[cig-commits] [commit] devel, master: added Mesh_Par_file_168 for 1D socal example to give the option of a 168-core simulation for a larger region. (bb91a32)
cig_noreply at geodynamics.org
cig_noreply at geodynamics.org
Wed Jun 18 16:48:39 PDT 2014
Repository : https://github.com/geodynamics/specfem3d
On branches: devel,master
Link : https://github.com/geodynamics/specfem3d/compare/6026e367984905ab133865f62fa6293b343759b9...47f703851338234f96397e7da9fbff63d8178b8a
>---------------------------------------------------------------
commit bb91a3250faf83edb1423425145559b5e154b72e
Author: Carl Tape <carltape at gi.alaska.edu>
Date: Tue Apr 5 00:07:57 2011 +0000
added Mesh_Par_file_168 for 1D socal example to give the option of a 168-core simulation for a larger region.
>---------------------------------------------------------------
bb91a3250faf83edb1423425145559b5e154b72e
homogeneous_halfspace/README | 4 ++-
.../socal1D/{Mesh_Par_file => Mesh_Par_file_168} | 24 ++++++-------
meshfem3D_examples/socal1D/README | 40 +++++++++++++++++++---
3 files changed, 50 insertions(+), 18 deletions(-)
diff --git a/homogeneous_halfspace/README b/homogeneous_halfspace/README
index 8d9536a..d2a94bf 100644
--- a/homogeneous_halfspace/README
+++ b/homogeneous_halfspace/README
@@ -24,11 +24,13 @@ step-by-step tutorial:
note: the same compiler used for SCOTCH must be used for SPECFEM3D (ifort, gfortran, pgf90)
- - copy three run scripts from utils/Cluster/ into SPECFEM3D/, e.g.,
+ - copy run scripts from utils/Cluster/ into SPECFEM3D/, e.g.,
pbs/go_decomposer_pbs.bash
pbs/go_generate_databases_pbs.bash
pbs/go_solver_pbs.bash
+ > cp utils/Cluster/pbs/*.bash .
+
note: you may need to adjust the commands (e.g., -q or -l)
diff --git a/meshfem3D_examples/socal1D/Mesh_Par_file b/meshfem3D_examples/socal1D/Mesh_Par_file_168
similarity index 73%
copy from meshfem3D_examples/socal1D/Mesh_Par_file
copy to meshfem3D_examples/socal1D/Mesh_Par_file_168
index acb120f..f6a0196 100644
--- a/meshfem3D_examples/socal1D/Mesh_Par_file
+++ b/meshfem3D_examples/socal1D/Mesh_Par_file_168
@@ -1,9 +1,9 @@
# coordinates of mesh block in latitude/longitude and depth in km
-LATITUDE_MIN = 33.790d0
-LATITUDE_MAX = 34.157d0
-LONGITUDE_MIN = -118.728d0
-LONGITUDE_MAX = -118.237d0
+LATITUDE_MIN = 32.2d0
+LATITUDE_MAX = 36.8d0
+LONGITUDE_MIN = -121.6d0
+LONGITUDE_MAX = -114.7d0
DEPTH_BLOCK_KM = 60.d0
UTM_PROJECTION_ZONE = 11
SUPPRESS_UTM_PROJECTION = .false.
@@ -14,12 +14,12 @@ INTERFACES_FILE = interfaces.dat
# number of elements at the surface along edges of the mesh at the surface
# (must be 8 * multiple of NPROC below if mesh is not regular and contains mesh doublings)
# (must be multiple of NPROC below if mesh is regular)
-NEX_XI = 24
-NEX_ETA = 24
+NEX_XI = 336
+NEX_ETA = 288
# number of MPI processors along xi and eta (can be different)
-NPROC_XI = 1
-NPROC_ETA = 1
+NPROC_XI = 14
+NPROC_ETA = 12
# Regular/irregular mesh
USE_REGULAR_MESH = .false.
@@ -51,8 +51,8 @@ NMATERIALS = 4
NREGIONS = 4
# define the different regions of the model as :
#NEX_XI_BEGIN #NEX_XI_END #NEX_ETA_BEGIN #NEX_ETA_END #NZ_BEGIN #NZ_END #material_id
-1 24 1 24 1 3 1
-1 24 1 24 4 6 2
-1 24 1 24 7 9 3
-1 24 1 24 10 12 4
+1 336 1 288 1 3 1
+1 336 1 288 4 6 2
+1 336 1 288 7 9 3
+1 336 1 288 10 12 4
diff --git a/meshfem3D_examples/socal1D/README b/meshfem3D_examples/socal1D/README
index 3431145..bba55d0 100644
--- a/meshfem3D_examples/socal1D/README
+++ b/meshfem3D_examples/socal1D/README
@@ -3,8 +3,16 @@ README
(/SPECFEM3D/examples/meshfem3D_examples/socal1D/README)
----------------------------------------------------------------------
-This is a very simple example to explain the use of the internal mesher meshfem3D.
-The example uses the standard 1D model used in southern California (SoCal-1D).
+This is an example of using the internal mesher meshfem3D for a 1D layered model.
+The 1D model is the standard model used in southern California:
+ Kanamori and Hadley (1975), Dreger and Helmberger (1990), Wald-Hutton,Given (1995)
+
+Figures for comparison:
+ socal1D_koma2004.png -- 1D model meshed in SPECFEM3D_BASIN (Komatitsch et al. 2004)
+ socal1D_sesame.png -- 1D model meshed here
+
+The resolution is similar to the the "2 s" accuracy mesh of Komatitsch et al. 2004 (DT = 0.011 s).
+It is the same size region as one slice (out of 168) of model m16 of Tape et al. (2009).
Option 1: run process.sh from local directory (must have access to mpi)
Option 2: follow instructions below for submitting separate jobs with qsub (or bsub)
@@ -22,7 +30,7 @@ Option 2: follow instructions below for submitting separate jobs with qsub (or b
> cp Par_file CMTSOLUTION STATIONS ..
> cd SPECFEM3D
- - adapt the submission scripts for the appropriate queues and number of cores:
+ - adapt the submission scripts for the appropriate queues (-q) and number of cores (-l):
go_mesher_pbs.bash, go_generate_databases_pbs.bash, go_solver_pbs.bash
- note: go_decomposer_pbs.bash is not used with the internal mesher
@@ -53,7 +61,7 @@ Option 2: follow instructions below for submitting separate jobs with qsub (or b
this will create binary mesh files, e.g. "proc000***_external_mesh.bin"
in directory in_out_files/DATABASES_MPI/.
- Optional: If you have paraview, load the .vtk files (see socal1D_sesame.png).
+ optional: If you have paraview, load the .vtk files (see socal1D_sesame.png).
4. run simulation:
@@ -63,7 +71,7 @@ Option 2: follow instructions below for submitting separate jobs with qsub (or b
- submit job script:
> qsub go_solver_pbs.bash
- - the job should take about 5 minutes
+ - the job should take about 2 minutes
- when the job is complete, you should have 3 sets (semd,semv,sema)
of 84 (ls -1 *semd | wc) seismogram files in the directory in_out_files/OUTPUT_FILES,
as well as 9 timestamp****** files
@@ -72,3 +80,25 @@ Option 2: follow instructions below for submitting separate jobs with qsub (or b
- xmgrace WTT*.semd &
----------------------------------------------------------------------
+To use the same 1D model but covering a larger region using 168 cores, follow these steps.
+
+1. modify Mesh_Par_file:
+
+ > cd in_data_files/meshfem3D_files
+ > cp Mesh_Par_file_168 Mesh_Par_file
+
+2. modify Par_file:
+
+ - open Par_file and modify NPROC: NPROC = 1 --> NPROC = 168
+
+ > cp Par_file CMTSOLUTION STATIONS ..
+ > cd ../../
+
+ - adapt the submission scripts for the appropriate queues (-q) and number of cores (-l):
+ go_mesher_pbs.bash, go_generate_databases_pbs.bash, go_solver_pbs.bash
+
+3. Follow steps 2-4 above.
+
+ note: slice 62 (out of 0-167) should contain the source and correspond to the 1-slice example above.
+
+----------------------------------------------------------------------
More information about the CIG-COMMITS
mailing list