There are a few things to note about BoxMG. First, it can only be used for domains that can be mapped to a square in 2D or a cube in 3D. It was brought into CMISS for solving large scale cardiac tissue problems as it is ideally suited for such domains. We often say that such domains are logically rectangular. Second, there are many options when using BoxMG but I have hidden most of them from the unexperienced user. The most important option to be aware of is the relaxation (or smoothing) method. If the problem is highly anisotropric, either from conductivities or a stretched mesh, then you must use line relaxation in 2D and plane relaxation in 3D. Finally, it should be noted that no benefit will be gained until the number of unknowns begins approaching 500,000. For less than 250,000 unknowns it will probably be faster to use either LU or CG with ILU. These are not hard and fast rules but something to keep in mind.
References:
Solving the cardiac bidomain equations for discontinuous conductivities, IEEE Trans. Biomedical Engineering, 53, 7, pp. 1265-1272, 2006
Fast elliptic solvers for cardiac bioelectric modeling on structured meshes, submitted to Computers in Biology and Medicine in 2007
# # Example b211h # # ------------ # ------------------------------------- # Set parameters and coors for problem # ------------------------------------- fem define para;r;cube;example fem define coor;r;cube;example # ------------------------------------- # Define elements and fibres # ------------------------------------- fem define node;r;cube;example fem define base;r;cube;example fem define elem;r;cube;example fem define fibr;d fem define elem;r;cube;example fibre # ------------------------------------- # Define grids # ------------------------------------- fem define grid;r;cube;example fem update grid geometry fem group grid external as boundary fem group grid xi1 low as stimulus fem group grid xi1 high as dirichlet # ------------------------------------- # Define equations, materials, and cells # ------------------------------------- fem define equa;r;cube;example class 1,2 fem define mate;r;cube;example class 1,2 fem define cell;r;cube;example class 1,2 # --------------------------------------------- # Define initial and boundary conds and solver # --------------------------------------------- fem define init;r;cube;example class 1,2 fem define solv;r;pcg;example class 1 fem define solv;r;bmgpl;example class 2 ############################################# # # With this example if you replace bmgpl in # the above line with bmtpt you will see the # poor convergence as a result of the stretched # mesh. Note that the domain is # # [0,10] X [0,1] X [0,1] # ############################################# # ------- # SOLVE # ------- fem solve to 1.0 class 1,2
#Testing example b212h fem li grid yq index v class 1
Name Modified Size
example_b212h.com 27-Apr-2007 1.6k bmgpl.ipsolv 27-Apr-2007 2.4k bmgpt.ipsolv 27-Apr-2007 2.5k cube.ipbase 27-Apr-2007 2.4k cube.ipcell 27-Apr-2007 1.7k cube.ipcoor 27-Apr-2007 570 cube.ipelem 27-Apr-2007 404 cube.ipelfb 27-Apr-2007 233 cube.ipequa 27-Apr-2007 1.5k cube.ipgrid 27-Apr-2007 600 cube.ipinit 27-Apr-2007 905 cube.ipmate 27-Apr-2007 1.7k cube.ipnode 27-Apr-2007 1.8k cube.ippara 27-Apr-2007 6.2k cube.ipsolv 27-Apr-2007 1.8k pcg.ipsolv 27-Apr-2007 2.1k test_output.com 27-Apr-2007 54
Name Modified Size
examples_b_b2_b21_b212_b212h.tar.gz 28-Apr-2007 10k
Status | Tested | Real time (s) | |
i686-linux | |||
cm | Failure | Sun Mar 6 00:06:59 2016 | 0 |
last break | Sun Apr 29 01:24:00 2007 | 2 | |
cm-debug | Failure | Sun Mar 6 00:07:09 2016 | 2 |
last break | Sat Apr 28 01:52:00 2007 | 10 | |
mips-irix | |||
cm | Failure | Tue Aug 21 01:47:33 2007 | 6 |
last break | Sun Apr 29 01:31:00 2007 | 7 | |
cm-debug | Failure | Tue Aug 21 02:03:22 2007 | 27 |
last break | Wed May 2 01:28:00 2007 | 27 | |
cm-debug-clear-malloc | Failure | Tue Aug 21 02:09:19 2007 | 25 |
last break | Sat Apr 28 01:26:00 2007 | 26 | |
cm-debug-clear-malloc7 | Failure | Tue Aug 21 02:09:45 2007 | 26 |
last break | Mon Apr 30 01:29:00 2007 | 25 | |
cm64 | Failure | Tue Aug 21 01:44:58 2007 | 6 |
last break | Sun Apr 29 01:31:00 2007 | 8 | |
cm64-debug | Failure | Tue Aug 21 02:14:02 2007 | 27 |
last break | Tue May 1 01:29:00 2007 | 32 | |
rs6000-aix | |||
cm | Failure | Wed Mar 4 01:10:39 2009 | 1 |
last break | Sun Apr 29 02:28:00 2007 | 2 | |
cm-debug | Failure | Wed Mar 4 01:18:13 2009 | 6 |
last break | Thu Sep 13 02:09:00 2007 | 6 | |
cm64 | Failure | Wed Mar 4 01:10:30 2009 | 1 |
last break | Sun Apr 29 02:28:00 2007 | 2 | |
cm64-debug | Failure | Wed Mar 4 01:19:05 2009 | 6 |
last break | Thu Sep 13 02:09:00 2007 | 6 | |
x86_64-linux | |||
cm | Failure | Sun Mar 6 00:01:38 2016 | 0 |
last break | Tue Mar 11 00:20:00 2008 | 1 | |
cm-debug | Failure | Sun Mar 6 00:01:39 2016 | 1 |
last break | Sat Apr 28 01:17:00 2007 | 3 |
rs6000-aix | |||
SIGSEGV | cm: | exit due to SIGSEGV signal. |
i686-linux | |||
Failure | cm: | ndiff test: significant differences with generic answer. | |
Failure | cm-debug: | ndiff test: significant differences with generic answer. | |
mips-irix | |||
Failure | cm: | ndiff test: significant differences with generic answer. | |
Failure | cm-debug: | ndiff test: significant differences with generic answer. | |
Failure | cm-debug-clear-malloc: | ndiff test: significant differences with generic answer. | |
Failure | cm-debug-clear-malloc7: | ndiff test: significant differences with generic answer. | |
Failure | cm64: | ndiff test: significant differences with generic answer. | |
Failure | cm64-debug: | ndiff test: significant differences with generic answer. | |
rs6000-aix | |||
Missing | cm: | output file not generated for ndiff; generic answer. | |
Failure | cm-debug: | ndiff test: significant differences with generic answer. | |
Failure | cm64: | ndiff test: significant differences with generic answer. | |
Failure | cm64-debug: | ndiff test: significant differences with generic answer. | |
x86_64-linux | |||
Failure | cm: | ndiff test: significant differences with generic answer. | |
Failure | cm-debug: | ndiff test: significant differences with generic answer. |
Html last generated: Sun Mar 6 05:51:12 2016
Input last modified: Fri Apr 27 08:54:23 2007