fortran-使用fortran实现的三线性插值算法.zip

# Cubic spline interpolation in Fortran A module for cubic spline interpolation in Fortran, based on: * [spline.f](https://www.netlib.org/fmm/spline.f) * [spline.f90](https://ww2.odu.edu/~agodunov/computing/programs/book2/Ch01/spline.f90), a Fortran 90 translation of `spline.f` ## Requirements A somewhat recent Fortran compiler (`gfortran` is the default compiler in the `Makefile`). ## Usage Usage looks like this: type(spline_t) :: spl ! Compute coefficients, n is size of input table call spline_set_coeffs(x_table, y_table, n, spl) ! Evaluate at x_eval using the elemental function spline_evaluate y = spline_evaluate(x_eval, spl) ## Example See `test.f90`, which can be compiled with: make and executed with for example: ./test 100 1000000 which uses a table with 100 points of a cubic function as input, which is then evaluated at 1000000 points using cubic spline interpolation. Example of output: Using n_table = 100 and n_eval = 1000000 Error (L1, L2, Linf): 0.91803402E-16 0.19963740E-15 0.17763568E-14 Evaluated points/second: 0.92936803E+08 Note that the error is so small here because a cubic spline can 'exactly' approximate a cubic function. For generic function interpolation, errors will be larger.