFD Forum | 4 | October 21, 2013 14:39 |
parallel support with CGNS format not yet implemented | kirkrich | SU2 | 3 | January 18, 2013 15:39 |
Parallel Computing Classes at San Diego Supercomputer Center Jan. 20-22 | Amitava Majumdar | Main CFD Forum | 0 | January 5, 1999 12:00 |