-
Configuration File
Mesh File
Restart File
Solver Setup
Physical Definition
Markers and Boundary Conditions
Convective Schemes
Slope Limiters and Shock Resolution
History and Solution Output
Linear Solvers and Preconditioners
Basics of Multi-Zone Computations
Execution
Aeroacoustic analysis of unsteady pressure
Flamelet Solver — Python Wrapper Wall Boundary Conditions
Output Fields Reference
Error Messages
-
Setup Guide: Steady RANS for External Aerodynamics
Setup Guide: Incompressible Flows
Setup Guide: Unsteady Flows and DES
Setup Guide: Laminar-Turbulent Transition
Setup Guide: Turbomachinery
Setup Guide: Heat Transfer and Conjugate Heat Transfer
Setup Guide: Species Transport and Combustion
Setup Guide: Shape Optimization with the Discrete Adjoint
Setup Guide: Non-Ideal Compressible Flows
-
Option Reference
Problem Definition
FluidModel
Data-driven fluid model parameters
Freestream Conditions
Reference Conditions
Boundary Markers
Time-marching
Linear solver definition
Convergence
Multi-grid
Spatial Discretization
Adjoint and Gradient
Sobolev Gradient Solver Parameters
Input/output files and formats
Dynamic mesh definition
Aeroelastic Simulation (Typical Section Model)
Wind Gust
Equivalent Area
Grid deformation
Deformable mesh
FEM flow solver definition
FEA solver
Multizone definition
Radiation solver
Heat solver
Visualize Control Volumes
Inverse design problem
Unsupported options
FFD point inversion
Automatic Differentation options
Python Options
Hybrid_RANSLES Options
Output Options
libROM options
Renamed and Removed Options
-
GIT-ting Started
Developing SU2 on GitHub (Internal Developers)
Running Regression Tests
Code Review
Code Structure
Style Guide
Advanced AD Techniques
Containers for Development
Writing Unit Tests
Integrating Tracy Profiler with SU2
Automatic Differentiation (AD)
Code Reference (Doxygen)
Debugging SU2
Documentation Improvements and Status
Glossary
Docs Check
Multi-grid
Generated from SU2 commit
6b54b78464on 2026-10-23. Do not edit this page. Edit the comment above the option inCommon/src/CConfig.cppinstead.
MGLEVEL
Multi-grid Levels. DEFAULT: 0
More in config_template.cfg: Max. number of multi-grid levels (0 = no multi-grid)
- Type: integer
- Default:
0 - In config_template.cfg:
MGLEVEL= 0 - Mentioned in: Changes in Version 8, Setup Guide: Incompressible Flows, Setup Guide: Non-Ideal Compressible Flows, Setup Guide: Shape Optimization with the Discrete Adjoint, Setup Guide: Species Transport and Combustion, Setup Guide: Steady RANS for External Aerodynamics and 4 more
- Used in 119 test case files, for example:
aeroelastic_NACA64A010.cfg(3),air_nozzle_species.cfg(1),inv_NACA0012.cfg(2) - Code:
CConfig.cppline 2082
MGCYCLE
Multi-grid cycle. OPTIONS: See MG_Cycle_Map. Defualt V_CYCLE
More in config_template.cfg: Multi-grid cycle (V_CYCLE, W_CYCLE, FULLMG_CYCLE)
- Type: one of the values below
- Values:
V_CYCLE,W_CYCLE,FULLMG_CYCLE - Default:
V_CYCLE - In config_template.cfg:
MGCYCLE= V_CYCLE - Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Setup Guide: Steady RANS for External Aerodynamics, Troubleshooting, Inviscid ONERA M6
- Used in 98 test case files, for example:
aeroelastic_NACA64A010.cfg(W_CYCLE),inv_wedge_ROE.cfg(V_CYCLE),inv_NACA0012.cfg(FULLMG_CYCLE) - Code:
CConfig.cppline 2084
MG_PRE_SMOOTH
Multi-grid pre-smoothing level
More in config_template.cfg: Max. iterations for multi-grid pre-smoothing level
- Type: list of integers
- In config_template.cfg:
MG_PRE_SMOOTH= ( 2, 2, 2, 2 ) - Mentioned in: Changes in Version 8, Setup Guide: Shape Optimization with the Discrete Adjoint, Inviscid ONERA M6
- Used in 94 test case files, for example:
aeroelastic_NACA64A010.cfg(( 1, 2, 3, 3 )),inv_NACA0012.cfg(( 2, 2, 2, 2 )),inv_NACA0012_FD.cfg(( 1, 2, 2, 2 )) - Code:
CConfig.cppline 2086
MG_POST_SMOOTH
Multi-grid post-smoothing level
More in config_template.cfg: Max. iterations for multi-grid post-smoothing level
- Type: list of integers
- In config_template.cfg:
MG_POST_SMOOTH= ( 2, 2, 2, 2 ) - Mentioned in: Changes in Version 8, Setup Guide: Shape Optimization with the Discrete Adjoint, Inviscid ONERA M6
- Used in 94 test case files, for example:
aeroelastic_NACA64A010.cfg(( 0, 0, 0, 0 )),inv_NACA0012.cfg(( 2, 2, 2, 2 )),inv_NACA0012_FD.cfg(( 1, 1, 1, 1 )) - Code:
CConfig.cppline 2088
MG_CORRECTION_SMOOTH
Jacobi implicit smoothing of the correction
- Type: list of integers
- In config_template.cfg:
MG_CORRECTION_SMOOTH= ( 1, 1, 1, 1 ) - Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Inviscid ONERA M6
- Used in 94 test case files, for example:
aeroelastic_NACA64A010.cfg(( 0, 0, 0, 0 )),inv_NACA0012.cfg(( 1, 1, 1, 1 )) - Code:
CConfig.cppline 2090
MG_DAMP_RESTRICTION
Damping factor for the residual restriction. DEFAULT: 0.75
- Type: real number
- Default:
0.5 - In config_template.cfg:
MG_DAMP_RESTRICTION= 0.5 - Mentioned in: Changes in Version 8, Setup Guide: Shape Optimization with the Discrete Adjoint, Setup Guide: Steady RANS for External Aerodynamics, Inviscid ONERA M6
- Used in 94 test case files, for example:
aeroelastic_NACA64A010.cfg(0.75),inv_NACA0012.cfg(0.5),inv_NACA0012_FD.cfg(1.0) - Code:
CConfig.cppline 2092
MG_DAMP_PROLONGATION
Damping factor for the correction prolongation. DEFAULT 0.75
- Type: real number
- Default:
0.5 - In config_template.cfg:
MG_DAMP_PROLONGATION= 0.5 - Mentioned in: Changes in Version 8, Setup Guide: Shape Optimization with the Discrete Adjoint, Setup Guide: Steady RANS for External Aerodynamics, Inviscid ONERA M6
- Used in 94 test case files, for example:
aeroelastic_NACA64A010.cfg(0.75),inv_NACA0012.cfg(0.5),inv_NACA0012_FD.cfg(1.0) - Code:
CConfig.cppline 2094
MG_SMOOTH_EARLY_EXIT
Enable early exit for MG smoothing when RMS drops below threshold. DEFAULT: NO
More in config_template.cfg: Enable early exit from multigrid smoothing when the residual RMS drops below MG_SMOOTH_RES_THRESHOLD * initial_rms (NO, YES)
- Type: YES / NO
- Default:
YES - In config_template.cfg:
MG_SMOOTH_EARLY_EXIT= YES - Mentioned in: Changes in Version 8
- Used in 3 test case files, for example:
inv_CRM_JST.cfg(NO) - Code:
CConfig.cppline 2096
MG_SMOOTH_RES_THRESHOLD
Early exit smoothing when current_rms drops below threshold * initial_rms. DEFAULT: 0.9
More in config_template.cfg: Relative RMS reduction threshold for the multigrid smoothing early exit. Smoothing stops when current_rms < threshold * initial_rms (default 0.5)
- Type: real number
- Default:
0.9 - In config_template.cfg:
MG_SMOOTH_RES_THRESHOLD= 0.9 - Mentioned in: Changes in Version 8
- Code:
CConfig.cppline 2098
MG_SMOOTH_OUTPUT
Print compact per-cycle smoothing iteration summary. DEFAULT: NO
More in config_template.cfg: Print a compact per-cycle summary of smoothing iterations and residuals (NO, YES)
- Type: YES / NO
- Default:
NO - In config_template.cfg:
MG_SMOOTH_OUTPUT= NO - Used in 2 test case files, for example:
inv_CRM_JST.cfg(NO) - Code:
CConfig.cppline 2100
MG_SMOOTH_STAGNATION_TOL
Stop smoothing if current_rms >= previous_rms this value. Values < 1.0 enable early exit on stagnation, 1.0 only exits on defect growth. DEFAULT: 0.99
More in config_template.cfg: Stop smoothing when current_rms >= previous_rms * this value. Values below 1.0 exit early on stagnation, 1.0 exits only when the defect grows, 0 disables the check (default 0.99)
- Type: real number
- Default:
0.99 - In config_template.cfg:
MG_SMOOTH_STAGNATION_TOL= 0.99 - Code:
CConfig.cppline 2103
MG_SMOOTH_COEFF
Smoothing coefficient for the correction prolongation Jacobi smoother. DEFAULT: 1.25
More in config_template.cfg: Smoothing coefficient for the Jacobi correction-prolongation smoother (default 1.25)
- Type: real number
- Default:
1.25 - In config_template.cfg:
MG_SMOOTH_COEFF= 1.25 - Mentioned in: Changes in Version 8
- Code:
CConfig.cppline 2105
MG_MIN_MESHSIZE
Minimum number of CVs on the coarsest multigrid level, counted over the whole level rather than per rank, so the hierarchy does not change with the number of ranks. DEFAULT: 500
More in config_template.cfg: Minimum number of control volumes on the coarsest multigrid level, counted over the whole level rather than per rank, so the hierarchy does not change with the number of MPI ranks. A level holding fewer control volumes than this is not created.
- Type: integer
- Default:
500 - In config_template.cfg:
MG_MIN_MESHSIZE= 500 - Mentioned in: Changes in Version 8
- Used in 3 test case files, for example:
inv_CRM_JST.cfg(100),inv_nozzle.cfg(500),multigrid.cfg(20) - Code:
CConfig.cppline 2108
MG_IMPLICIT_LINES
Pave the coarse grid with advancing fronts raised from boundaries that carry a stretched layer normal to themselves. DEFAULT: NO
More in config_template.cfg: Pave the coarse grid with advancing fronts raised from boundaries that carry a stretched layer normal to themselves (NO, YES)
- Type: YES / NO
- Default:
NO - In config_template.cfg:
MG_IMPLICIT_LINES= NO - Used in 17 test case files, for example:
inv_CRM_JST.cfg(YES) - Code:
CConfig.cppline 2111
MG_STARTUP_ITER
Max number of iterations spent on each mesh during the Full Multigrid (FMG) startup phase. DEFAULT: 100
More in config_template.cfg: Number of iterations spent on each mesh during the Full Multigrid (FMG) startup phase. After this many iterations the solution is prolongated to the next finer mesh. It is also the length of the CFL ramp every level is brought up over, the finest grid included, which is what keeps the CFL from jumping when the startup hands the solution over. 0 removes the iteration budget, leaving MG_STARTUP_CONVERGENCE and MG_STARTUP_STAGNATION to decide, and disables the ramp (default 100).
- Type: integer
- Default:
100 - In config_template.cfg:
MG_STARTUP_ITER= 100 - Used in 1 test case file, for example:
inv_NACA0012.cfg(10) - Code:
CConfig.cppline 2114
MG_STARTUP_CONVERGENCE
During the startup phase of Full-MG, leave the current level once CONV_FIELD has dropped by this many orders of magnitude relative to its value when the level became active. DEFAULT: -2
More in config_template.cfg: Full-MG promotion on convergence: orders of magnitude (log10) that CONV_FIELD must drop on the active level, relative to its value when that level became active, before promoting to the next finer level early. Negative for a drop, on the same scale as CONV_RESIDUAL_MINVAL, so -2 asks for a fall to 1% of the starting residual. Only residual convergence fields are used. Make it more negative to demand a deeper drop; a large negative value effectively disables this criterion (default -2).
- Type: real number
- Default:
-2.0 - In config_template.cfg:
MG_STARTUP_CONVERGENCE= -2 - Used in 1 test case file, for example:
inv_NACA0012.cfg(-2) - Code:
CConfig.cppline 2118
MG_STARTUP_STAGNATION
Full-MG promotion on stagnation. If the active level’s residual ratio between successive iterations exceeds this value for MG_STARTUP_STAGNATION_ITER consecutive iterations, promote to the next finer level without waiting out MG_STARTUP_ITER. 0 disables it. DEFAULT: 0.99
More in config_template.cfg: Full-MG promotion on stagnation: if the active level’s residual ratio between successive iterations exceeds this for MG_STARTUP_STAGNATION_ITER consecutive iterations, promote to the next finer level early. 0 disables (default 0.99).
- Type: real number
- Default:
0.99 - In config_template.cfg:
MG_STARTUP_STAGNATION= 0.99 - Used in 1 test case file, for example:
inv_NACA0012.cfg(0.99) - Code:
CConfig.cppline 2122
MG_STARTUP_STAGNATION_ITER
Consecutive stalled iterations required before Full-MG promotes on stagnation. 0 disables it, as MG_STARTUP_STAGNATION= 0 does. DEFAULT: 5
More in config_template.cfg: Consecutive stalled iterations required before Full-MG promotes on stagnation. 0 disables the criterion, as MG_STARTUP_STAGNATION= 0 does (default 5).
- Type: integer
- Default:
5 - In config_template.cfg:
MG_STARTUP_STAGNATION_ITER= 5 - Code:
CConfig.cppline 2125
MG_CFL_SCALING
Per-level CFL scaling factors for coarse MG levels. Entry i is the ratio CFL(i+1)/CFL(i). If fewer values than nMGLevels are given, the last value is repeated. DEFAULT: 0.25 (i.e., 1/4 per level)
More in config_template.cfg: Per-level CFL scaling factors for the coarse multigrid levels. Entry i is the ratio CFL(i+1)/CFL(i). If fewer values than MGLEVEL are given the last one is repeated (default 0.25, i.e. a quarter of the CFL per level)
- Type: list of real numbers
- In config_template.cfg:
MG_CFL_SCALING= ( 0.25 ) - Code:
CConfig.cppline 2128