diff options
Diffstat (limited to 'src/Lagrange-tensor-product/README')
| -rw-r--r-- | src/Lagrange-tensor-product/README | 69 |
1 files changed, 69 insertions, 0 deletions
diff --git a/src/Lagrange-tensor-product/README b/src/Lagrange-tensor-product/README new file mode 100644 index 0000000..44cd090 --- /dev/null +++ b/src/Lagrange-tensor-product/README @@ -0,0 +1,69 @@ +$Header$ + +This directory contains code for Lagrange polynomial interpolation, +using the tensor-product definition for multiple dimensions. + +makefile + This has targets to create the subdirectories for the + Maple-generated coefficient files, and to run Maple to + generated those files. + +make.code.defn + This tells the Cactus build system which files to compile. + +[123]d.maple + These files contain the Maple code to generate the interpolation + coefficients. + +fns.maple + These files define the interpolation functions, coefficients, etc, + for the Maple code in [123]d.maple + +[123]d.cube.order?.smooth?.c + These files are the individual cases for the interpolation. + Each file defines a separate function. Each file just #defines + a bunch of macros, then #includes "../template.c" to get the + actual interpolation code. Unfortunately, the ISO C standard + only guarantees 31 characters for function names (in fact, it + doesn't even guarantee that for external names -- they may be + limited by a system linker -- but I quietly ignore that and + assume I can use 31), so the functions are named in a cryptic + manner: + 123456789 123456789 123456789 1 + AEILocalInterp_U_LagTP_1cube_10 + AEILocalInterp_U_LagTP_1cube_20 + AEILocalInterp_U_LagTP_1cube_30 + AEILocalInterp_U_LagTP_1cube_40 + AEILocalInterp_U_LagTP_1cube_50 + AEILocalInterp_U_LagTP_1cube_60 + AEILocalInterp_U_LagTP_2cube_10 + AEILocalInterp_U_LagTP_2cube_20 + AEILocalInterp_U_LagTP_2cube_30 + AEILocalInterp_U_LagTP_2cube_40 + AEILocalInterp_U_LagTP_3cube_10 + AEILocalInterp_U_LagTP_3cube_20 + AEILocalInterp_U_LagTP_3cube_30 + AEILocalInterp_U_LagTP_3cube_40 + + The naming convention is as follows (eg for the last function) + AEILocalInterp name of this thorn + U these functions are all for the + CCTK_InterpLocalUniform() API + LagTP these functions are all for the + "Lagrange polynomial interpolation (tensor product)" + interpolation operator + 3 this function does 3-d interpolation + cube this function uses a cube-shaped + interpolation molecule + 4 this function does 4th order interpolation + 0 this function does smoothing=0 interpolation + + The same naming convention is used for the other interpolation + operators, except that for Hermite the smoothing part is dropped: + LagTP "Lagrange polynomial interpolation (tensor product)" + LagMD "Lagrange polynomial interpolation (maximum degree)" + Herm "Hermite polynomial interpolation" + +all_prototypes.h + This file contains C prototypes for each of the functions + in [123]d.cube.order?.smooth0.c. |