[msahin.ae00]

I am trying to test the capabilities of the SU2 code for a rotor. I have installed SU2-4.1.0 with the following options.

./configure --prefix=/home/evren/libs/SU2-latest
--with-cxx=/home/evren/libs/openmpi-1.10.2/bin/mpicxx
--with-cc=/home/evren/libs/openmpi-1.10.2/bin/mpicc --enable-mpi

SU2 works fine for small size meshes (approximately 2 million). Then l increased mesh resolution to 4 million and l get the error below. I can run the same mesh on WIndows 8 using the SU2-3.2.2 version. It requires approximately 12GB. SU2-4.1.0 under cygwin also fails on Windows 8 for the same mesh. This may be memory problem since l have only 16GB on Windows machine. However the unix machine has a relatively large memory (250GB).

More recently l have installed the version 4.1.0 on another parallel machine with the following option:

./configure --prefix=/AKDENIZ/users/msahin/SU2-4.1.0 --with-cxx=/usr/mpi/gcc/openmpi-1.8.8/bin/mpicxx --with-cc=/usr/mpi/gcc/openmpi-1.8.8/bin/mpicc --enable-mpi

and it works for 6million element approximately consuming 97 GB.

I also need to mention that l use all hexa elements with local conformal refinement. Sometimes the number of elements connected to a node can be relatively high in 3d. This number is 8 for a Cartesian mesh in 3d.

Best regards

Mehmet SAHIN


ROTOR_TAI]$ ../../libs/openmpi-1.10.2/bin/mpirun -np 4 SU2_CFD rot_caradonna_tung_4.1.0_EU.cfg

-------------------------------------------------------------------------
| ___ _ _ ___ |
| / __| | | |_ ) Release 4.1.0 "Cardinal" |
| \__ \ |_| |/ / |
| |___/\___//___| Suite (Computational Fluid Dynamics Code) |
|
| |
-------------------------------------------------------------------------
| SU2 Lead Dev.: Dr. Francisco Palacios, Francisco.D.Palacios@boeing.com|
| Dr. Thomas D. Economon, economon@stanford.edu |
-------------------------------------------------------------------------
| SU2 Developers: |
| - Prof. Juan J. Alonso's group at Stanford University. |
| - Prof. Piero Colonna's group at Delft University of Technology. |
| - Prof. Nicolas R. Gauger's group at Kaiserslautern U. of Technology.
| |
| - Prof. Alberto Guardone's group at Polytechnic University of Milan. |
| - Prof. Rafael Palacios' group at Imperial College London. |
-------------------------------------------------------------------------
| Copyright (C) 2012-2015 SU2, the open-source CFD code. |
| |
| SU2 is free software; you can redistribute it and/or |
| modify it under the terms of the GNU Lesser General Public |
| License as published by the Free Software Foundation; either |
| version 2.1 of the License, or (at your option) any later version. |
| |
| SU2 is distributed in the hope that it will be useful, |
| but WITHOUT ANY WARRANTY; without even the implied warranty of |
| MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| Lesser General Public License for more details. |
| |
| You should have received a copy of the GNU Lesser General Public |
| License along with SU2. If not, see <http://www.gnu.org/licenses/>. |
-------------------------------------------------------------------------

------------------------ Physical Case Definition ----------------------- Compressible Euler equations.
Mach number: 0.
Angle of attack (AoA): 0 deg, and angle of sideslip (AoS): 0 deg.
Performing a dynamic mesh simulation: rotating reference frame.
No restart solution, use the values at infinity (freestream).
Dimensional simulation.
The reference length/area (force coefficient) is 1.
The reference length (moment computation) is 1.
Reference origin (moment computation) is (0.25, 0, 0).
Surface(s) where the force coefficients are evaluated: airfoil.
Surface(s) plotted in the output file: airfoil.
Surface(s) belonging to the Fluid-Structure Interaction problem:
Input mesh file name: TAI_ROTOR_MESH.su2

---------------------- Space Numerical Integration ---------------------- Jameson-Schmidt-Turkel scheme for the flow inviscid terms.
JST viscous coefficients (1st, 2nd & 4th): 0.15, 0.5, 0.02.
The method includes a grid stretching correction (p = 0.3).
Second order integration.
Gradient computation using Green-Gauss theorem.

---------------------- Time Numerical Integration ----------------------- Local time stepping (steady state simulation).
Euler implicit method for the flow equations.
CFL adaptation. Factor down: 1.5, factor up: 0.5,
lower limit: 1, upper limit: 100.
Courant-Friedrichs-Lewy number: 6

------------------------- Convergence Criteria -------------------------- Maximum number of iterations: 99999.
Reduce the density residual 6 orders of magnitude.
The minimum bound for the density residual is 10^(-12).
Start convergence criteria at iteration 10.

-------------------------- Output Information --------------------------- Writing a flow solution every 100 iterations.
Writing the convergence history every 1 iterations.
The output file format is Tecplot ASCII (.dat).
Convergence history file name: history.
Forces breakdown file name: forces_breakdown.dat.
Surface flow coefficients file name: surface_flow.
Flow variables file name: flow.
Restart flow file name: restart_flow.dat.

------------------- Config File Boundary Information -------------------- Euler wall boundary marker(s): airfoil.
Far-field boundary marker(s): farfield.

---------------------- Read Grid File Information ----------------------- Three dimensional problem.
3882215 points before parallel partitioning.
Performing linear partitioning of the grid nodes.
3816766 interior elements before parallel partitioning.
Distributing elements across all ranks.
Calling the partitioning functions.
Building the graph adjacency structure.
2 surface markers.
118736 boundary elements in index 0 (Marker = airfoil).
7526 boundary elements in index 1 (Marker = farfield).
Calling ParMETIS...
[ferrel:23100] *** Process received signal *** [ferrel:23100] Signal: Segmentation fault (11) [ferrel:23100] Signal code: Address not mapped (1) [ferrel:23100] Failing at address: 0xffffffff33b87320 [ferrel:23100] [ 0] /lib64/libpthread.so.0[0x3b2400f500]
[ferrel:23100] [ 1] SU2_CFD[0xdc2df8]
[ferrel:23100] [ 2] SU2_CFD[0xdaf049]
[ferrel:23100] [ 3] SU2_CFD[0xcb4d07]
[ferrel:23100] [ 4] SU2_CFD[0xa1be6a]
[ferrel:23100] [ 5] /lib64/libc.so.6(__libc_start_main+0xfd)[0x3b2381ecdd]
[ferrel:23100] [ 6] SU2_CFD[0x407919]
[ferrel:23100] *** End of error message ***
--------------------------------------------------------------------------
mpirun noticed that process rank 0 with PID 23100 on node ferrel exited on signal 11 (Segmentation fault).
--------------------------------------------------------------------------

[fpalacios]

Dear Mehmet,

As you noticed, the memory issue is in the partitioning step.

Few days ago, Tom Economon (Stanford) fixed an important memory issue in the partitioning. Could you please try again using the version that you have in

https://github.com/su2code/SU2/tree/develop

Please let us know if this new version works.

Best,
Francisco

PS.- Despite you have a lot of memory available in your machine, the problem is that it is possible that a computational node is taking too much of his own memory (I'm guessing that your machine has distributed memory).

[msahin.ae00]

Dear Francisco

I have reinstalled the latest version and the problem still persists on that machine. The error message is bellow.

Best

Mehmet

---------------------- Read Grid File Information -----------------------
Three dimensional problem.
4793727 points before parallel partitioning.
Performing linear partitioning of the grid nodes.
4659974 interior elements before parallel partitioning.
Distributing elements across all ranks.
Calling the partitioning functions.
Building the graph adjacency structure.
2 surface markers.
215476 boundary elements in index 0 (Marker = airfoil).
45702 boundary elements in index 1 (Marker = farfield).
Calling ParMETIS...
*** glibc detected *** /home/evren/libs/SU2-develop/bin/SU2_CFD: malloc(): memory corruption: 0x0000000007b1a4c0 ***
*** glibc detected *** /home/evren/libs/SU2-develop/bin/SU2_CFD: malloc(): memory corruption: 0x0000000007b1a4c0 ***
[ferrel:18997] *** An error occurred in MPI_Waitall
[ferrel:18997] *** reported by process [256088539137,140733193388042]
[ferrel:18997] *** on communicator MPI COMMUNICATOR 3 DUP FROM 0
[ferrel:18997] *** MPI_ERR_TRUNCATE: message truncated
[ferrel:18997] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[ferrel:18997] *** and potentially your MPI job)
[ferrel.:18985] 10 more processes have sent help message help-mpi-errors.txt / mpi_errors_are_fatal
[ferrel.:18985] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
ROTOR_S76]$

[mhd_mrt]

Hello All,

I was wondering why am I getting an error after my parallel simulation on my laptop. The simulations runs without a problem in SERIAL but when I do :
parallel_computation.py -f XXXXX.cfg -n 4 it finished with the CFD calculations but apparently gets stuck at the SU2_SOL ??

Any comments or suggestions are appreciated.

Mehdi






-------------------------------------------------------------------------

History file, closed.

------------------------- Solver Postprocessing -------------------------
Deleted CNumerics container.
Deleted CIntegration container.
Deleted CSolver container.
Deleted CIteration container.
Deleted CInterpolator container.
Deleted CTransfer container.
Deleted CGeometry container.
Deleted CFreeFormDefBox class.
Deleted CSurfaceMovement class.
Deleted CVolumetricMovement class.
Deleted CConfig container.
Deleted COutput class.
-------------------------------------------------------------------------

Completed in 2.809731 seconds on 2 cores.

------------------------- Exit Success (SU2_CFD) ------------------------


-------------------------------------------------------------------------
| ___ _ _ ___ |
| / __| | | |_ ) Release 5.0.0 "Raven" |
| \__ \ |_| |/ / |
| |___/\___//___| Suite (Solution Exporting Code) |
| |
-------------------------------------------------------------------------
| SU2 Lead Dev.: Dr. Francisco Palacios, Francisco.D.Palacios@boeing.com|
| Dr. Thomas D. Economon, economon@stanford.edu |
-------------------------------------------------------------------------
| SU2 Developers: |
| - Prof. Juan J. Alonso's group at Stanford University. |
| - Prof. Piero Colonna's group at Delft University of Technology. |
| - Prof. Nicolas R. Gauger's group at Kaiserslautern U. of Technology. |
| - Prof. Alberto Guardone's group at Polytechnic University of Milan. |
| - Prof. Rafael Palacios' group at Imperial College London. |
| - Prof. Edwin van der Weide's group at the University of Twente. |
| - Prof. Vincent Terrapon's group at the University of Liege. |
-------------------------------------------------------------------------
| Copyright (C) 2012-2017 SU2, the open-source CFD code. |
| |
| SU2 is free software; you can redistribute it and/or |
| modify it under the terms of the GNU Lesser General Public |
| License as published by the Free Software Foundation; either |
| version 2.1 of the License, or (at your option) any later version. |
| |
| SU2 is distributed in the hope that it will be useful, |
| but WITHOUT ANY WARRANTY; without even the implied warranty of |
| MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| Lesser General Public License for more details. |
| |
| You should have received a copy of the GNU Lesser General Public |
| License along with SU2. If not, see <http://www.gnu.org/licenses/>. |
-------------------------------------------------------------------------

------------------------ Physical Case Definition -----------------------
Input mesh file name: Driven.su2

-------------------------- Output Information ---------------------------
The output file format is Paraview ASCII (.vtk).
Flow variables file name: flow.

------------------- Config File Boundary Information --------------------
Euler wall boundary marker(s): wall.
Supersonic inlet boundary marker(s): left, right.

---------------------- Read Grid File Information -----------------------
Two dimensional problem.
1500 points before parallel partitioning.
Performing linear partitioning of the grid nodes.
998 interior elements before parallel partitioning.
Calling the partitioning functions.
Building the graph adjacency structure.
Distributing elements across all ranks.
3 surface markers.
2 boundary elements in index 0 (Marker = left).
2 boundary elements in index 1 (Marker = right).
998 boundary elements in index 2 (Marker = wall).
Calling ParMETIS...
Finished partitioning using ParMETIS (3 edge cuts).
Communicating partition data and creating halo layers.
1000 interior elements including halo cells.
1000 quadrilaterals.
1506 vertices including ghost points.
Establishing MPI communication patterns.
Identify vertices.

------------------------- Solution Postprocessing -----------------------

================================================== =================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 77625 RUNNING AT Mehdi.local
= EXIT CODE: 11
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
================================================== =================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault: 11 (signal 11)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions
Traceback (most recent call last):
File "/Users/mhd/SU2/bin/parallel_computation.py", line 112, in <module>
main()
File "/Users/mhd/SU2/bin/parallel_computation.py", line 63, in main
options.compute )
File "/Users/mhd/SU2/bin/parallel_computation.py", line 98, in parallel_computation
info = SU2.run.merge(config)
File "/Users/mhd/SU2/bin/SU2/run/merge.py", line 76, in merge
merge_solution(konfig)
File "/Users/mhd/SU2/bin/SU2/run/merge.py", line 102, in merge_solution
SU2_SOL( config )
File "/Users/mhd/SU2/bin/SU2/run/interface.py", line 234, in SOL
run_command( the_Command )
File "/Users/mhd/SU2/bin/SU2/run/interface.py", line 297, in run_command
raise exception , message
RuntimeError: Path = /Users/mhd/SU2_Tutorials/shocktube/SU2_Real Shock tube/V9/DrivenV3/1-Domain_Best/AD_V6/,
Command = mpirun -n 2 /Users/mhd/SU2/bin/SU2_SOL config_SOL.cfg
SU2 process returned error '11'

@Mehdi/AD_V6 :

[mhd_mrt]

Problem solved!

The compiling step missed the MPI config.

Thanks!