Study guide: Finite difference methods for wave motion

$$ \newcommand{\uex}{{u_{\small\mbox{e}}}} \newcommand{\half}{\frac{1}{2}} \newcommand{\halfi}{{1/2}} \newcommand{\xpoint}{\boldsymbol{x}} \newcommand{\normalvec}{\boldsymbol{n}} \newcommand{\Oof}[1]{\mathcal{O}(#1)} \newcommand{\Ix}{\mathcal{I}_x} \newcommand{\Iy}{\mathcal{I}_y} \newcommand{\It}{\mathcal{I}_t} \newcommand{\setb}[1]{#1^0} % set begin \newcommand{\sete}[1]{#1^{-1}} % set end \newcommand{\setl}[1]{#1^-} \newcommand{\setr}[1]{#1^+} \newcommand{\seti}[1]{#1^i} \newcommand{\Real}{\mathbb{R}} $$

How to avoid array copying

Two-dimensional arrays are stored row by row in Python and C

Two-dimensional arrays are stored column by column in Fortran

f2py takes a copy of a numpy (C) array and transposes it when calling Fortran

Such copies are time and memory consuming

Remedy: declare numpy arrays with Fortran storage

order = 'Fortran' if version == 'f77' else 'C'
u   = zeros((Nx+1,Ny+1), order=order)
u_1 = zeros((Nx+1,Ny+1), order=order)
u_2 = zeros((Nx+1,Ny+1), order=order)

Option -DF2PY_REPORT_ON_ARRAY_COPY=1 makes f2py write out array copying:

Terminal> f2py -c wave2D_u0_loop_f77.pyf --build-dir build_f77 \ 
          -DF2PY_REPORT_ON_ARRAY_COPY=1 wave2D_u0_loop_f77.f