olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR > Class Template Reference

Device-side view of a block lattice. More...

#include <context.hh>

Inheritance diagram for olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >:

Collaboration diagram for olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >:

Public Member Functions
	DeviceBlockLattice (ConcreteBlockLattice< T, DESCRIPTOR, Platform::GPU_CUDA > &lattice) __host__
CellID	getCellId (LatticeR< DESCRIPTOR::d > loc) const __device__
CellDistance	getNeighborDistance (LatticeR< DESCRIPTOR::d > dir) const __device__
Cell< T, DESCRIPTOR >	get (CellID iCell) __device__
Cell< T, DESCRIPTOR >	get (LatticeR< DESCRIPTOR::d > loc) __device__
LatticeR< DESCRIPTOR::d >	getLatticeR (CellID iCell) const __device__
bool	isInside (LatticeR< DESCRIPTOR::d > loc) const __device__
bool	isPadding (LatticeR< DESCRIPTOR::d > loc) const __device__
bool	isPadding (CellID iCell) const __device__
Public Member Functions inherited from olb::gpu::cuda::DeviceContext< T, DESCRIPTOR >
template<concepts::ConcreteFieldsContainer CONTAINER>
	DeviceContext (CONTAINER &lattice) __host__
	DeviceContext (ConcreteBlockD< T, DESCRIPTOR, Platform::GPU_CUDA > &lattice) __host__
std::size_t	getNcells () const any_platform
template<typename FIELD >
FIELD::template value_type< T > **	getField () __device__

Additional Inherited Members
Public Types inherited from olb::gpu::cuda::DeviceContext< T, DESCRIPTOR >
using	value_t = T
using	descriptor_t = DESCRIPTOR

Detailed Description

template<typename T, typename DESCRIPTOR>
class olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >

Device-side view of a block lattice.

Used for non-local operators such as post processors

Definition at line 244 of file context.hh.

Constructor & Destructor Documentation

DeviceBlockLattice()

template<typename T , typename DESCRIPTOR >

olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::DeviceBlockLattice ( ConcreteBlockLattice< T, DESCRIPTOR, Platform::GPU_CUDA > & lattice )

inline

Definition at line 251 of file context.hh.

                                                                                           :
    DeviceContext<T,DESCRIPTOR>(lattice),
    _core(lattice.getExtent()),
    _padding{lattice.getPadding()}
  {
    auto size = lattice.getExtent() + 2*_padding;
    if constexpr (DESCRIPTOR::d == 3) {
      _projection = {size[1]*size[2], size[2], 1};
    } else {
      _projection = {size[1], 1};
    }
  }

Member Function Documentation

get() [1/2]

template<typename T , typename DESCRIPTOR >

Cell< T, DESCRIPTOR > olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::get ( CellID iCell )

inline

Definition at line 272 of file context.hh.

                                                  {
    return Cell<T,DESCRIPTOR>(*this, iCell);
  }

Here is the caller graph for this function:

get() [2/2]

template<typename T , typename DESCRIPTOR >

Cell< T, DESCRIPTOR > olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::get ( LatticeR< DESCRIPTOR::d > loc )

inline

Definition at line 276 of file context.hh.

                                                               {
    return get(getCellId(loc));
  }

References olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::get(), and olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::getCellId().

Here is the call graph for this function:

getCellId()

template<typename T , typename DESCRIPTOR >

CellID olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::getCellId ( LatticeR< DESCRIPTOR::d > loc ) const

inline

Definition at line 264 of file context.hh.

                                                               {
    return (loc + _padding) * _projection;
  }

Here is the caller graph for this function:

getLatticeR()

template<typename T , typename DESCRIPTOR >

LatticeR< DESCRIPTOR::d > olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::getLatticeR ( CellID iCell ) const

inline

Definition at line 280 of file context.hh.

                                                                   {
    if constexpr (DESCRIPTOR::d == 3) {
      const std::int32_t iX = iCell / _projection[0];
      const std::int32_t remainder = iCell % _projection[0];
      const std::int32_t iY = remainder / _projection[1];
      const std::int32_t iZ = remainder % _projection[1];
      return LatticeR<DESCRIPTOR::d>{iX - _padding, iY - _padding, iZ - _padding};
    } else {
      const std::int32_t iX = iCell / _projection[0];
      const std::int32_t iY = iCell % _projection[0];
      return LatticeR<DESCRIPTOR::d>{iX - _padding, iY - _padding};
    }
  };

Here is the caller graph for this function:

getNeighborDistance()

template<typename T , typename DESCRIPTOR >

CellDistance olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::getNeighborDistance ( LatticeR< DESCRIPTOR::d > dir ) const

inline

Definition at line 268 of file context.hh.

                                                                               {
    return dir * _projection;
  }

isInside()

template<typename T , typename DESCRIPTOR >

bool olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::isInside ( LatticeR< DESCRIPTOR::d > loc ) const

inline

Definition at line 294 of file context.hh.

                                                            {
    return loc >= -_padding && loc < _core + _padding;
  }

Here is the caller graph for this function:

isPadding() [1/2]

template<typename T , typename DESCRIPTOR >

bool olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::isPadding ( CellID iCell ) const

inline

Definition at line 302 of file context.hh.

                                                {
    return isPadding(getLatticeR(iCell));
  }

References olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::getLatticeR(), and olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::isPadding().

Here is the call graph for this function:

isPadding() [2/2]

template<typename T , typename DESCRIPTOR >

bool olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::isPadding ( LatticeR< DESCRIPTOR::d > loc ) const

inline

Definition at line 298 of file context.hh.

                                                             {
    return isInside(loc) && !(loc >= 0 && loc < _core);
  }

References olb::gpu::cuda::DeviceBlockLattice< T, DESCRIPTOR >::isInside().

Here is the call graph for this function:

Here is the caller graph for this function:

---

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

• src/core/platform/gpu/cuda/context.hh