Backend-specific implementation of EulerP Evaluator kernels. More...

#include <EulerP_Evaluator_impl.hpp>

Classes
struct	Cons2Prim_Arg
	Device-side argument struct for conservative-to-primitive conversion (no gradients/viscosity). More...

struct	Cons2PrimMu_Arg
	Device-side argument struct for conservative-to-primitive + viscosity kernel. More...

struct	EstEigenDt_Arg
	Device-side argument struct for eigenvalue estimation and time-step computation. More...

struct	Flux2nd_Arg
	Device-side argument struct for 2nd-order flux evaluation and RHS accumulation. More...

struct	RecFace2nd_Arg
	Device-side argument struct for 2nd-order face value reconstruction. More...

struct	RecGradient_Arg
	Device-side argument struct for gradient reconstruction kernels. More...

Public Types
using	t_Scalar_deviceViewVector_sup = std::unique_ptr< deviceViewVector< TUScalar::t_deviceView< B >, B > >
	Unique pointer to a device view vector of scalar array views.

using	t_ScalarGrad_deviceViewVector_sup = std::unique_ptr< deviceViewVector< TUScalarGrad::t_deviceView< B >, B > >
	Unique pointer to a device view vector of scalar-gradient array views.

Public Member Functions
void	RecGradient_GGRec (RecGradient_Arg &arg)
	Host specialization: Green-Gauss gradient reconstruction.

void	RecGradient_BarthLimiter (RecGradient_Arg &arg)
	Host specialization: Barth-Jespersen gradient limiter.

void	Cons2PrimMu (Cons2PrimMu_Arg &arg)
	Host specialization: conservative-to-primitive conversion with viscosity.

void	Cons2Prim (Cons2Prim_Arg &arg)
	Host specialization: conservative-to-primitive conversion (no gradients/viscosity).

void	EstEigenDt_GetFaceLam (EstEigenDt_Arg &arg)
	Host specialization: per-face eigenvalue estimation.

void	EstEigenDt_FaceLam2CellDt (EstEigenDt_Arg &arg)
	Host specialization: face eigenvalue accumulation to cell time steps.

void	RecFace2nd (RecFace2nd_Arg &arg)
	Host specialization: 2nd-order face value reconstruction.

void	Flux2nd (Flux2nd_Arg &arg)
	Host specialization: 2nd-order Roe flux evaluation and face-to-cell RHS scatter.

Static Public Member Functions
static void	RecGradient_GGRec (RecGradient_Arg &arg)
	Green-Gauss gradient reconstruction: boundary ghost values + cell gradient computation.

static void	RecGradient_BarthLimiter (RecGradient_Arg &arg)
	Barth-Jespersen gradient limiter applied to reconstructed gradients.

static void	Cons2PrimMu (Cons2PrimMu_Arg &arg)
	Executes conservative-to-primitive conversion with viscosity computation.

static void	Cons2Prim (Cons2Prim_Arg &arg)
	Executes conservative-to-primitive conversion (no gradients/viscosity).

static void	EstEigenDt_GetFaceLam (EstEigenDt_Arg &arg)
	First pass: computes per-face eigenvalue estimates from cell states.

static void	EstEigenDt_FaceLam2CellDt (EstEigenDt_Arg &arg)
	Second pass: accumulates face eigenvalues to cells and computes local dt.

static void	RecFace2nd (RecFace2nd_Arg &arg)
	Executes 2nd-order face value reconstruction from cell-centered data.

static void	Flux2nd (Flux2nd_Arg &arg)
	Evaluates 2nd-order Roe flux per face and scatters to cell RHS.

Detailed Description

template<DeviceBackend B>
struct DNDS::EulerP::Evaluator_impl

Backend-specific implementation of EulerP Evaluator kernels.

Each kernel (RecGradient, Cons2PrimMu, EstEigenDt, RecFace2nd, Flux2nd) is wrapped in an inner Arg struct that converts host-side shared_ptr arrays into device views, and a static method that dispatches the actual computation. The inner Portable struct within each Arg is trivially copyable for CUDA kernel launch.

Template Parameters

B	The DeviceBackend (Host or CUDA). Explicit specializations live in EulerP_Evaluator_impl.cpp (Host) and the CUDA compilation unit.

Definition at line 78 of file EulerP_Evaluator_impl.hpp.

Member Typedef Documentation

◆ t_Scalar_deviceViewVector_sup

template<DeviceBackend B>

using DNDS::EulerP::Evaluator_impl::t_Scalar_deviceViewVector_sup = std::unique_ptr<deviceViewVector<TUScalar::t_deviceView, B> >

Unique pointer to a device view vector of scalar array views.

Definition at line 81 of file EulerP_Evaluator_impl.hpp.

◆ t_ScalarGrad_deviceViewVector_sup

template<DeviceBackend B>

using DNDS::EulerP::Evaluator_impl::t_ScalarGrad_deviceViewVector_sup = std::unique_ptr<deviceViewVector<TUScalarGrad::t_deviceView, B> >

Unique pointer to a device view vector of scalar-gradient array views.

Definition at line 83 of file EulerP_Evaluator_impl.hpp.

Member Function Documentation

◆ Cons2Prim() [1/2]

void DNDS::EulerP::Evaluator_impl::Cons2Prim ( Cons2Prim_Arg & arg )

Host specialization: conservative-to-primitive conversion (no gradients/viscosity).

OpenMP-parallel loop over all points in the u array, calling Cons2Prim_Kernel for each cell.

Definition at line 157 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ Cons2Prim() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::Cons2Prim ( Cons2Prim_Arg & arg )

static

Executes conservative-to-primitive conversion (no gradients/viscosity).

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ Cons2PrimMu() [1/2]

void DNDS::EulerP::Evaluator_impl::Cons2PrimMu ( Cons2PrimMu_Arg & arg )

Host specialization: conservative-to-primitive conversion with viscosity.

OpenMP-parallel loop over all points in the u array, calling Cons2PrimMu_Kernel for each cell.

Definition at line 125 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ Cons2PrimMu() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::Cons2PrimMu ( Cons2PrimMu_Arg & arg )

static

Executes conservative-to-primitive conversion with viscosity computation.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ EstEigenDt_FaceLam2CellDt() [1/2]

void DNDS::EulerP::Evaluator_impl::EstEigenDt_FaceLam2CellDt ( EstEigenDt_Arg & arg )

Host specialization: face eigenvalue accumulation to cell time steps.

OpenMP-parallel loop over all owned cells, calling EstEigenDt_FaceLam2CellDt_Kernel for each cell.

Definition at line 227 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ EstEigenDt_FaceLam2CellDt() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::EstEigenDt_FaceLam2CellDt ( EstEigenDt_Arg & arg )

static

Second pass: accumulates face eigenvalues to cells and computes local dt.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ EstEigenDt_GetFaceLam() [1/2]

void DNDS::EulerP::Evaluator_impl::EstEigenDt_GetFaceLam ( EstEigenDt_Arg & arg )

Host specialization: per-face eigenvalue estimation.

OpenMP-parallel loop over all processor-local faces, calling EstEigenDt_GetFaceLam_Kernel for each face.

Definition at line 189 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ EstEigenDt_GetFaceLam() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::EstEigenDt_GetFaceLam ( EstEigenDt_Arg & arg )

static

First pass: computes per-face eigenvalue estimates from cell states.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ Flux2nd() [1/2]

void DNDS::EulerP::Evaluator_impl::Flux2nd ( Flux2nd_Arg & arg )

Host specialization: 2nd-order Roe flux evaluation and face-to-cell RHS scatter.

First loop (OpenMP parallel): computes per-face numerical flux via Flux2nd_Kernel_FluxFace. Second loop (OpenMP parallel): scatters face fluxes to cell RHS via Flux2nd_Kernel_Face2Cell.

Definition at line 288 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ Flux2nd() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::Flux2nd ( Flux2nd_Arg & arg )

static

Evaluates 2nd-order Roe flux per face and scatters to cell RHS.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ RecFace2nd() [1/2]

void DNDS::EulerP::Evaluator_impl::RecFace2nd ( RecFace2nd_Arg & arg )

Host specialization: 2nd-order face value reconstruction.

OpenMP-parallel loop over all processor-local faces, calling RecFace2nd_Kernel for each face.

Definition at line 256 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ RecFace2nd() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::RecFace2nd ( RecFace2nd_Arg & arg )

static

Executes 2nd-order face value reconstruction from cell-centered data.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ RecGradient_BarthLimiter() [1/2]

void DNDS::EulerP::Evaluator_impl::RecGradient_BarthLimiter ( RecGradient_Arg & arg )

Host specialization: Barth-Jespersen gradient limiter.

First loop (OpenMP parallel): applies the flow-variable limiter to all owned cells. Second loop (OpenMP parallel, conditional): applies the scalar-variable limiter if there are transported scalars.

Definition at line 77 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ RecGradient_BarthLimiter() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::RecGradient_BarthLimiter ( RecGradient_Arg & arg )

static

Barth-Jespersen gradient limiter applied to reconstructed gradients.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

◆ RecGradient_GGRec() [1/2]

void DNDS::EulerP::Evaluator_impl::RecGradient_GGRec ( RecGradient_Arg & arg )

Host specialization: Green-Gauss gradient reconstruction.

First loop (serial): generates boundary ghost values for all boundary faces. Second loop (OpenMP parallel): computes the Green-Gauss cell gradient for all owned cells.

Definition at line 27 of file EulerP_Evaluator_impl.cpp.

Here is the call graph for this function:

◆ RecGradient_GGRec() [2/2]

template<DeviceBackend B>

static void DNDS::EulerP::Evaluator_impl::RecGradient_GGRec ( RecGradient_Arg & arg )

static

Green-Gauss gradient reconstruction: boundary ghost values + cell gradient computation.

Parameters

arg	Device-side argument struct with all required views.

Here is the caller graph for this function:

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

/home/runner/work/DNDSR/DNDSR/src/EulerP/EulerP_Evaluator_impl.hpp

Classes

Public Types

Public Member Functions

Static Public Member Functions

Detailed Description

Member Typedef Documentation

◆ t_Scalar_deviceViewVector_sup

◆ t_ScalarGrad_deviceViewVector_sup

Member Function Documentation

◆ Cons2Prim() [1/2]

◆ Cons2Prim() [2/2]

◆ Cons2PrimMu() [1/2]

◆ Cons2PrimMu() [2/2]

◆ EstEigenDt_FaceLam2CellDt() [1/2]

◆ EstEigenDt_FaceLam2CellDt() [2/2]

◆ EstEigenDt_GetFaceLam() [1/2]

◆ EstEigenDt_GetFaceLam() [2/2]

◆ Flux2nd() [1/2]

◆ Flux2nd() [2/2]

◆ RecFace2nd() [1/2]

◆ RecFace2nd() [2/2]

◆ RecGradient_BarthLimiter() [1/2]

◆ RecGradient_BarthLimiter() [2/2]

◆ RecGradient_GGRec() [1/2]

◆ RecGradient_GGRec() [2/2]