[CSMDakota]

Is anybody having issues initializing solutions with cental schemes in 7.1.1?

I just updated from v7.0.0 to v7.1.1, and to check my installation I re-ran a case I had just solved in the former (but I did build a fresh .cfg file from the 7.1.1 template).

• My setup uses the JST central scheme, which now fails in 7.1.1 during solution startup right after the Jacobian initialization (see below).
• The case starts OK if I use an Upwind scheme like ROE or SLAU2, but again fails using Lax-Friedrich.

Is there something with the new vectorization that could be causing this?

• It crashes regardless if USE_VECTORIZATION= Yes or No
• I also tried a range of CENTAL_JACOBIAN_FIX_FACTOR= values (0 to 8)

After reading https://github.com/su2code/SU2/pull/1022, I rebuilt with the following:

export CXXFLAGS="-O2 -funroll-loops -ffast-math -march=ivybridge -mtune=ivybridge" (I have ivybridge processors which I think support 256 bit AVX?)
./meson.py build --reconfigure -Dwith-omp=true

But I still find the same behavior, here is the crash-point during startup when running in parallel:

Initialize Jacobian structure (Navier-Stokes). MG level: 0.
Initialize Jacobian structure (SA model).
[z820:12833] *** Process received signal ***
[z820:12833] Signal: Segmentation fault (11)
[z820:12833] Signal code: Address not mapped (1)
[z820:12833] Failing at address: (nil)

If I run in serial I get a bit more information:

Initialize Jacobian structure (SA model).

------------------- Numerics Preprocessing ( Zone 0 ) -------------------

----------------- Integration Preprocessing ( Zone 0 ) ------------------

------------------- Iteration Preprocessing ( Zone 0 ) ------------------
Euler/Navier-Stokes/RANS fluid iteration.

------------------------------ Begin Solver -----------------------------

Simulation Run using the Single-zone Driver


Error in "virtual void CEulerSolver::InstantiateEdgeNumerics(const CSolver* const*, const CConfig*)":
-------------------------------------------------------------------------
Low-Mach correction is not supported with vectorization.
------------------------------ Error Exit -------------------------------

This error is interesting in that I had set USE_VECTORIZATION = NO.

I also suppose my build configuration for AVX did not work; when a ROE setup starts solving it reports (but otherwise begins solving OK):

"WARNING: SU2 was not compiled for an AVX-capable architecture."

Thanks,
Brandon

[pcg]

Hi Brandon,
Do you have low Mach correction options in the config when you try to use JST?
That feature is not compatible with centered schemes (or rather it does not affect them), and at the moment not compatible with vectorized upwind schemes, but all centered schemes are vectorized now... And so I took the blue pill and threw an error if low Mach + vectorization.
It should work if you disable that option, low Mach preconditioning is compatible with any scheme.

Regards,
Pedro

[flavio73]

Hi Pedro
I have a similar issue. Installed the latest version on a machine running linux mint.
I first used the provided pre-compiled version and then I used the source code and compiled it o my machine. I am using the JST central scheme. It works only if I run it in serial. When I try to run it using mpi I got an error during the initialisation of the Jacobian.
Both low Mach correction and vectorisation are disabled. On the other hand if I select ROE the code runs.

I attach the error I got.
What do you suggest ?

Thanks for the help
Flavio


-------------------------------------------------------------------------
| ___ _ _ ___ |
| / __| | | |_ ) Release 7.1.1 "Blackbird" |
| \__ \ |_| |/ / |
| |___/\___//___| Suite (Computational Fluid Dynamics Code) |
| |
-------------------------------------------------------------------------
| SU2 Project Website: https://su2code.github.io |
| |
| The SU2 Project is maintained by the SU2 Foundation |
| (http://su2foundation.org) |
-------------------------------------------------------------------------
| Copyright 2012-2020, SU2 Contributors |
| |
| 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/>. |
-------------------------------------------------------------------------

Parsing config file for zone 0

----------------- Physical Case Definition ( Zone 0 ) -------------------
Compressible RANS equations.
Turbulence model: Spalart Allmaras
Hybrid RANS/LES: No Hybrid RANS/LES
Mach number: 0.35.
Angle of attack (AoA): 1 deg, and angle of sideslip (AoS): 0 deg.
Reynolds number: 6.983e+06. Reference length 1.
No restart solution, use the values at infinity (freestream).
Dimensional simulation.
The reference area is 1 m^2.
The semi-span will be computed using the max y(3D) value.
The reference length is 1 m.
Reference origin for moment evaluation is (0.25, 0, 0).
Surface(s) where the force coefficients are evaluated: Wing.

Surface(s) plotted in the output file: Wing.
Surface(s) to be analyzed in detail: Wing.
Input mesh file name: Mesh_28012021.su2

--------------- Space Numerical Integration ( Zone 0 ) ------------------
Jameson-Schmidt-Turkel scheme (2nd order in space) for the flow inviscid terms.
JST viscous coefficients (2nd & 4th): 0.5, 0.02.
The method includes a grid stretching correction (p = 0.3).
Scalar upwind solver for the turbulence model.
First order integration in space.
Average of gradients with correction (viscous flow terms).
Average of gradients with correction (viscous turbulence terms).
Gradient for upwind reconstruction: unweighted Least-Squares.
Gradient for viscous and source terms: Green-Gauss.

--------------- Time Numerical Integration ( Zone 0 ) ------------------
Local time stepping (steady state simulation).
Euler implicit method for the flow equations.
FGMRES is used for solving the linear system.
Using a ILU(0) preconditioning.
Convergence criteria of the linear solver: 1e-06.
Max number of linear iterations: 5.
CFL adaptation. Factor down: 0.1, factor up: 2,
lower limit: 10, upper limit: 1e+10,
acceptable linear residual: 0.001.
Courant-Friedrichs-Lewy number: 1
Euler implicit time integration for the turbulence model.

------------------ Convergence Criteria ( Zone 0 ) ---------------------
Maximum number of solver subiterations: 999999.
Begin convergence monitoring at iteration 10.
Residual minimum value: 1e-8.
Cauchy series min. value: 1e-10.
Number of Cauchy elements: 100.
Begin windowed time average at iteration 500.
Begin time convergence monitoring at iteration 500.
Time cauchy series min. value: 0.001.
Number of Cauchy elements: 10.

-------------------- Output Information ( Zone 0 ) ----------------------
Writing solution files every 100 iterations.
Writing the convergence history file every 1 inner iterations.
Writing the screen convergence history every 1 inner iterations.
The tabular file format is Tecplot (.dat).
Convergence history file name: history.
Forces breakdown file name: forces_breakdown.dat.
Surface file name: surface_flow.
Volume file name: flow.
Restart file name: restart_flow.dat.

------------- Config File Boundary Information ( Zone 0 ) ---------------
+-----------------------------------------------------------------------+
| Marker Type| Marker Name|
+-----------------------------------------------------------------------+
| Far-field| FF|
+-----------------------------------------------------------------------+
| Symmetry plane| Symmetry|
+-----------------------------------------------------------------------+
| Heat flux wall| Wing|
+-----------------------------------------------------------------------+

-------------------- Output Preprocessing ( Zone 0 ) --------------------
Screen output fields: INNER_ITER, DRAG, LIFT, AVG_CFL, RMS_DENSITY, RMS_MOMENTUM-X, RMS_ENERGY
History output group(s): ITER, AERO_COEFF, RMS_RES
Convergence field(s): RMS_DENSITY
Ignoring Time Convergence Field(s): TAVG_DRAG TAVG_LIFT
Warning: No (valid) fields chosen for time convergence monitoring. Time convergence monitoring inactive.
Volume output fields: COORDINATES, SOLUTION, PRIMITIVE

------------------- Geometry Preprocessing ( Zone 0 ) -------------------
Three dimensional problem.
6731542 grid points before partitioning.
10300559 volume elements before partitioning.
3 surface markers.
7114 boundary elements in index 0 (Marker = FF).
12897 boundary elements in index 1 (Marker = Symmetry).
211803 boundary elements in index 2 (Marker = Wing).
Executing the partitioning functions.
Building the graph adjacency structure.
Calling ParMETIS... graph partitioning complete (136138 edge cuts).
Distributing ParMETIS coloring.
Rebalancing vertices.
Rebalancing volume element connectivity.
Rebalancing markers and surface elements.
6942857 vertices including ghost points.
10455167 interior elements including halo cells.
3255746 tetrahedra.
5273806 hexahedra.
1762172 prisms.
163443 pyramids.
Establishing MPI communication patterns.
Setting point connectivity.
Renumbering points (Reverse Cuthill McKee Ordering).
Recomputing point connectivity.
Setting element connectivity.
Checking the numerical grid orientation.
There has been a re-orientation of 1762172 PRISM volume elements.
All surface elements are correctly orientend.
Identifying edges and vertices.
Setting the control volume structure.
Volume of the computational grid: 23898.2.
Searching for the closest normal neighbors to the surfaces.
Storing a mapping from global to local point index.
Compute the surface curvature.
Max K: 186167. Mean K: 110.52. Standard deviation K: 1508.87.
Checking for periodicity.
Computing mesh quality statistics for the dual control volumes.
+--------------------------------------------------------------+
| Mesh Quality Metric| Minimum| Maximum|
+--------------------------------------------------------------+
| Orthogonality Angle (deg.)| 4.84843| 89.9999|
| CV Face Area Aspect Ratio| 1.0066| 61327.6|
| CV Sub-Volume Ratio| 1| 15713.3|
+--------------------------------------------------------------+
Finding max control volume width.
Communicating number of neighbors.
Semi-span length = 4.72553 m.
Wetted area = 10.2427 m^2.
Area projection in the x-plane = 0.811782 m^2, y-plane = 0.960273 m^2, z-plane = 4.30359 m^2.
Max. coordinate in the x-direction = 5.87959 m, y-direction = 4.72553 m, z-direction = 2.5882 m.
Min. coordinate in the x-direction = 3.65827 m, y-direction = -1.578e-10 m, z-direction = 1.49489 m.
Checking if Euler & Symmetry markers are straight/plane:
Boundary marker Symmetry is NOT a single plane.
Computing wall distances.

-------------------- Solver Preprocessing ( Zone 0 ) --------------------
Viscous flow: Computing pressure using the ideal gas law
based on the free-stream temperature and a density computed
from the Reynolds number.
Force coefficients computed using free-stream values.

-- Models:
+------------------------------------------------------------------------------+
| Viscosity Model| Conductivity Model| Fluid Model|
+------------------------------------------------------------------------------+
| SUTHERLAND| CONSTANT_PRANDTL| STANDARD_AIR|
+------------------------------------------------------------------------------+
-- Fluid properties:
+------------------------------------------------------------------------------+
| Name| Dim. value| Ref. value| Unit|Non-dim. value|
+------------------------------------------------------------------------------+
| Ref. Viscosity| 1.716e-05| 1| N.s/m^2| 1.716e-05|
| Sutherland Temp.| 273.15| 1| K| 273.15|
| Sutherland Const.| 110.4| 1| K| 110.4|
+------------------------------------------------------------------------------+
| Prandtl (Lam.)| -| -| -| 0.72|
| Prandtl (Turb.)| -| -| -| 0.9|
+------------------------------------------------------------------------------+
| Gas Constant| 287.058| 1| N.m/kg.K| 287.058|
| Spec. Heat Ratio| -| -| -| 1.4|
+------------------------------------------------------------------------------+
-- Initial and free-stream conditions:
+------------------------------------------------------------------------------+
| Name| Dim. value| Ref. value| Unit|Non-dim. value|
+------------------------------------------------------------------------------+
| Static Pressure| 86773.4| 1| Pa| 86773.4|
| Density| 1.04906| 1| kg/m^3| 1.04906|
| Temperature| 288.15| 1| K| 288.15|
| Total Energy| 213882| 1| m^2/s^2| 213882|
| Velocity-X| 119.086| 1| m/s| 119.086|
| Velocity-Y| 0| 1| m/s| 0|
| Velocity-Z| 2.07865| 1| m/s| 2.07865|
| Velocity Magnitude| 119.104| 1| m/s| 119.104|
+------------------------------------------------------------------------------+
| Viscosity| 1.7893e-05| 1| N.s/m^2| 1.7893e-05|
| Conductivity| -| 1| W/m^2.K| -|
| Turb. Kin. Energy| 53.1966| 1| m^2/s^2| 53.1966|
| Spec. Dissipation| 311889| 1| 1/s| 311889|
+------------------------------------------------------------------------------+
| Mach Number| -| -| -| 0.35|
| Reynolds Number| -| -| -| 6.983e+06|
+------------------------------------------------------------------------------+
Initialize Jacobian structure (Navier-Stokes). MG level: 0.
Initialize Jacobian structure (SA model).
[fd2:20952] *** Process received signal ***
[fd2:20952] Signal: Segmentation fault (11)
[fd2:20952] Signal code: Address not mapped (1)
[fd2:20952] Failing at address: (nil)
[fd2:20952] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7fe5a49e7980]
[fd2:20952] [ 1] SU2_CFD(+0x48921b)[0x55744993721b]
[fd2:20952] [ 2] SU2_CFD(+0x3b15df)[0x55744985f5df]
[fd2:20952] [ 3] SU2_CFD(+0x35a0c3)[0x5574498080c3]
[fd2:20952] [ 4] SU2_CFD(+0x3aafa1)[0x557449858fa1]
[fd2:20952] [ 5] SU2_CFD(+0x398683)[0x557449846683]
[fd2:20952] [ 6] SU2_CFD(+0x24b4af)[0x5574496f94af]
[fd2:20952] [ 7] SU2_CFD(+0x24bfd4)[0x5574496f9fd4]
[fd2:20952] [ 8] SU2_CFD(+0x24cd6b)[0x5574496fad6b]
[fd2:20952] [ 9] SU2_CFD(+0x11f12c)[0x5574495cd12c]
[fd2:20952] [10] SU2_CFD(+0x12f25c)[0x5574495dd25c]
[fd2:20952] [11] SU2_CFD(+0x134b1f)[0x5574495e2b1f]
[fd2:20952] [12] SU2_CFD(+0xe0066)[0x55744958e066]
[fd2:20952] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7fe5a4605bf7]
[fd2:20952] [14] SU2_CFD(+0xf0aca)[0x55744959eaca]
[fd2:20952] *** End of error message ***
[fd2:20881] *** Process received signal ***
[fd2:20881] Signal: Segmentation fault (11)
[fd2:20881] Signal code: Address not mapped (1)
[fd2:20881] Failing at address: (nil)
[fd2:20881] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f9f95f95980]
[fd2:20881] [ 1] SU2_CFD(+0x48921b)[0x556715cb721b]
[fd2:20881] [ 2] SU2_CFD(+0x3b15df)[0x556715bdf5df]
[fd2:20881] [ 3] SU2_CFD(+0x35a0c3)[0x556715b880c3]
[fd2:20881] [ 4] SU2_CFD(+0x3aafa1)[0x556715bd8fa1]
[fd2:20881] [ 5] SU2_CFD(+0x398683)[0x556715bc6683]
[fd2:20881] [ 6] SU2_CFD(+0x24b4af)[0x556715a794af]
[fd2:20881] [ 7] SU2_CFD(+0x24bfd4)[0x556715a79fd4]
[fd2:20881] [ 8] SU2_CFD(+0x24cd6b)[0x556715a7ad6b]
[fd2:20881] [ 9] SU2_CFD(+0x11f12c)[0x55671594d12c]
[fd2:20881] [10] SU2_CFD(+0x12f25c)[0x55671595d25c]
[fd2:20881] [11] SU2_CFD(+0x134b1f)[0x556715962b1f]
[fd2:20881] [12] SU2_CFD(+0xe0066)[0x55671590e066]
[fd2:20881] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7f9f95bb3bf7]
[fd2:20881] [14] SU2_CFD(+0xf0aca)[0x55671591eaca]
[fd2:20881] *** End of error message ***
[fd2:20882] *** Process received signal ***
[fd2:20882] Signal: Segmentation fault (11)
[fd2:20882] Signal code: Address not mapped (1)
[fd2:20882] Failing at address: (nil)
[fd2:20882] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f14f10a9980]
[fd2:20882] [ 1] SU2_CFD(+0x48921b)[0x55caa715c21b]
[fd2:20882] [ 2] SU2_CFD(+0x3b15df)[0x55caa70845df]
[fd2:20882] [ 3] SU2_CFD(+0x35a0c3)[0x55caa702d0c3]
[fd2:20882] [ 4] SU2_CFD(+0x3aafa1)[0x55caa707dfa1]
[fd2:20882] [ 5] SU2_CFD(+0x398683)[0x55caa706b683]
[fd2:20882] [ 6] SU2_CFD(+0x24b4af)[0x55caa6f1e4af]
[fd2:20882] [ 7] SU2_CFD(+0x24bfd4)[0x55caa6f1efd4]
[fd2:20882] [ 8] SU2_CFD(+0x24cd6b)[0x55caa6f1fd6b]
[fd2:20882] [ 9] SU2_CFD(+0x11f12c)[0x55caa6df212c]
[fd2:20882] [10] SU2_CFD(+0x12f25c)[0x55caa6e0225c]
[fd2:20882] [11] SU2_CFD(+0x134b1f)[0x55caa6e07b1f]
[fd2:20882] [12] SU2_CFD(+0xe0066)[0x55caa6db3066]
[fd2:20882] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7f14f0cc7bf7]
[fd2:20882] [14] SU2_CFD(+0xf0aca)[0x55caa6dc3aca]
[fd2:20882] *** End of error message ***
[fd2:21126] *** Process received signal ***
[fd2:21126] Signal: Segmentation fault (11)
[fd2:21126] Signal code: Address not mapped (1)
[fd2:21126] Failing at address: (nil)
[fd2:21126] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f21fec94980]
[fd2:21126] [ 1] SU2_CFD(+0x48921b)[0x5646ecf9e21b]
[fd2:21126] [ 2] SU2_CFD(+0x3b15df)[0x5646ecec65df]
[fd2:21126] [ 3] SU2_CFD(+0x35a0c3)[0x5646ece6f0c3]
[fd2:21126] [ 4] SU2_CFD(+0x3aafa1)[0x5646ecebffa1]
[fd2:21126] [ 5] SU2_CFD(+0x398683)[0x5646ecead683]
[fd2:21126] [ 6] SU2_CFD(+0x24b4af)[0x5646ecd604af]
[fd2:21126] [ 7] SU2_CFD(+0x24bfd4)[0x5646ecd60fd4]
[fd2:21126] [ 8] SU2_CFD(+0x24cd6b)[0x5646ecd61d6b]
[fd2:21126] [ 9] SU2_CFD(+0x11f12c)[0x5646ecc3412c]
[fd2:21126] [10] SU2_CFD(+0x12f25c)[0x5646ecc4425c]
[fd2:21126] [11] SU2_CFD(+0x134b1f)[0x5646ecc49b1f]
[fd2:21126] [12] SU2_CFD(+0xe0066)[0x5646ecbf5066]
[fd2:21126] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7f21fe8b2bf7]
[fd2:21126] [14] SU2_CFD(+0xf0aca)[0x5646ecc05aca]
[fd2:21126] *** End of error message ***
[fd2:21208] *** Process received signal ***
[fd2:21208] Signal: Segmentation fault (11)
[fd2:21208] Signal code: Address not mapped (1)
[fd2:21208] Failing at address: (nil)
[fd2:21208] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7fc4efd5a980]
[fd2:21208] [ 1] SU2_CFD(+0x48921b)[0x55df349d921b]
[fd2:21208] [ 2] SU2_CFD(+0x3b15df)[0x55df349015df]
[fd2:21208] [ 3] SU2_CFD(+0x35a0c3)[0x55df348aa0c3]
[fd2:21208] [ 4] SU2_CFD(+0x3aafa1)[0x55df348fafa1]
[fd2:21208] [ 5] SU2_CFD(+0x398683)[0x55df348e8683]
[fd2:21208] [ 6] SU2_CFD(+0x24b4af)[0x55df3479b4af]
[fd2:21208] [ 7] SU2_CFD(+0x24bfd4)[0x55df3479bfd4]
[fd2:21208] [ 8] SU2_CFD(+0x24cd6b)[0x55df3479cd6b]
[fd2:21208] [ 9] SU2_CFD(+0x11f12c)[0x55df3466f12c]
[fd2:21208] [10] SU2_CFD(+0x12f25c)[0x55df3467f25c]
[fd2:21208] [11] SU2_CFD(+0x134b1f)[0x55df34684b1f]
[fd2:21208] [12] SU2_CFD(+0xe0066)[0x55df34630066]
[fd2:21208] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7fc4ef978bf7]
[fd2:21208] [14] SU2_CFD(+0xf0aca)[0x55df34640aca]
[fd2:21208] *** End of error message ***
[fd2:20880] *** Process received signal ***
[fd2:20880] Signal: Segmentation fault (11)
[fd2:20880] Signal code: Address not mapped (1)
[fd2:20880] Failing at address: (nil)
[fd2:20880] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7fad22745980]
[fd2:20880] [ 1] SU2_CFD(+0x48921b)[0x55fd4384e21b]
[fd2:20880] [ 2] SU2_CFD(+0x3b15df)[0x55fd437765df]
[fd2:20880] [ 3] SU2_CFD(+0x35a0c3)[0x55fd4371f0c3]
[fd2:20880] [ 4] SU2_CFD(+0x3aafa1)[0x55fd4376ffa1]
[fd2:20880] [ 5] SU2_CFD(+0x398683)[0x55fd4375d683]
[fd2:20880] [ 6] SU2_CFD(+0x24b4af)[0x55fd436104af]
[fd2:20880] [ 7] SU2_CFD(+0x24bfd4)[0x55fd43610fd4]
[fd2:20880] [ 8] SU2_CFD(+0x24cd6b)[0x55fd43611d6b]
[fd2:20880] [ 9] SU2_CFD(+0x11f12c)[0x55fd434e412c]
[fd2:20880] [10] SU2_CFD(+0x12f25c)[0x55fd434f425c]
[fd2:20880] [11] SU2_CFD(+0x134b1f)[0x55fd434f9b1f]
[fd2:20880] [12] SU2_CFD(+0xe0066)[0x55fd434a5066]
[fd2:20880] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7fad22363bf7]
[fd2:20880] [14] SU2_CFD(+0xf0aca)[0x55fd434b5aca]
[fd2:20880] *** End of error message ***
[fd2:20902] *** Process received signal ***
[fd2:20902] Signal: Segmentation fault (11)
[fd2:20902] Signal code: Address not mapped (1)
[fd2:20902] Failing at address: (nil)
[fd2:20902] [ 0] [fd2:21150] *** Process received signal ***
[fd2:21150] Signal: Segmentation fault (11)
[fd2:21150] Signal code: Address not mapped (1)
[fd2:21150] Failing at address: (nil)
/lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f3abb0a0980]
[fd2:20902] [ 1] SU2_CFD(+0x48921b)[0x5596a8a4e21b]
[fd2:20902] [ 2] SU2_CFD(+0x3b15df)[0x5596a89765df]
[fd2:20902] [ 3] SU2_CFD(+0x35a0c3)[0x5596a891f0c3]
[fd2:20902] [ 4] SU2_CFD(+0x3aafa1)[0x5596a896ffa1]
[fd2:20902] [ 5] SU2_CFD(+0x398683)[0x5596a895d683]
[fd2:20902] [ 6] SU2_CFD(+0x24b4af)[0x5596a88104af]
[fd2:20902] [ 7] SU2_CFD(+0x24bfd4)[0x5596a8810fd4]
[fd2:20902] [ 8] SU2_CFD(+0x24cd6b)[0x5596a8811d6b]
[fd2:20902] [ 9] SU2_CFD(+0x11f12c)[0x5596a86e412c]
[fd2:20902] [10] SU2_CFD(+0x12f25c)[0x5596a86f425c]
[fd2:20902] [11] SU2_CFD(+0x134b1f)[0x5596a86f9b1f]
[fd2:20902] [12] SU2_CFD(+0xe0066)[0x5596a86a5066]
[fd2:20902] [13] [fd2:21150] [ 0] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7f3abacbebf7]
[fd2:20902] [14] SU2_CFD(+0xf0aca)[0x/lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7fd5b7f9f980]
5596a86b5aca]
[fd2:20902] *** End of error message ***
[fd2:21150] [ 1] SU2_CFD(+0x48921b)[0x55de1d66f21b]
[fd2:21150] [ 2] SU2_CFD(+0x3b15df)[0x55de1d5975df]
[fd2:21150] [ 3] SU2_CFD(+0x35a0c3)[0x55de1d5400c3]
[fd2:21150] [ 4] SU2_CFD(+0x3aafa1)[0x55de1d590fa1]
[fd2:21150] [ 5] SU2_CFD(+0x398683)[0x55de1d57e683]
[fd2:21150] [ 6] SU2_CFD(+0x24b4af)[0x55de1d4314af]
[fd2:21150] [ 7] SU2_CFD(+0x24bfd4)[0x55de1d431fd4]
[fd2:21150] [ 8] SU2_CFD(+0x24cd6b)[0x55de1d432d6b]
[fd2:21150] [ 9] SU2_CFD(+0x11f12c)[0x55de1d30512c]
[fd2:21150] [10] SU2_CFD(+0x12f25c)[0x55de1d31525c]
[fd2:21150] [11] SU2_CFD(+0x134b1f)[0x55de1d31ab1f]
[fd2:21150] [12] SU2_CFD(+0xe0066)[0x55de1d2c6066]
[fd2:21150] [13] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7)[0x7fd5b7bbdbf7]
[fd2:21150] [14] SU2_CFD(+0xf0aca)[0x55de1d2d6aca]
[fd2:21150] *** End of error message ***
--------------------------------------------------------------------------
mpirun noticed that process rank 4 with PID 0 on node fd2 exited on signal 11 (Segmentation fault).

[CSMDakota]

Originally Posted by pcg

Hi Brandon,
Do you have low Mach correction options in the config when you try to use JST?
That feature is not compatible with centered schemes (or rather it does not affect them), and at the moment not compatible with vectorized upwind schemes, but all centered schemes are vectorized now... And so I took the blue pill and threw an error if low Mach + vectorization.
It should work if you disable that option, low Mach preconditioning is compatible with any scheme.

Regards,
Pedro

Thanks, Pedro,


I do observe the same behavior whether I have low Mach correction on or off.


Could it be related to the AVX setup? Was this a valid way to configure the compilation for vectorization?

> export CXXFLAGS="-O2 -funroll-loops -ffast-math -march=ivybridge -mtune=ivybridge"



> ./meson.py build --reconfigure -Dwith-omp=true

Thanks for all your hard development work on SU2!
--Brandon--

[flavio73]

I made some progress. By comparing the new configuration file with an old one (the oneraM6 one), I found that the following new option was active:
NUM_METHOD_GRAD_RECON=LEAST_SQUARES.
Just commenting it solved the problem. I am still testing it. I will let you know
Thanks
Flavio

[CSMDakota]

Originally Posted by flavio73

I made some progress. By comparing the new configuration file with an old one (the oneraM6 one), I found that the following new option was active:
NUM_METHOD_GRAD_RECON=LEAST_SQUARES.
Just commenting it solved the problem. I am still testing it. I will let you know
Thanks
Flavio

Nice troubleshooting, Flavio! For my case, I went from GREEN_GAUSS to LEAST_SQUARES to "commented out", but neither fixed the problem for me.


Thanks,
Brandon

[pcg]

Hi Flavio, the NUM_METHOD_GRAD_RECON option also does not affect centered schemes (other than crashing them apparently) I will work on fix to avoid that.

Hi Brandon, can you share the config file? Your compilation options should be valid.
(avx errors, if there are any, happen as soon as you launch SU2, the OS detects illegal instructions for the architecture).

[CSMDakota]

Thanks, Pedro,


My *.cfg is attached here (appended with ".txt" to make the attachment uploader happy).



--Brandon--

[pcg]

So I was able to replicate the problem with your config, modified to use the mesh from the quickstart case.
But there is no crash if I comment out the line with NUM_METHOD_GRAD_RECON.
And there are also no memory errors.

[CSMDakota]

Pedro,


OK it also worked for me when I commented NUM_METHOD_GRAD_RECON out. Not sure why it didn't work when I tried it before, but it worked OK just now.


Thanks,
Brandon

[flavio73]

Hi guys,
just an update. I still have random failures which I cannot understand.
They could be related to the use of the compilation option -Dwith-omp=true.
I will try to recompile it with AVX support but without omp.

Here is the message that I got randomly.
Anyone experienced a similar behaviour ?
Thanks
Flavio


Traceback (most recent call last):
File "/home/flavio/bin/parallel_computation.py", line 110, in <module>
main()
File "/home/flavio/bin/parallel_computation.py", line 57, in main
options.compute )
File "/home/flavio/bin/parallel_computation.py", line 88, in parallel_computation
info = SU2.run.CFD(config)
File "/home/flavio/bin/SU2/run/interface.py", line 112, in CFD
run_command( the_Command )
File "/home/flavio/bin/SU2/run/interface.py", line 270, in run_command
raise exception(message)
SU2.EvaluationFailure: Path = /home/flavio/JST_MAT_ADAPT/,
Command = mpirun -n 10 /home/flavio/bin/SU2_CFD config_CFD.cfg
SU2 process returned error '1'

[pcg]

If you run SU2 directly without the parallel_computation script you might get more info.
Note that if you compiled SU2 with MPI+OpenMP you need to be a bit more careful launching the code.
The total number of cores that will be used is "number of mpi ranks" (10 in your case) times the "number of threads per rank", this is specified in one of two ways:
- Environment variable OMP_NUM_THREADS=?
- SU2_CFD -t ? ...
In most systems the default OMP_NUM_THREADS is the number of cores.
In your case if you have 10 cores and launch 10 mpi ranks you might end up with 100 threads.
Oversubscribing the system like that often results in errors.

[flavio73]

Thanks Pedro
You are right ! I will give it a try to see if this was the problem
Thanks a lot for the hint
Flavio