[CIG-SHORT] benchmarks

[Carl W. Gable]

Good job Greg!!! I have the general impression
that the grids I created for the benchmark problems (reverse,
strike slip) are not being used. Or to turn that around, is
anyone using or going to use the grids I created for the
benchmarks and put on the web?

Also, has anyone or is anyone going to use the interpolation
utility that interpolates scalar values from one grid to
another. It seem that up to now all (nearly all???) the
benchmark runs (Greg and Charles) have been done on exactly
the same grid. Have any calculations been run on different
grids and solutions compared?

Carl

-------------------------------------
Hi Everybody:
  I've now performed the GeoFEST runs of the reverse-slip benchmark 
problem with the intermediate-resolution (500 meter) grid spacing.  As 
in my earlier postings, the input and output files will be found here:
  http://www.physics.hmc.edu/GL/workshop/
The GeoFEST-format input file is named tet4_500m.gft ; this and the 
previous 1000 meter input file have not yet been posted at CIG, but if 
somebody wants to do that, feel free.

I did just two runs of the 500m case:

tet4_halfkmA.out - 0.10 yr time step, 1.0 implicit parameter, 10^-7 
solver convergence criterion.
CPU time: 17880 s. on 2.4 GHz AMD Opteron Processor 250; 1.8 GB memory 
used

tet4_halfkmB.out - 0.10 yr time step, 1.0 implicit parameter, 10^-9 
solver convergence criterion.
CPU time: 21561 s. on 2.4 GHz AMD Opteron Processor 250; 1.8 GB memory 
used

The mesh coordinates and connectivity are identical to Charles' PyLith 
input.

                       - Greg

-------------------------------------
Gregory A. Lyzenga  <lyzenga at hmc.edu>           ***     (909) 621-8378
Dept. of Physics, Harvey Mudd College           *** fax (909) 621-8887
Claremont, CA 91711-5990 
<http://www.physics.hmc.edu/faculty/Lyzenga/>