| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
| |
The interpolation code in template.c is now split up into subfunctions
(in common/, Lagrange/, and Hermite/), so it should now compile in
finite cpu/memory even with optimization.
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@113 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
parameter-table entry to allow the min/max ends of the grid to be
handled differently. (We'll need this for the multiprocessor interpolation.)
***** IMPORTANT *****
This change is *NOT* backwards-compatible: if you are calling
CCTK_InterpLocalUniform() from this thorn (CactusBase/LocalInterp),
and you have set out_of_range_tolerance[] to a non-default value,
you must (slightly) change your code, or else you'll get an error
return from the interpolator.
If you don't use out_of_range_tolerance[] , then you don't need
to do anything.
The semantics of out_of_range_tolerance[] are now as follows:
@var out_of_range_tolerance
@vdesc Specifies how out-of-range interpolation points should
be handled. The array elements are matched up with
the axes and minimum/maximum ends of the grid in the
order [x_min, x_max, y_min, y_max, z_min, z_max, ...].
An array value TOL is interpreted as follows:
If TOL >= 0.0,
then an interpolation point is considered to be
"out of range" if and only if the interpolation
point is > TOL * coord_delta[axis]
outside the grid in this coordinate direction.
If TOL == -1.0,
then an interpolation point is considered to be
"out of range" if and only if a centered molecule
(or more generally, a molecule whose centering
is chosen pretending that the grid is of infinite
extent), would require data from outside the grid
in this direction.
Other values of TOL are illegal (reserved for future use).
@vtype const CCTK_REAL out_of_range_tolerance[2*N_dims]
@endvar
The change is that it used to have only N_dims elements, one for
each axis -- now there are 2*N_dims elements, one for each min/max
end of each axis.
----------------------------------------------------------------------
Modified Files:
doc/documentation.tex
src/GeneralizedPolynomial-Uniform/InterpLocalUniform.c
src/GeneralizedPolynomial-Uniform/InterpLocalUniform.h
src/GeneralizedPolynomial-Uniform/interpolate.maple
src/GeneralizedPolynomial-Uniform/molecule_posn.c
src/GeneralizedPolynomial-Uniform/template.c
src/GeneralizedPolynomial-Uniform/test_molecule_posn.c
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@88 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
|
|
| |
coefficients, and the coefficient files themselves, to properly support
multiple interpolation operators, and adds some skeleton support for
Hermite interpolation. I will add full support for Hermite in a future
checkin.
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@81 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@63 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@45 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
| |
-- no actual Jacobian query code yet, though
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@40 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
| |
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@34 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
| |
--> reduces the number of args quite a bit
* streamline notation for molecule min/max m and position queries
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@29 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
|
|
| |
back into CVS:
main change is to Jacobian structure query API (code+docs),
also implement about 2/3 of this in the code
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@22 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
|
|
| |
1.0e4*DBL_EPSILON
--> this at least scales with machine precision,
abeit it does suffer from the serious problem that C double
might not be the right type to use (CCTK_REAL)
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@20 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
|
|
|
|
|
|
|
| |
since Ihad a test case where rounding errors pushed a point which
should have been just at the edge of the domain, out to about 1.5e-14
grid points outside it. :(
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@17 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|
|
git-svn-id: http://svn.cactuscode.org/arrangements/CactusBase/LocalInterp/trunk@2 df1f8a13-aa1d-4dd4-9681-27ded5b42416
|