CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > OpenFOAM > OpenFOAM Running, Solving & CFD

AMG versus ICCG

Register Blogs Members List Search Today's Posts Mark Forums Read

Like Tree2Likes
  • 2 Post By hjasak

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   November 7, 2006, 13:46
Default For my approx 1.2 million cell
  #1
Senior Member
 
Srinath Madhavan (a.k.a pUl|)
Join Date: Mar 2009
Location: Edmonton, AB, Canada
Posts: 703
Rep Power: 21
msrinath80 is on a distinguished road
For my approx 1.2 million cells (single-phase) icoFoam vortex shedding case, here is the output after the first three time steps and the corresponding fvSolution pressure solver entries:

ICCG: (p ICCG 1e-06 0;)

/*---------------------------------------------------------------------------*\
| ========= | |
| \ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \ / O peration | Version: 1.3 |
| \ / A nd | Web: http://www.openfoam.org |
| \/ M anipulation | |
\*---------------------------------------------------------------------------*/

Exec : /home/madhavan/OpenFOAM/OpenFOAM-1.3/applications/bin/linuxAMD64Gcc4DPOpt/icoFoa m /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/. one_sq_cyl_3d_unsteady
Date : Nov 07 2006
Time : 12:27:17
Host : masternode
PID : 21882
Root : /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/.
Case : one_sq_cyl_3d_unsteady
Nprocs : 1
Create time

Create mesh for time = 0

Reading transportProperties

Reading field p

Reading field U

Reading/calculating face flux field phi

Reading probeLocations

Constructing probes


Starting time loop

Time = 0.1

Mean and max Courant Numbers = 0 0.0741732
BICCG: Solving for Ux, Initial residual = 1, Final residual = 7.60531e-08, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0, Final residual = 0, No Iterations 0
BICCG: Solving for Uz, Initial residual = 0, Final residual = 0, No Iterations 0
ICCG: Solving for p, Initial residual = 1, Final residual = 9.73006e-07, No Iterations 825
time step continuity errors : sum local = 4.55914e-10, global = -5.75299e-13, cumulative = -5.75299e-13
ICCG: Solving for p, Initial residual = 0.00108292, Final residual = 9.32198e-07, No Iterations 685
time step continuity errors : sum local = 3.75695e-07, global = -7.83026e-10, cumulative = -7.83601e-10

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.01178e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.01178e-05
LiftCoefficient = 1.41566e-11


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 4.79068e-06
pressureDragCoefficient = 4.62468e-06
viscDragCoefficient = 1.66004e-07
LiftCoefficient = 8.15603e-14

ExecutionTime = 347.02 s ClockTime = 347 s

Time = 0.2

Mean and max Courant Numbers = 0.108516 4.04533
BICCG: Solving for Ux, Initial residual = 0.823434, Final residual = 4.45773e-06, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0.329552, Final residual = 3.46137e-06, No Iterations 4
BICCG: Solving for Uz, Initial residual = 0.362595, Final residual = 4.31759e-06, No Iterations 3
ICCG: Solving for p, Initial residual = 0.00371997, Final residual = 9.84174e-07, No Iterations 719
time step continuity errors : sum local = 1.53182e-07, global = 3.88971e-10, cumulative = -3.9463e-10
ICCG: Solving for p, Initial residual = 0.00419475, Final residual = 9.92882e-07, No Iterations 709
time step continuity errors : sum local = 7.36119e-08, global = -2.94289e-10, cumulative = -6.88919e-10

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.11014e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.11014e-05
LiftCoefficient = 2.18023e-11


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = -9.70849e-07
pressureDragCoefficient = -1.11824e-06
viscDragCoefficient = 1.47396e-07
LiftCoefficient = -2.70848e-11

ExecutionTime = 639.75 s ClockTime = 640 s

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

AMG (p AMG 1e-06 0 1000;):

/*---------------------------------------------------------------------------*\
| ========= | |
| \ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \ / O peration | Version: 1.3 |
| \ / A nd | Web: http://www.openfoam.org |
| \/ M anipulation | |
\*---------------------------------------------------------------------------*/

Exec : /home/madhavan/OpenFOAM/OpenFOAM-1.3/applications/bin/linuxAMD64Gcc4DPOpt/icoFoa m /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/. one_sq_cyl_3d_unsteady
Date : Nov 07 2006
Time : 11:29:35
Host : masternode
PID : 21242
Root : /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/.
Case : one_sq_cyl_3d_unsteady
Nprocs : 1
Create time

Create mesh for time = 0

Reading transportProperties

Reading field p

Reading field U

Reading/calculating face flux field phi

Reading probeLocations

Constructing probes


Starting time loop

Time = 0.1

Mean and max Courant Numbers = 0 0.0741732
BICCG: Solving for Ux, Initial residual = 1, Final residual = 7.60531e-08, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0, Final residual = 0, No Iterations 0
BICCG: Solving for Uz, Initial residual = 0, Final residual = 0, No Iterations 0
AMG: Solving for p, Initial residual = 1, Final residual = 9.80164e-07, No Iterations 105
time step continuity errors : sum local = 4.59248e-10, global = 2.47297e-14, cumulative = 2.47297e-14
AMG: Solving for p, Initial residual = 0.00108292, Final residual = 9.69421e-07, No Iterations 48
time step continuity errors : sum local = 3.90696e-07, global = -1.15826e-11, cumulative = -1.15579e-11

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.01179e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.01179e-05
LiftCoefficient = -5.64623e-09


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 4.79158e-06
pressureDragCoefficient = 4.62556e-06
viscDragCoefficient = 1.6602e-07
LiftCoefficient = 1.45852e-10

ExecutionTime = 146.7 s ClockTime = 147 s

Time = 0.2

Mean and max Courant Numbers = 0.108519 4.04592
BICCG: Solving for Ux, Initial residual = 0.823441, Final residual = 4.46106e-06, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0.329408, Final residual = 3.4621e-06, No Iterations 4
BICCG: Solving for Uz, Initial residual = 0.362669, Final residual = 4.339e-06, No Iterations 3
AMG: Solving for p, Initial residual = 0.0037213, Final residual = 9.14087e-07, No Iterations 82
time step continuity errors : sum local = 1.42274e-07, global = -4.78435e-13, cumulative = -1.20363e-11
AMG: Solving for p, Initial residual = 0.00419589, Final residual = 9.94429e-07, No Iterations 75
time step continuity errors : sum local = 7.37282e-08, global = 1.93804e-13, cumulative = -1.18425e-11

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.11013e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.11013e-05
LiftCoefficient = 8.69834e-09


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = -9.72347e-07
pressureDragCoefficient = -1.11973e-06
viscDragCoefficient = 1.47384e-07
LiftCoefficient = -2.83295e-10

ExecutionTime = 294.14 s ClockTime = 295 s

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


AMG (p AMG 1e-06 0 100;):

/*---------------------------------------------------------------------------*\
| ========= | |
| \ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \ / O peration | Version: 1.3 |
| \ / A nd | Web: http://www.openfoam.org |
| \/ M anipulation | |
\*---------------------------------------------------------------------------*/

Exec : /home/madhavan/OpenFOAM/OpenFOAM-1.3/applications/bin/linuxAMD64Gcc4DPOpt/icoFoa m /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/. one_sq_cyl_3d_unsteady
Date : Nov 07 2006
Time : 12:20:05
Host : masternode
PID : 21848
Root : /home/madhavan/OpenFOAM/madhavan-1.3/run/tutorials/icoFoam/.
Case : one_sq_cyl_3d_unsteady
Nprocs : 1
Create time

Create mesh for time = 0

Reading transportProperties

Reading field p

Reading field U

Reading/calculating face flux field phi

Reading probeLocations

Constructing probes


Starting time loop

Time = 0.1

Mean and max Courant Numbers = 0 0.0741732
BICCG: Solving for Ux, Initial residual = 1, Final residual = 7.60531e-08, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0, Final residual = 0, No Iterations 0
BICCG: Solving for Uz, Initial residual = 0, Final residual = 0, No Iterations 0
AMG: Solving for p, Initial residual = 1, Final residual = 8.07098e-07, No Iterations 102
time step continuity errors : sum local = 3.78169e-10, global = 1.46199e-14, cumulative = 1.46199e-14
AMG: Solving for p, Initial residual = 0.00108292, Final residual = 8.70596e-07, No Iterations 48
time step continuity errors : sum local = 3.50868e-07, global = 3.39341e-10, cumulative = 3.39356e-10

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.01179e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.01179e-05
LiftCoefficient = -4.35326e-09


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 4.79156e-06
pressureDragCoefficient = 4.62554e-06
viscDragCoefficient = 1.6602e-07
LiftCoefficient = 8.678e-11

ExecutionTime = 108.75 s ClockTime = 109 s

Time = 0.2

Mean and max Courant Numbers = 0.10852 4.04587
BICCG: Solving for Ux, Initial residual = 0.823445, Final residual = 4.46093e-06, No Iterations 3
BICCG: Solving for Uy, Initial residual = 0.329439, Final residual = 3.4626e-06, No Iterations 4
BICCG: Solving for Uz, Initial residual = 0.362502, Final residual = 4.33196e-06, No Iterations 3
AMG: Solving for p, Initial residual = 0.00372114, Final residual = 7.40435e-07, No Iterations 68
time step continuity errors : sum local = 1.15246e-07, global = -2.12212e-12, cumulative = 3.37234e-10
AMG: Solving for p, Initial residual = 0.00419577, Final residual = 9.12739e-07, No Iterations 60
time step continuity errors : sum local = 6.76724e-08, global = 1.6342e-11, cumulative = 3.53576e-10

Wall patch = 0
Wall patch name = ChannelWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = 2.11013e-05
pressureDragCoefficient = 0
viscDragCoefficient = 2.11013e-05
LiftCoefficient = 7.399e-09


Wall patch = 1
Wall patch name = ObstacleWalls
Uav = (1 0 0)
Aref = 1
nu = nu [0 2 -1 0 0 0 0] 1.004e-06
DragCoefficient = -9.72242e-07
pressureDragCoefficient = -1.11963e-06
viscDragCoefficient = 1.47384e-07
LiftCoefficient = -2.05608e-10

ExecutionTime = 191.77 s ClockTime = 192 s

Clearly, the AMG solver is better than ICCG for this case. Are there any guidelines on what number to use at the coarsest level for the AMG solver?
msrinath80 is offline   Reply With Quote

Old   November 7, 2006, 14:49
Default On the top level of AMG you ha
  #2
Senior Member
 
Hrvoje Jasak
Join Date: Mar 2009
Location: London, England
Posts: 1,906
Rep Power: 33
hjasak will become famous soon enough
On the top level of AMG you have a system of equations that needs to be solved to a tight tolerance and the question is how much work to leave for the top level. In serial runs, anything between a dozen and a couple of hundred cells will give you the same result (1000 is too much). In parallel, it is the communications that matter, which means you need to be more careful: make sure there's a sufficient number of cells for each CPU. You also need to worry about the case where you have multiple separate domains. In general, 20-30 cells per CPU in parallel will do you fine (until I do some more tricks to speed this up). :-)

Hrv
mgg and lourencosm like this.
__________________
Hrvoje Jasak
Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk
hjasak is offline   Reply With Quote

Old   November 7, 2006, 15:15
Default Thanks for that insightful rep
  #3
Senior Member
 
Srinath Madhavan (a.k.a pUl|)
Join Date: Mar 2009
Location: Edmonton, AB, Canada
Posts: 703
Rep Power: 21
msrinath80 is on a distinguished road
Thanks for that insightful reply Hrv. Much appreciated.
msrinath80 is offline   Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Dissimilarity in compiling new BC in 32bit OpenSuSE 103 versus 64bit CentOS 50 versus 64bit SLED 100 sponiar OpenFOAM Running, Solving & CFD 3 February 26, 2008 06:31
RIPPLE ICCG diverge brian CFX 0 February 24, 2007 00:26
AMG vs ICCG eugene OpenFOAM Running, Solving & CFD 2 June 7, 2005 14:07
[ICCG, CGSTAB ?] name Main CFD Forum 1 September 13, 2001 11:13
ICCG / ILU routines Jin Li Main CFD Forum 0 June 24, 2000 05:45


All times are GMT -4. The time now is 18:55.