CurrentModule = IncompressibleNavierStokes
All discrete operators are built using KernelAbstractions.jl and Cartesian indices, similar to WaterLily.jl. This allows for dimension- and backend-agnostic code. See this blog post for how to write kernels. IncompressibleNavierStokes previously relied on assembling sparse operators to perform the same operations. While being very efficient and also compatible with CUDA (CUSPARSE), storing these matrices in memory is expensive for large 3D problems.
Offset
avg
divergence!
divergence
vorticity
vorticity!
convection!
diffusion!
bodyforce!
momentum!
momentum
laplacian!
laplacian
pressuregradient!
pressuregradient
interpolate_u_p
interpolate_u_p!
interpolate_ω_p
interpolate_ω_p!
Dfield!
Dfield
Qfield!
Qfield
eig2field!
eig2field
kinetic_energy
kinetic_energy!
total_kinetic_energy
tensorbasis
divoftensor!
tensorbasis!
get_scale_numbers
dissipation_from_strain!