orb5x::MassMatrix Class Reference

Constructs small element-local B-spline overlap operators. More...

#include "mass_matrix.hpp"

Static Public Member Functions
static void	compute_1d_mass_matrix (int nidbas, Real dx, Kokkos::View< Real ** > &mass_mat)
	Compute \(M_{ij}=\int B_i(x)B_j(x)\,dx\) on one uniform cell.
static void	compute_2d_mass_matrix (int nidbas_s, int nidbas_chi, Real ds, Real dchi, Kokkos::View< Real **** > &mass_mat_2d)
	Form the tensor product of radial and poloidal 1-D mass matrices.
static void	compute_inverse_mass_fourier (const Kokkos::View< Complex * > &mass_fourier, Kokkos::View< Complex * > &inv_mass_fourier, int nphi)
	Compute a guarded reciprocal for each Fourier coefficient.
static void	compute_mass_matrix_toroidal_deriv (int nidbas, Real dx, Kokkos::View< Real ** > &mass_mat_deriv)
	Compute \(D_{ij}=\int B_i(x)\,dB_j/dx\,dx\) on one cell.
static void	project_to_fourier (const Kokkos::View< Real ** > &mass_bspl, Kokkos::View< Complex * > &mass_fourier, int nphi)
	Populate per-mode coefficients from a B-spline mass matrix.

Detailed Description

Constructs small element-local B-spline overlap operators.

These helpers create host-computed Kokkos views used to map between local B-spline coefficients and integral weights. They are separate from the production distributed matrices assembled by MatrixCoeff/SolverModule. Quadratic coefficients and Fourier projection in this utility are currently approximate, as stated by the implementation.

Member Function Documentation

compute_1d_mass_matrix()

void orb5x::MassMatrix::compute_1d_mass_matrix ( int nidbas, Real dx, Kokkos::View< Real ** > & mass_mat )

static

Compute \(M_{ij}=\int B_i(x)B_j(x)\,dx\) on one uniform cell.

Parameters

[in]	nidbas	B-spline degree; explicit coefficients exist for 1–3.
[in]	dx	Coordinate-space cell width; the matrix has the same units.
[in,out]	mass_mat	Output (nidbas+1)^2 device view, allocated when empty.

Note

Unsupported degrees produce a diagonal matrix scaled by dx.

compute_2d_mass_matrix()

void orb5x::MassMatrix::compute_2d_mass_matrix ( int nidbas_s, int nidbas_chi, Real ds, Real dchi, Kokkos::View< Real **** > & mass_mat_2d )

static

Form the tensor product of radial and poloidal 1-D mass matrices.

Parameters

[in]	nidbas_s	Radial B-spline degree.
[in]	nidbas_chi	Poloidal B-spline degree.
[in]	ds	Radial cell width in normalized s coordinates.
[in]	dchi	Poloidal cell width [rad].
[in,out]	mass_mat_2d	Four-index tensor, allocated when empty.

compute_inverse_mass_fourier()

void orb5x::MassMatrix::compute_inverse_mass_fourier ( const Kokkos::View< Complex * > & mass_fourier, Kokkos::View< Complex * > & inv_mass_fourier, int nphi )

static

Compute a guarded reciprocal for each Fourier coefficient.

Parameters

[in]	mass_fourier	Complex mode coefficients.
[in,out]	inv_mass_fourier	Output reciprocals, allocated when empty.
[in]	nphi	Number of coefficients to process.

Note

Magnitudes at or below 1e-12 map to zero.

compute_mass_matrix_toroidal_deriv()

void orb5x::MassMatrix::compute_mass_matrix_toroidal_deriv ( int nidbas, Real dx, Kokkos::View< Real ** > & mass_mat_deriv )

static

Compute \(D_{ij}=\int B_i(x)\,dB_j/dx\,dx\) on one cell.

Parameters

[in]	nidbas	B-spline degree.
[in]	dx	Cell width; retained for interface parity although the implemented degree 1–3 reference coefficients are dimensionless.
[in,out]	mass_mat_deriv	Output square device view.

project_to_fourier()

void orb5x::MassMatrix::project_to_fourier ( const Kokkos::View< Real ** > & mass_bspl, Kokkos::View< Complex * > & mass_fourier, int nphi )

static

Populate per-mode coefficients from a B-spline mass matrix.

Parameters

[in]	mass_bspl	Square local mass matrix.
[in,out]	mass_fourier	Complex output of length nphi, allocated when empty.
[in]	nphi	Number of toroidal Fourier modes.

Warning

The current implementation assigns the same element average to every mode; it is not a full discrete Fourier transform.

---

The documentation for this class was generated from the following file:

• mass_matrix.hpp