Sobolev Gradient Solver Parameters

Generated from SU2 commit 6b54b78464 on 2026-10-23. Do not edit this page. Edit the comment above the option in Common/src/CConfig.cpp instead.

SMOOTH_GRADIENT

Switch to activate gradient smoothing

More in config_template.cfg: Activate the gradient smoothing solver for the discrete adjoint driver (NO, YES) see TestCases/grad_smooth/naca0012/inv_NACA0012_gradsmooth.cfg for a detailed explanantion on the config options shown here.

• Type: YES / NO
• Default: NO
• In config_template.cfg: SMOOTH_GRADIENT= NO
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (YES)
• Code: CConfig.cpp line 2314

SMOOTHING_EPSILON1

Epsilon of the identity term in the Laplace Beltrami Operator

More in config_template.cfg: Scaling factor for the identity part of the Laplace-Beltrami operator

• Type: real number
• Default: 1.0
• In config_template.cfg: SMOOTHING_EPSILON1= 1.0
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (1.0), ONERAM6_gradsmooth.cfg (51.8116714655595)
• Code: CConfig.cpp line 2316

SMOOTHING_EPSILON2

Epsilon of the Laplace term in the Laplace Beltrami Operator

More in config_template.cfg: Scaling factor for the Laplace part of the Laplace-Beltrami operator

• Type: real number
• Default: 1.0
• In config_template.cfg: SMOOTHING_EPSILON2= 1.0
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (0.625), ONERAM6_gradsmooth.cfg (0.700463534712351)
• Code: CConfig.cpp line 2318

SEPARATE_DIMENSIONS

Switch to calculate for each dimension separately

More in config_template.cfg: Switch for running the smoothing procedure seperately in each space dimension (NO, YES).

• Type: YES / NO
• Default: NO
• In config_template.cfg: SEPARATE_DIMENSIONS= NO
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (NO)
• Code: CConfig.cpp line 2320

SMOOTH_ON_SURFACE

Switch to activate working on the design surfaces only

More in config_template.cfg: Switch for applying the Sobolev smoothing to the design surface mesh, i.e., instead of the whole volume mesh (NO, YES).

• Type: YES / NO
• Default: NO
• In config_template.cfg: SMOOTH_ON_SURFACE= NO
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (YES)
• Code: CConfig.cpp line 2322

DIRICHLET_SURFACE_BOUNDARY

Switch to activate zero Dirichlet boundary for surface mode

More in config_template.cfg: Use Dirichlet boundary conditions when working on the design surface, only used for SMOOTH_ON_SURFACE= YES (NO, YES).

• Type: YES / NO
• Default: NO
• In config_template.cfg: DIRICHLET_SURFACE_BOUNDARY= NO
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (NO)
• Code: CConfig.cpp line 2324

SOBOLEV_MODE

Switch to activate the debbuging modus

More in config_template.cfg: Mode how the Sobolev method is applied to the discrete adjoint gradient.

• Type: one of the values below
• Values: NONE, PARAM_LEVEL_COMPLETE, MESH_LEVEL, ONLY_GRADIENT
• Default: NONE
• In config_template.cfg: SOBOLEV_MODE= NO_MODUS
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (PARAM_LEVEL_COMPLETE)
• Code: CConfig.cpp line 2326

HESS_OBJFUNC_FILENAME

• Output filename for the Sobolev Hessian approximation.

More in config_template.cfg: Output filename for the assembled Sobolev smoothing system matrix

• Type: string
• Default: string("of_hess.dat")
• In config_template.cfg: HESS_OBJFUNC_FILENAME= of_hess.dat
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (of_hess.dat)
• Code: CConfig.cpp line 2329

GRAD_LINEAR_SOLVER

Linear solver for the gradient smoothing

More in config_template.cfg: Linear solver or smoother for implicit formulations (same options as LINEAR_SOLVER)

• Type: one of the values below
• Values: CONJUGATE_GRADIENT, BCGSTAB, FGMRES, RESTARTED_FGMRES, FGCRODR, SMOOTHER, PASTIX_LDLT, PASTIX_LU
• Default: FGMRES
• In config_template.cfg: GRAD_LINEAR_SOLVER= FGMRES
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (FGMRES)
• Code: CConfig.cpp line 2332

GRAD_LINEAR_SOLVER_PREC

Preconditioner for the Krylov linear solvers

More in config_template.cfg: Preconditioner of the Krylov linear solver (ILU, LU_SGS, JACOBI)

• Type: one of the values below
• Values: NONE, IDENTITY, JACOBI, LU_SGS, LINELET, ILU, Q_LU_SGS, Q_JACOBI, Q_IDENTITY, PASTIX_ILU, PASTIX_LU, PASTIX_LDLT
• Default: ILU
• In config_template.cfg: GRAD_LINEAR_SOLVER_PREC= ILU
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (ILU)
• Code: CConfig.cpp line 2334

GRAD_LINEAR_SOLVER_ERROR

Minimum error threshold for the linear solver for the implicit formulation

More in config_template.cfg: Minimum residual criteria for the linear solver convergence of the Sobolev smoothing solver

• Type: real number
• Default: 1E-14
• In config_template.cfg: GRAD_LINEAR_SOLVER_ERROR= 1E-14
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (1E-14)
• Code: CConfig.cpp line 2336

GRAD_LINEAR_SOLVER_ITER

Maximum number of iterations of the linear solver for the implicit formulation

More in config_template.cfg: Number of linear solver iterations for the Sobolev smoothing solver

• Type: integer
• Default: 1000
• In config_template.cfg: GRAD_LINEAR_SOLVER_ITER= 1000
• Used in 2 test case files, for example: inv_NACA0012_gradsmooth.cfg (1000)
• Code: CConfig.cpp line 2338