With OGS_FEM, OGS_FEM_SP:
1 Gauss
2 BiCGStab (OGS_FEM_MPI)
3 BiCG (OGS_FEM_MPI)
4 QMRGG Stab
5 CG (OGS_FEM_MPI)
6 CGNR
7 CGS (OGS_FEM_MPI)
8 Richardson
9 JOR
10 SOR
11 AMG1R5
12 UMF
13 GMRes

With OGS_FEM_MKL 805 Pardiso

0 No
1 Jacobi (for OGS_FEM_MPI)
100 ILU

1. vollbesetzte Matrix
   ( param1 = Dimension )
   (2) nur A[i,j]!=0 werden gespeichert (Sparse)
   ( param1 = Dimension )
   (3) symmetrische sparse Matrix
   fuer Preconditioner "incomplete Cholesky"
   nur Aik !=0 mit k>=i werden gespeichert
   ( param1 = Dimension )
   (4) unsymmetrische sparse Matrix
   fuer Preconditioner "incomplete LDU-Zerlegung"
   nur Aik !=0 werden gespeichert
   ( param1 = Dimension )
   

Example *.num file:

$OVERALL_COUPLING
2 10    ; minNumberOfIterations maxNumberOfIterations 
#NUMERICS
$PCS_TYPE
LIQUID_FLOW
$ELE_MASS_LUMPING
1
$LINEAR_SOLVER
; method error_tolerance max_iterations theta precond storage
2 6 1.e-014 5000 1.0 1 2
$NON_LINEAR_ITERATIONS
;type -- error_method -- max_iterations -- relaxation -- tolerance(s)
PICARD LMAX 2 0.0 1.
$ELE_GAUSS_POINTS
2
$COUPLING_CONTROL
LMAX 1.  ; Error tolerance in coupling loop

#NUMERICS
$PCS_TYPE
MASS_TRANSPORT
$ELE_GAUSS_POINTS
2
$LINEAR_SOLVER
2 6 1.e-014 5000 1.0 1 2
$NON_LINEAR_ITERATIONS
;type -- error_method -- max_iterations -- relaxation -- tolerance(s)
PICARD LMAX .1 0.0 1.0e-7
$COUPLING_CONTROL
LMAX 1e-1  ; Error tolerance in coupling loop

#STOP

Solver bcgs, gmres, etc.
Preconditioner jacobi, bjacobi, sor, asm, mg

$LINEAR_SOLVER
petsc solver preconditioner errorTolerance maxInterations Theta

0 r<e 
1 r/b<e
2 r_n1/r_n<e
3 if r_n1>1 then r_n1/r_n < e else r<e
4 r/x<e
5 r_n1/max(x,b,r_n)<e 
6 ??? 

In *num:

$FEM_FCT
method prelimiterType const_alpha

method 1: linearized FCT
prelimiterType 0: just cancel, 1: minmod, 2: superbee
fct_const_apha -1: off (is default), [0.0, 1.0] 0: upwind, 1: galerkin

Example *tim file:

#TIME_STEPPING
$PCS_TYPE
LIQUID_FLOW
$TIME_END
1e10
$TIME_CONTROL
SELF_ADAPTIVE
4 1.3 ; multiply time step sitze by 1.3 if number of iterations <= 4 (if condition satisfied, this is selected, the following ignored)
7 1.0 ; multiply time step sitze by 1. if number of iterations <= 7
10 0.7 ; else multiply time step sitze by 0.7 ( the 10 in last line does nothing, can give more than three lines)
MAX_TIME_STEP
1e10
MIN_TIME_STEP
1e3
INITIAL_STEP_SIZE
1e3
ITERATIVE_TYPE
COUPLED ; for outer coupling loop (take keyword LINEAR for solver, NONLINEAR for Picard)
#STOP

You can store primary variable values with the DAT_TYPE PRIMARY_VARIABLES and reload them as initial conditions with the DIS_TYPE DIRECT into your next simulation

Example:

• Generate output file with pressure from LIQUID_FLOW at time 1000:

#OUTPUT
$PCS_TYPE
LIQUID_FLOW
$NOD_VALUES
PRESSURE1
$GEO_TYPE
DOMAIN
$DAT_TYPE
PRIMARY_VARIABLES
$TIM_TYPE
1000

• Read the file:

#INITIAL_CONDITION
$PCS_TYPE
LIQUID_FLOW
$PRIMARY_VARIABLE
PRESSURE1
$DIS_TYPE
DIRECT LIQUID_FLOW_domain_primary_variables.txt

See wiki page
Example:

#MEDIUM PROPERTIES
…
$ELEMENT_VOLUME_MULTIPLYER
1 1 8 ; 2D case x, y (z) - first number is mode, then 1 parameter if 1D, 2 parameter if 2D, 3 parameter if 3D

modes
0: increase storage only
1: increase storage and adapt Laplace, advection, velocity

$DIS_TYPE
GRADIENT ref_depth ref_depth_value ref_depth_gradient

Darcy flux for LIQUID_FLOW, GROUNDWATER_FLOW, RICHARDS_FLOW
at nodes: VELOCITY_X1 VELOCITY_Y1 VELOCITY_Z1
at gauss point nr 0: VELOCITY1_X VELOCITY1_Y VELOCITY1_Z

To get fick flux, add

$PCS_TYPE
  MASS_TRANSPORT

and to get fourier flux, add

$PCS_TYPE
  HEAT_TRANSPORT

to output instance.

In *ic: set flag with keyword

$STORE_VALUES 

in *.out: Use prefix DELTA_, e.g.

$DAT_TYPE
DELTA_PRESSURE1
DELTA_TEMPERATURE1 
DELTA_CONCENTRATION1

Implemented for surfaces, e.g. to get an average surface temperature

$NOD_VALUES
TEMPERATURE1
$GEO_TYPE
 SURFACE sfc_1
$DIS_TYPE
 AVERAGE

See wiki page

Example:

#OUTPUT
$PCS_TYPE
LIQUID_FLOW
$DAT_TYPE
CONTENT -1 ; or mmp index
$TIM_TYPE
STEPS 1
#OUTPUT
$PCS_TYPE
LIQUID_FLOW
$DAT_TYPE
TOTAL_FLUX
$TIM_TYPE
STEPS 1

Copy GP velocities from one element to a number of other elements:

#PROCESS
 $PCS_TYPE
  LIQUID_FLOW
 $NUM_TYPE
  NEW
 $COPY_GP_VELOCITIES  ; Give total number of indices, then index to copy from, the indices to copy to.
   2  44  46

See [wiki page] (https://github.com/drjod/ogs_kb1/wiki/NNNC)

Example:

#SOURCE_TERM
$PCS_TYPE
HEAT_TRANSPORT
$PRIMARY_VARIABLE
TEMPERATURE1
$GEO_TYPE
SURFACE SURF_BOTTOM_LEFT
$DIS_TYPE
CONSTANT_NEUMANN 1.0
$CONNECTED_GEOMETRY
SURFACE SURF_BOTTOM_RIGHT
$CONNECT_PARAMETERS
1e6 1 ; exchange coefficient verbosity level = 0, 1
$CONNECT_MODE
2 226 0 0 1 1.e-10
; mode ; 0: symmetric, 1: non-symmetric (downwind fixed), 2 variable (dependent on velocity in reference element)
; 2 ref_element_number n_ref_x, n_ref_y, n_ref_z minimum_velocity_abs

<<<<<<< HEAD

Flux for RICHARDS_FLOW to couple to OVERLAND_FLOW:

#SOURCE_TERM
 $PCS_TYPE
  RICHARDS_FLOW
 $PRIMARY_VARIABLE
  PRESSURE1
 $GEO_TYPE
  POLYLINE INTERFACE
 $DIS_TYPE_CONDITION
  CONSTANT_NEUMANN 1
  PCS OVERLAND_FLOW
  HEAD
  2.83333333e-2 1.e-3 0.0 0.0 ; leakage, rill height, given value, residual permeability 

Couple HEAT_TRANSPORT to a given temperature of 10:

#SOURCE_TERM
 $PCS_TYPE
  HEAT_TRANSPORT
 $PRIMARY_VARIABLE
  TEMPERATURE1
 $GEO_TYPE
  POLYLINE INTERFACE
 $DIS_TYPE_CONDITION
  CONSTANT_NEUMANN 1
  HEAT_TRANSPORT
  GIVEN_TEMPERATURE
  0.1 0.0 10.0 0.0 ; leakage, rill height, given value, residual permeability 

=======

develop

#PROCESS
 $PCS_TYPE
  LIQUID_FLOW
 $DEACTIVATED_SUBDOMAIN
  2  ; number of deactivated subdomains
  1 2   ; patch indices (mmp groups)

Different fluid density models for different patch indices (mmp groups). Supported are density models 1-8 and 14, e.g.:

#FLUID_PROPERTIES
 $FLUID_TYPE
  LIQUID
 $MATERIAL_DENSITY
  0 1 1000    ; density model 1 for patch index 0
  1 8 0       ; density model 8 for patch index 1

Different fluid properties for different patch indices (mmp groups).

Example:

Generate two liquids

#FLUID_PROPERTIES
 $FLUID_TYPE
   LIQUID_WELL
   
#FLUID_PROPERTIES
 $FLUID_TYPE
   LIQUID_AQ  

and select liquids as:

#MEDIUM_PROPERTIES
 $DEPENDENT_FLUID
   WELL
   
#MEDIUM_PROPERTIES
 $DEPENDENT_FLUID
   AQ

0: Curve rho(x)
1: rho = const
2: rho(p) = rho_0*(1+beta_p*(p-p_0))
3: rho(C) = rho_0*(1+beta_C*(C-C_0))
4: rho(T) = rho_0*(1+beta_T*(T-T_0))
5: rho(C,T) = rho_0*(1+beta_C*(C-C_0)+beta_T*(T-T_0))
6: rho(p,T) = rho_0*(1+beta_p*(p-p_0)+beta_T*(T-T_0))
7: Pefect gas
8: Density output AB-model
10: density from fct-file with temperature-pressure values
11: Redlich-Kwong EOS for different fluids
12: Peng-Robinson EOS for different fluids
13: Helmholtz free Energy
14: Exponential law
15: Amagat's law for mixture
18: Density at nodes from the phase transition model
19: Density from GEMS
20: rho(p,T, C) for water, range p < 100 MPa, 0 <= T <= 350 °C
21: rho(p,C,T) = rho_0*(1+beta_p*(p-p_0)+beta_C*(C-C_0)+beta_T*(T-T_0))
23: Density depends on concentration of salt and dissolved CO2 26: Dalton's law + ideal gas for use with TNEQ/TES

Example specifications in input file *mfp:

1. Model 0 - rho = rho(x):
   

$DENSITY 
 0 1 

Parameter "1" refers to the first data table in a *.rfd ascii input file, which you will have to provide in this case:

;Curve 1  Temp Density
#CURVES
 273.15  999.8675792
 274.15  999.9265054
 ...
#STOP

More information:
CFluidProperties::Read(std::ifstream*)
CFluidProperties::Density(double*) [separate calculation for solving PDEs and density output]

0: Curve my(x)
1: my = const
2: my(p) = my_0*(1+gamma_p*(p-p_0))
3: my^l(T) acoording toYaws et al. (1976)
4: my^g(T) acoording to Marsily (1986)
5: my^g(p,T) acoording toReichenberg (1971)
6: my(C,T)
8: my(p,C,T)
9: my(rho,T)
15: VTPR-EoS
18: ViscositY at nodes from the phase transition model
19: Viscosity for GEMS
26: Wilke (see Poling, B. E.; Prausnitz, J. M.; John Paul, O. & Reid, R. C. The properties of gases and liquids McGraw-Hill New York, 2001, 5: page 9.21)

More information:
CFluidProperties::Read(std::ifstream*)
CFluidProperties::Viscosity(double* variables)

qstat
qstat –u $LOGIN
qstat –q angus
qdel –f $JOBID
mpi: #PBS -l select=1:ncpus=2:mpiprocs=2:mem=1gb:place=scatter
omp: #PBS -l select=1:ncpus=1:ompthreads=4:mem=1gb:place=group=host

An example bash script for an MPI-Job with 8 cores on the angus queue (rz cluster):

#!/bin/bash
#PBS -o screenout.txt
#PBS -j oe
#PBS -r n
#PBS -l walltime=2:00:00
#PBS -l select=1:ncpus=8:mem=3gb
#PBS -l place=scatter
#PBS -q angus
#PBS -N test

cd $PBS_O_WORKDIR

. /usr/share/Modules/init/bash

. /cluster/Software/intel1502/composer_xe_2015.2.164/bin/compilervars.sh  intel64
. /cluster/Software/intel1502/composer_xe_2015.2.164/mkl/bin/intel64/mklvars_intel64.sh
. /cluster/Software/intel1502/impi/5.0.3.048/intel64/bin/mpivars.sh

time mpirun -r rsh -machinefile $PBS_NODEFILE -n 8 ogs_OGS_FEM_MPI testCase

qstat -f $PBS_JOBID
exit

In this example, name of ogs is 'ogs_OGS_FEM_MPI' and of input files 'testCase'. Put this script with ogs and input files into one folder, cd there and execute the script. Output will be written into screenout.txt.
To change the number of cores, you modify the number (8) in two locations of the script. These are in the PBS script command ncpus=8 and the mpirun command $PBS_NODEFILE -n 8. Important: Each node hosts 16 cores. You can select more than 16 cores by taking more nodes. For instance, you get with the PBS- command

#PBS -l select=3:ncpus=8:mem=64gb

24 cores (set also $PBS_NODEFILE -n 24).
The example script sets maximum wall time as 2 hours, 3G of memory are available for the simulation (each cluster node has 64 (128)GB available in total).

You find the newest script compileInKiel.sh for RZ cluster, NEC Cluster, Lokstedt server Kiel in repository tUNIX. Instructions are in the script.

Script partition.sh for domain decomposition with METIS in repository icbc/remote/easyPeasy (calls partmesh from ufz/mesh_partition)

For OGE_FEM_MPI: ./partition.sh numberOfPartitions –e –asci path (or no path)
For OGS_FEM_PETSC: ./partition.sh numberOfPartitions –n –bin path (or no path)

numberOfPartitions = 2, 3, 4,...