Namespace muda
Namespaces
Type | Name |
---|---|
namespace | config |
namespace | details |
namespace | distance |
namespace | eigen |
namespace | lbvh |
namespace | spatial_hash |
Classes
Type | Name |
---|---|
class | BSRMatrixViewT <IsConst, typename Ty, N> |
class | Buffer2DViewT <IsConst, typename T> |
class | Buffer3DViewT <IsConst, typename T> |
class | BufferLaunch |
class | BufferViewT <IsConst, typename T> |
class | CFieldEntryView <typename T, Layout, M, N> |
class | CFieldEntryViewer <typename T, Layout, M, N> |
class | CMatrixMapInfo <typename T, M, N> |
class | COOMatrixViewT <IsConst, typename Ty> |
class | COOVectorViewBase <IsConst, typename T> |
class | ClosureId |
class | ComputeGraph |
class | ComputeGraphBuilder |
class | ComputeGraphCaptureNode |
class | ComputeGraphClosure |
class | ComputeGraphDependency |
class | ComputeGraphGraphvizOptions |
class | ComputeGraphNode <typename NodeT, Type> |
class | ComputeGraphNodeBase |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVar <typename T> |
class | ComputeGraphVarBase |
class | ComputeGraphVarManager |
class | CubWrapper <typename Derive> |
class | Debug |
struct | Default |
class | Dense1DT <IsConst, typename T> |
class | Dense2DBase <IsConst, typename T> |
class | Dense3DBase <IsConst, typename T> |
class | DenseMatrixViewT <IsConst, typename Ty> |
class | DenseMatrixViewerT <IsConst, typename T> |
class | DenseVectorViewT <IsConst, typename T> |
class | DenseVectorViewerT <IsConst, typename T> |
class | DenseViewerT <IsConst, typename T> |
class | DeviceBCOOMatrix <typename T, N> |
class | DeviceBCOOMatrix <typename T, N> |
class | DeviceBCOOVector <typename T, N> |
class | DeviceBCOOVector <typename T, N> |
class | DeviceBSRMatrix <typename Ty, N> |
class | DeviceBuffer <typename T> A std::vector like wrapper of cuda device memory, allows user to: |
class | DeviceBuffer2D <typename T> |
class | DeviceBuffer3D <typename T> |
class | DeviceDenseMatrix <typename Ty> |
class | DeviceDenseVector <typename T> |
class | DeviceDoubletVector <typename T, N> |
class | DeviceHistogram |
class | DeviceMergeSort |
class | DevicePartition |
class | DeviceRadixSort |
class | DeviceReduce |
class | DeviceRunLengthEncode |
class | DeviceScan |
class | DeviceSegmentedRadixSort |
class | DeviceSegmentedReduce |
class | DeviceSegmentedSort |
class | DeviceSelect |
class | DeviceSpmv |
class | DeviceTripletMatrix <typename T, N> |
class | DeviceVar <typename T> |
class | DeviceVector <typename T> |
class | DoubletVectorViewT <IsConst, typename T, N> |
class | DoubletVectorViewT <IsConst, typename T, N> |
class | DoubletVectorViewerT <IsConst, typename T, N> |
class | Empty |
class | Event RAII wrapper for cudaEvent. |
class | EventRecordNode |
class | EventWaitNode |
class | Extent2D |
class | Extent3D |
class | Field |
class | FieldBuildOptions |
class | FieldBuilder <Layout> |
class | FieldEntry <typename T, Layout, M, N> |
class | FieldEntryBase |
class | FieldEntryBaseData |
class | FieldEntryCore |
class | FieldEntryLaunch |
class | FieldEntryLayoutInfo |
class | FieldEntryView <typename T, Layout, M, N> |
class | FieldEntryViewBase <IsConst, typename T, Layout, M, N> |
class | FieldEntryViewCore <IsConst, typename T, Layout, M, N> |
class | FieldEntryViewer <typename T, Layout, M, N> |
class | FieldEntryViewerBase <IsConst, typename T, Layout, M, N> |
class | FieldEntryViewerCore <IsConst, typename T, Layout, M, N> |
class | Flags <typename BitType> |
class | Graph |
class | GraphExec |
class | GraphLaunch |
class | GraphNode |
class | GraphViewer |
class | HostCall |
class | HostDeviceConfig <typename T> |
class | HostDeviceConfigView <typename T> |
class | HostNode |
class | HostNodeParms <typename T> |
class | HostVector <typename T> |
class | IdWithType <typename T> |
class | Kernel <typename F> |
class | KernelLabel |
class | KernelNode |
class | KernelNodeParms <typename U> |
class | Launch A wrapper of raw cuda kernel launch in muda style , removing the <<<>>> usage, for better intellisense support. |
class | LaunchBase <typename T> |
class | LaunchCore |
class | LinearSystemAlgorithm |
class | LinearSystemContext |
class | LinearSystemContextCreateInfo |
class | LinearSystemHandles |
class | LinearSystemSolveReorder |
class | LinearSystemSolveTolerance |
class | LogProxy |
class | Logger |
class | LoggerDataContainer |
class | LoggerMetaData |
class | LoggerViewer |
class | MatrixFormatConverter |
class | MatrixMapInfo <typename T, M, N> For MapMatrix e.g. Eigen::Map< ... > |
class | MemcpyNode |
class | Memory |
class | MemsetNode |
class | NDReshaper |
class | NodeId |
class | NodeParms |
class | Offset2D |
class | Offset3D |
class | ParallelFor a frequently used parallel for loop, DynamicBlockDim andGridStrideLoop strategy are provided, and can be switched seamlessly to each other. |
class | ParallelForDetails |
class | Profile |
class | RangeName |
class | Stream RAII wrapper for cudaStream. |
class | SubField |
class | SubFieldImpl <Layout> |
class | SubFieldImpl <Layout> |
class | SubFieldImpl <Layout> |
class | SubFieldImpl <Layout> |
class | SubFieldInterface |
struct | Tag <typename T> |
class | TripletMatrixViewT <IsConst, typename Ty, N> |
class | TripletMatrixViewT <IsConst, typename Ty, N> |
class | TripletMatrixViewerT <IsConst, typename T, N> |
class | VarId |
class | VarViewT <IsConst, typename T> |
class | ViewBase <IsConst_> |
class | ViewerBase <IsConst_> |
struct | always_false <typename T> |
struct | always_true <typename T> |
class | cuda_error <typename T> |
class | exception |
struct | force_trivial <typename T> |
struct | force_trivially_constructible <typename T> |
struct | force_trivially_copy_assignable <typename T> |
struct | force_trivially_copy_constructible <typename T> |
struct | force_trivially_destructible <typename T> |
class | invalid_argument |
struct | is_cuda_arch |
class | logic_error |
class | not_implemented |
class | out_of_range |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_only_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
struct | read_write_view <typename T> |
class | runtime_error |
Public Types
Public Attributes
Type | Name |
---|---|
bool | COMPUTE_GRAPH_ON = = MUDA_COMPUTE_GRAPH_ON |
bool | DEBUG_VIEWER = = config::on(true) |
bool | DEFAULT_ASYNC_ALLOC_FREE = = false |
int | HEAVY_WORKLOAD_BLOCK_SIZE = = 64 |
int | LIGHT_WORKLOAD_BLOCK_SIZE = = 256 |
int | MIDDLE_WORKLOAD_BLOCK_SIZE = = 128 |
bool | RUNTIME_CHECK_ON = = MUDA_CHECK_ON |
bool | TRAP_ON_ERROR = = config::on(true) |
bool | always_false_v = = always_false<T>::value |
bool | force_trivial_v = = force_trivial<T>::value |
bool | force_trivially_constructible_v = = force_trivially_constructible<T>::value |
bool | force_trivially_copy_assignable_v = = |
force_trivially_copy_assignable<T>::value |
|
bool | force_trivially_copy_constructible_v = = |
force_trivially_copy_constructible<T>::value |
|
bool | force_trivially_destructible_v = = force_trivially_destructible<T>::value |
bool | is_cuda_arch_v = = is_cuda_arch::value |
bool | is_read_only_view_v = = std::is_same_v<T, read_only_view_t<T>> |
bool | is_read_write_view_v = = std::is_same_v<T, read_write_view_t<T>> |
bool | is_trivial_v = = std::is_trivial_v<T> || force_trivial_v<T> |
bool | is_trivially_constructible_v = = |
std::is_trivially_constructible_v<T> || force_trivially_constructible_v<T> |
|
bool | is_trivially_copy_assignable_v = = |
std::is_trivially_copy_assignable_v<T> || force_trivially_copy_assignable_v<T> |
|
bool | is_trivially_copy_constructible_v = = |
std::is_trivially_copy_constructible_v<T> || force_trivially_copy_constructible_v<T> |
|
bool | is_trivially_destructible_v = = |
std::is_trivially_destructible_v<T> || force_trivially_destructible_v<T> |
|
bool | is_uniform_view_v = = is_read_only_view_v<T> && is_read_write_view_v<T> |
Public Functions
Type | Name |
---|---|
MUDA_DEFINE_ARITHMATIC_OPERATOR (+) |
|
MUDA_DEFINE_ARITHMATIC_OPERATOR (-) |
|
MUDA_DEFINE_COMPARISON_OPERATOR (<=) |
|
MUDA_DEFINE_COMPARISON_OPERATOR () |
|
MUDA_INLINE MUDA_GENERIC Extent2D | as_extent (const Offset2D & offset) |
MUDA_INLINE MUDA_GENERIC Extent3D | as_extent (const Offset3D & offset) |
MUDA_INLINE MUDA_GENERIC Offset2D | as_offset (const Extent2D & extent) |
MUDA_INLINE MUDA_GENERIC Offset3D | as_offset (const Extent3D & extent) |
__forceinline__ __device__ T | atomic_add (T * address, T val) |
__forceinline__ __device__ T | atomic_and (T * address, T val) |
__forceinline__ __device__ T | atomic_cas (T * address, T compare, T val) |
__forceinline__ __device__ T | atomic_exch (T * address, T val) |
__forceinline__ __device__ T | atomic_max (T * address, T val) |
__forceinline__ __device__ T | atomic_min (T * address, T val) |
__forceinline__ __device__ T | atomic_or (T * address, T val) |
__forceinline__ __device__ T | atomic_sub (T * address, T val) |
__forceinline__ __device__ T | atomic_xor (T * address, T val) |
MUDA_INLINE MUDA_GENERIC dim3 | block_dim () |
MUDA_INLINE MUDA_GENERIC dim3 | block_idx () |
MUDA_INLINE MUDA_GENERIC void | brkpt () |
__host__ __device__ void | check (T result, char const *const func, const char *const file, int const line) |
dim3 | cube (int x) |
cublasOperation_t | cublas_trans_operation (bool b) |
cudaDataType_t | cuda_data_type () |
cusparseIndexType_t | cusparse_index_type () |
std::string_view | enum_name (ComputeGraphNodeType t) |
MUDA_INLINE MUDA_GENERIC dim3 | grid_dim () |
MUDA_GLOBAL void | grid_stride_loop_kernel (ParallelForCallable< F > f) |
MUDA_INLINE MUDA_GENERIC decltype(auto) | invoke (FHost && host, FDevice && device) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense (const T * data) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_1d (const T * data, int dimx) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_1d (const T(&) data) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_2d (const T * data, const int2 & dim) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_2d (const T * data, int dimx, int dimy) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_3d (const T * data, const int3 & dim) |
MUDA_INLINE MUDA_GENERIC auto | make_cdense_3d (const T * data, int dimx, int dimy, int dimz) |
MUDA_INLINE MUDA_GENERIC auto | make_dense (T * data) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_1d (T * data, int dimx) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_1d (T(&) data) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_2d (T * data, const int2 & dim) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_2d (T * data, int dimx, int dimy) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_3d (T * data, const int3 & dim) |
MUDA_INLINE MUDA_GENERIC auto | make_dense_3d (T * data, int dimx, int dimy, int dimz) |
MUDA_INLINE MUDA_GENERIC Extent2D | max (const Extent2D & lhs, const Extent2D & rhs) |
MUDA_INLINE MUDA_GENERIC Extent3D | max (const Extent3D & lhs, const Extent3D & rhs) |
MUDA_INLINE MUDA_GENERIC Offset2D | min (const Offset2D & lhs, const Offset2D & rhs) |
MUDA_INLINE MUDA_GENERIC Offset3D | min (const Offset3D & lhs, const Offset3D & rhs) |
MUDA_INLINE MUDA_GENERIC const char * | mudaCudaGetErrorEnum (cusolverStatus_t error) |
Empty | on (::cudaStream_t stream) |
Empty | on () |
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags< BitType > | operator& (BitType bit, Flags< BitType > const & flags) |
MUDA_INLINE MUDA_GENERIC int2 | operator+ (const int2 & a, const int2 & b) |
bool | operator< (ComputeGraphVarUsage lhs, ComputeGraphVarUsage rhs) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::Matrix< T, M, N > & val) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::Map< Eigen::Matrix< T, M, N >, MapOptions, StrideType > & val) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::Map< const Eigen::Matrix< T, M, N >, MapOptions, StrideType > & val) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::MatrixX< T > & val) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::VectorX< T > & val) |
MUDA_DEVICE LogProxy | operator<< (LogProxy o, const Eigen::RowVectorX< T > & val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, char1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, char2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, char3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, char4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, short1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, short2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, short3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, short4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, int1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, int2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, int3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, int4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, long1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, long2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, long3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, long4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, longlong1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, longlong2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, longlong3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, longlong4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uchar1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uchar2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uchar3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uchar4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uint1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uint2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uint3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, uint4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulong1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulong2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulong3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulong4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulonglong1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulonglong2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulonglong3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, ulonglong4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, float1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, float2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, float3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, float4 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, double1 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, double2 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, double3 val) |
MUDA_INLINE MUDA_DEVICE LogProxy | operator<< (LogProxy proxy, double4 val) |
bool | operator<= (ComputeGraphVarUsage lhs, ComputeGraphVarUsage rhs) |
bool | operator== (ComputeGraphVarUsage lhs, ComputeGraphVarUsage rhs) |
bool | operator> (ComputeGraphVarUsage lhs, ComputeGraphVarUsage rhs) |
bool | operator>= (ComputeGraphVarUsage lhs, ComputeGraphVarUsage rhs) |
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags< BitType > | operator^ (BitType bit, Flags< BitType > const & flags) |
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags< BitType > | operator| (BitType bit, Flags< BitType > const & flags) |
MUDA_GLOBAL void | parallel_for_kernel (ParallelForCallable< F > f) |
MUDA_INLINE MUDA_GENERIC void | print (const char *const fmt, Args &&... arg) |
MUDA_INLINE MUDA_GENERIC const T & | print_check (const T & t) |
MUDA_INLINE MUDA_GENERIC auto | print_check (const char * t) |
MUDA_INLINE MUDA_GENERIC OutType | print_convert (const InType & v) |
MUDA_INLINE MUDA_GENERIC auto | print_convert (const char * v) |
MUDA_INLINE MUDA_GENERIC const T & | print_convert (const Dense< T > & v) |
MUDA_INLINE MUDA_GENERIC const T & | print_convert (const CDense< T > & v) |
MUDA_HOST double | profile_host (F && f) |
MUDA_GENERIC MUDA_INLINE constexpr T * | remove_const (const T * ptr) noexcept |
MUDA_GENERIC MUDA_INLINE constexpr T & | remove_const (const T & ref) noexcept |
dim3 | square (int x) |
MUDA_INLINE MUDA_GENERIC dim3 | thread_idx () |
MUDA_INLINE MUDA_GENERIC void | trap () |
void | wait_device () |
void | wait_event (cudaEvent_t event) |
void | wait_stream (::cudaStream_t stream) |
Public Types Documentation
typedef BCOOMatrixView
typedef BCOOMatrixViewer
typedef BCOOVectorView
typedef BCOOVectorViewer
typedef BSRMatrixView
typedef Buffer2DView
typedef Buffer3DView
typedef BufferView
typedef CBCOOMatrixView
typedef CBCOOMatrixViewer
typedef CBCOOVectorView
typedef CBCOOVectorViewer
typedef CBSRMatrixView
typedef CBuffer2DView
typedef CBuffer3DView
typedef CBufferView
typedef CCOOMatrixView
typedef CCOOMatrixViewer
typedef CCOOVectorView
typedef CCOOVectorViewer
typedef CCSRMatrixView
typedef CDense
typedef CDense1D
typedef CDense2D
typedef CDense3D
typedef CDenseMatrixView
typedef CDenseMatrixViewer
typedef CDenseVectorView
typedef CDenseVectorViewer
typedef CDoubletVectorView
typedef CDoubletVectorViewer
typedef COOMatrixView
typedef COOMatrixViewer
typedef COOVectorView
typedef COOVectorViewer
typedef CSRMatrixView
typedef CSRMatrixViewT
typedef CTripletMatrixView
typedef CTripletMatrixViewer
typedef CVarView
typedef ComputeGraphEventRecordNode
using muda::ComputeGraphEventRecordNode =
ComputeGraphNode<EventRecordNode, ComputeGraphNodeType::EventRecordNode>;
typedef ComputeGraphEventWaitNode
using muda::ComputeGraphEventWaitNode =
ComputeGraphNode<EventWaitNode, ComputeGraphNodeType::EventWaitNode>;
enum ComputeGraphFlag
typedef ComputeGraphKernelNode
using muda::ComputeGraphKernelNode =
ComputeGraphNode<KernelNode, ComputeGraphNodeType::KernelNode>;
typedef ComputeGraphMemcpyNode
using muda::ComputeGraphMemcpyNode =
ComputeGraphNode<MemcpyNode, ComputeGraphNodeType::MemcpyNode>;
typedef ComputeGraphMemsetNode
using muda::ComputeGraphMemsetNode =
ComputeGraphNode<MemsetNode, ComputeGraphNodeType::MemsetNode>;
enum ComputeGraphNodeType
enum muda::ComputeGraphNodeType {
None,
KernelNode,
MemcpyNode,
MemsetNode,
CaptureNode,
EventRecordNode,
EventWaitNode,
Max
};
enum ComputeGraphPhase
enum ComputeGraphVarUsage
typedef DefaultTag
typedef Dense
typedef Dense1D
typedef Dense2D
typedef Dense3D
typedef DenseMatrixView
typedef DenseMatrixViewer
typedef DenseVectorView
typedef DenseVectorViewer
typedef DeviceCOOMatrix
typedef DeviceCOOVector
typedef DeviceCSRMatrix
typedef DoubletVectorView
typedef DoubletVectorViewer
enum FieldEntryLayout
enum FieldEntryType
enum GraphInstantiateFlagBit
enum muda::GraphInstantiateFlagBit {
FreeOnLaunch = CUgraphInstantiate_flags::CUDA_GRAPH_INSTANTIATE_FLAG_AUTO_FREE_ON_LAUNCH,
Upload = 2,
DeviceLaunch = 4,
UseNodePriority = 8
};
typedef I32IdWithType
typedef I64IdWithType
typedef KeyValuePair
enum LinearSystemReorderMethod
enum LoggerBasicType
enum muda::LoggerBasicType {
None,
Int8,
Int16,
Int,
Int32 = Int,
Int64,
Long,
LongLong,
UInt8,
UInt16,
UInt,
UInt32 = UInt,
UInt64,
ULong,
ULongLong,
Float,
Double,
String,
FmtString,
Object
};
typedef LoggerFmtArg
enum ParallelForType
typedef TripletMatrixView
typedef TripletMatrixViewer
typedef U32IdWithType
typedef U64IdWithType
typedef VarView
typedef raw_type_t
typedef read_only_view_t
typedef read_write_view_t
typedef span
Public Attributes Documentation
variable COMPUTE_GRAPH_ON
variable DEBUG_VIEWER
variable DEFAULT_ASYNC_ALLOC_FREE
variable HEAVY_WORKLOAD_BLOCK_SIZE
variable LIGHT_WORKLOAD_BLOCK_SIZE
variable MIDDLE_WORKLOAD_BLOCK_SIZE
variable RUNTIME_CHECK_ON
variable TRAP_ON_ERROR
variable always_false_v
variable force_trivial_v
variable force_trivially_constructible_v
variable force_trivially_copy_assignable_v
variable force_trivially_copy_constructible_v
variable force_trivially_destructible_v
variable is_cuda_arch_v
variable is_read_only_view_v
variable is_read_write_view_v
variable is_trivial_v
variable is_trivially_constructible_v
variable is_trivially_copy_assignable_v
variable is_trivially_copy_constructible_v
variable is_trivially_destructible_v
variable is_uniform_view_v
Public Functions Documentation
function MUDA_DEFINE_ARITHMATIC_OPERATOR
function MUDA_DEFINE_ARITHMATIC_OPERATOR
function MUDA_DEFINE_COMPARISON_OPERATOR
function MUDA_DEFINE_COMPARISON_OPERATOR
function as_extent
function as_extent
function as_offset
function as_offset
function atomic_add
function atomic_and
function atomic_cas
template<typename T>
__forceinline__ __device__ T muda::atomic_cas (
T * address,
T compare,
T val
)
function atomic_exch
function atomic_max
function atomic_min
function atomic_or
function atomic_sub
function atomic_xor
function block_dim
function block_idx
function brkpt
function check
template<typename T>
inline __host__ __device__ void muda::check (
T result,
char const *const func,
const char *const file,
int const line
)
function cube
function cublas_trans_operation
function cuda_data_type
function cusparse_index_type
function enum_name
function grid_dim
function grid_stride_loop_kernel
template<typename F, typename UserTag>
MUDA_GLOBAL void muda::grid_stride_loop_kernel (
ParallelForCallable< F > f
)
function invoke
template<typename FHost, typename FDevice>
MUDA_INLINE MUDA_GENERIC decltype(auto) muda::invoke (
FHost && host,
FDevice && device
)
function make_cdense
function make_cdense_1d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_cdense_1d (
const T * data,
int dimx
)
function make_cdense_1d
function make_cdense_2d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_cdense_2d (
const T * data,
const int2 & dim
)
function make_cdense_2d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_cdense_2d (
const T * data,
int dimx,
int dimy
)
function make_cdense_3d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_cdense_3d (
const T * data,
const int3 & dim
)
function make_cdense_3d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_cdense_3d (
const T * data,
int dimx,
int dimy,
int dimz
)
function make_dense
function make_dense_1d
function make_dense_1d
function make_dense_2d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_dense_2d (
T * data,
const int2 & dim
)
function make_dense_2d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_dense_2d (
T * data,
int dimx,
int dimy
)
function make_dense_3d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_dense_3d (
T * data,
const int3 & dim
)
function make_dense_3d
template<typename T>
MUDA_INLINE MUDA_GENERIC auto muda::make_dense_3d (
T * data,
int dimx,
int dimy,
int dimz
)
function max
function max
function min
function min
function mudaCudaGetErrorEnum
function on
function on
function operator&
template<typename BitType>
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags < BitType > muda::operator& (
BitType bit,
Flags < BitType > const & flags
)
function operator+
function operator<
function operator<<
template<typename T, int M, int N>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::Matrix< T, M, N > & val
)
function operator<<
template<typename T, int M, int N, int MapOptions, typename StrideType>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::Map< Eigen::Matrix< T, M, N >, MapOptions, StrideType > & val
)
function operator<<
template<typename T, int M, int N, int MapOptions, typename StrideType>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::Map< const Eigen::Matrix< T, M, N >, MapOptions, StrideType > & val
)
function operator<<
template<typename T>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::MatrixX< T > & val
)
function operator<<
template<typename T>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::VectorX< T > & val
)
function operator<<
template<typename T>
MUDA_DEVICE LogProxy muda::operator<< (
LogProxy o,
const Eigen::RowVectorX< T > & val
)
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<<
function operator<=
function operator==
function operator>
function operator>=
function operator^
template<typename BitType>
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags < BitType > muda::operator^ (
BitType bit,
Flags < BitType > const & flags
)
function operator|
template<typename BitType>
MUDA_INLINE MUDA_GENERIC MUDA_CONSTEXPR Flags < BitType > muda::operator| (
BitType bit,
Flags < BitType > const & flags
)
function parallel_for_kernel
template<typename F, typename UserTag>
MUDA_GLOBAL void muda::parallel_for_kernel (
ParallelForCallable< F > f
)
function print
template<typename... Args>
MUDA_INLINE MUDA_GENERIC void muda::print (
const char *const fmt,
Args &&... arg
)
function print_check
function print_check
function print_convert
template<typename InType, typename OutType>
MUDA_INLINE MUDA_GENERIC OutType muda::print_convert (
const InType & v
)
function print_convert
function print_convert
template<typename T>
MUDA_INLINE MUDA_GENERIC const T & muda::print_convert (
const Dense < T > & v
)
function print_convert
template<typename T>
MUDA_INLINE MUDA_GENERIC const T & muda::print_convert (
const CDense < T > & v
)
function profile_host
function remove_const
template<typename T>
MUDA_GENERIC MUDA_INLINE constexpr T * muda::remove_const (
const T * ptr
) noexcept
function remove_const
template<typename T>
MUDA_GENERIC MUDA_INLINE constexpr T & muda::remove_const (
const T & ref
) noexcept
function square
function thread_idx
function trap
function wait_device
function wait_event
function wait_stream
The documentation for this class was generated from the following file src/muda/assert.h