|
MUDA_HOST | BufferLaunch (cudaStream_t s=nullptr) MUDA_NOEXCEPT |
|
MUDA_HOST | BufferLaunch (int block_dim, cudaStream_t s=nullptr) MUDA_NOEXCEPT |
|
MUDA_HOST | BufferLaunch (int grid_dim, int block_dim, cudaStream_t s=nullptr) MUDA_NOEXCEPT |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer< T > &buffer, size_t size) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer2D< T > &buffer, Extent2D extent) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer3D< T > &buffer, Extent3D extent) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | reserve (DeviceBuffer< T > &buffer, size_t capacity) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | reserve (DeviceBuffer2D< T > &buffer, Extent2D capacity) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | reserve (DeviceBuffer3D< T > &buffer, Extent3D capacity) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer< T > &buffer, size_t size, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer2D< T > &buffer, Extent2D extent, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | resize (DeviceBuffer3D< T > &buffer, Extent3D extent, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | clear (DeviceBuffer< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | clear (DeviceBuffer2D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | clear (DeviceBuffer3D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | alloc (DeviceBuffer< T > &buffer, size_t n) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | alloc (DeviceBuffer2D< T > &buffer, Extent2D extent) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | alloc (DeviceBuffer3D< T > &buffer, Extent3D extent) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | free (DeviceBuffer< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | free (DeviceBuffer2D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | free (DeviceBuffer3D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | shrink_to_fit (DeviceBuffer< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | shrink_to_fit (DeviceBuffer2D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | shrink_to_fit (DeviceBuffer3D< T > &buffer) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (VarView< T > dst, CVarView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (BufferView< T > dst, CBufferView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer2DView< T > dst, CBuffer2DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer3DView< T > dst, CBuffer3DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (VarView< T > dst, VarView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (BufferView< T > dst, BufferView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer2DView< T > dst, Buffer2DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer3DView< T > dst, Buffer3DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< VarView< T >> &dst, const ComputeGraphVar< VarView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< BufferView< T >> &dst, const ComputeGraphVar< BufferView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< Buffer2DView< T >> &dst, const ComputeGraphVar< Buffer2DView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< Buffer3DView< T >> &dst, const ComputeGraphVar< Buffer3DView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, CVarView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, CBufferView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, CBuffer2DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, CBuffer3DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, VarView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, BufferView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, Buffer2DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (T *dst, Buffer3DView< T > src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< T *> &dst, const ComputeGraphVar< BufferView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< T *> &dst, const ComputeGraphVar< Buffer2DView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< T *> &dst, const ComputeGraphVar< VarView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< T *> &dst, const ComputeGraphVar< Buffer3DView< T >> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (VarView< T > dst, const T *src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (BufferView< T > dst, const T *src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer2DView< T > dst, const T *src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (Buffer3DView< T > dst, const T *src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< BufferView< T >> &dst, const ComputeGraphVar< T *> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< Buffer2DView< T >> &dst, const ComputeGraphVar< T *> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< VarView< T >> &dst, const ComputeGraphVar< T *> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | copy (ComputeGraphVar< Buffer3DView< T >> &dst, const ComputeGraphVar< T *> &src) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (VarView< T > buffer, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (BufferView< T > buffer, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (Buffer2DView< T > buffer, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (Buffer3DView< T > buffer, const T &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (ComputeGraphVar< VarView< T >> &buffer, const ComputeGraphVar< T > &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (ComputeGraphVar< BufferView< T >> &buffer, const ComputeGraphVar< T > &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (ComputeGraphVar< Buffer2DView< T >> &buffer, const ComputeGraphVar< T > &val) |
|
template<typename T > |
MUDA_HOST BufferLaunch & | fill (ComputeGraphVar< Buffer3DView< T >> &buffer, const ComputeGraphVar< T > &val) |
|
MUDA_GENERIC | LaunchBase (::cudaStream_t stream) MUDA_NOEXCEPT |
|
MUDA_GENERIC | LaunchBase (cudaStream_t stream) MUDA_NOEXCEPT |
|
BufferLaunch & | push_range (const std::string &name) |
|
BufferLaunch & | pop_range () |
|
BufferLaunch & | kernel_name (std::string_view name) |
|
std::string_view | kernel_name () const |
|
BufferLaunch & | record (cudaEvent_t e, int flag=cudaEventRecordDefault) |
|
BufferLaunch & | record (ComputeGraphVar< cudaEvent_t > &e, const std::vector< ComputeGraphVarBase * > &vars) |
|
BufferLaunch & | record (ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|
BufferLaunch & | when (cudaEvent_t e, int flag=cudaEventWaitDefault) |
|
BufferLaunch & | wait (cudaEvent_t e, int flag=cudaEventWaitDefault) |
|
BufferLaunch & | wait (const ComputeGraphVar< cudaEvent_t > &e, const std::vector< ComputeGraphVarBase * > &vars) |
|
BufferLaunch & | wait (const ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|
BufferLaunch & | wait () |
|
BufferLaunch & | callback (const std::function< void(::cudaStream_t, ::cudaError)> &callback) |
|
Next | next (Next n) |
|
Next | next (Args &&... args) |
|
MUDA_GENERIC | LaunchCore (::cudaStream_t stream) MUDA_NOEXCEPT |
|
void | init_stream (::cudaStream_t s) |
|
void | push_range (const std::string &name) |
|
void | pop_range () |
|
void | record (cudaEvent_t e, int flag=cudaEventRecordDefault) |
|
void | record (ComputeGraphVar< cudaEvent_t > &e, const std::vector< ComputeGraphVarBase *> &vars) |
|
template<typename... ViewT> |
void | record (ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|
void | when (cudaEvent_t e, int flag=cudaEventWaitDefault) |
|
void | wait (cudaEvent_t e, int flag=cudaEventWaitDefault) |
|
void | wait (const ComputeGraphVar< cudaEvent_t > &e, const std::vector< ComputeGraphVarBase *> &vars) |
|
template<typename... ViewT> |
void | wait (const ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|
void | wait () |
|
void | callback (const std::function< void(::cudaStream_t, ::cudaError)> &callback) |
|
template<typename... ViewT> |
MUDA_INLINE void | record (ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|
template<typename... ViewT> |
MUDA_INLINE void | wait (const ComputeGraphVar< cudaEvent_t > &e, ComputeGraphVar< ViewT > &... vars) |
|