tvm
Class Hierarchy

Go to the graphical class hierarchy

This inheritance list is sorted roughly, but not completely, alphabetically:
[detail level 123456]
 Ctvm::runtime::memory::Allocator
 Ctvm::arith::AnalyzerAnalyzer that contains bunch of sub-analyzers
 Ctvm::runtime::relax_vm::Instruction::ArgThe auxiliary data structure for instruction argument
 Ctvm::runtime::metadata::ArrayAccessor< C, Ref >A span-like class which permits access to Array fields with complex elements. These array fields should be accessed from C++ using the Metadata wrapper classes. This class lazily instantiates those wrappers as they are accessed
 Ctvm::runtime::metadata::ArrayAccessor< const char *, ::tvm::runtime::String >A specialization of ArrayAccessor for String. This class is needed because the String constructor signature is different from the typical Metadata subclass
 Ctvm::runtime::SimpleObjAllocator::ArrayHandler< ArrayType, ElemType >
 Ctvm::runtime::metadata::ArrayIterator< C, Ref >An iterator implementation that lazily instantiates the C++ wrapping Metadata class
 Ctvm::detail::AttrDocEntry
 Ctvm::detail::AttrDocVisitor
 Ctvm::detail::AttrExistVisitor
 Ctvm::detail::AttrInitEntry< T >
 Ctvm::detail::AttrInitVisitor< FFind >
 Ctvm::detail::AttrNonDefaultVisitor
 Ctvm::detail::AttrNopEntry
 Ctvm::detail::AttrNormalVisitor
 Ctvm::AttrRegistry< typename, typename >
 Ctvm::AttrRegistryMap< KeyType, ValueType >Map<Key, ValueType> used to store meta-data
 Ctvm::AttrRegistryMap< Op, ValueType >
 Ctvm::AttrRegistryMap< TargetKind, ValueType >
 Ctvm::AttrRegistryMapContainerMap< KeyType >Generic attribute map
 Ctvm::detail::AttrsSEqualVisitor
 Ctvm::detail::AttrsSHashVisitor
 Ctvm::detail::AttrTriggerNonDefaultEntry< T >
 Ctvm::AttrVisitorVisitor class to get the attributes of an AST/IR node. The content is going to be called for each field
 Ctvm::relax::distributed::AxisTensor axis
 Ctvm::relax::distributed::AxisGroupGraphA graph whose nodes are tensor axes, and the edge means some information can be propagated through the two axes. Although it only does sharding propagation, this data structure can be extended to perform all kinds of propagation that happens on tensor axes
 Ctvm::relax::distributed::AxisGroupHash
 Ctvm::relax::distributed::AxisHash
 Ctvm::relax::distributed::AxisShardingSpecEqual
 Ctvm::relax::distributed::AxisShardingSpecHash
 Ctvm::BaseValueEqualEquality definition of base value class
 Ctvm::BaseValueHashHash definition of base value classes
 Ctvm::tir::BlockInfoThe information about a TensorIR block, it contains two categories of information 1) Info on the block scope rooted at a specific block, including dependency tracking, flags indicating if the scope is a stage pipeline, etc. 2) Info on the block itself, including if the block has a quasi-affine binding, if the regions it reads are completely covered by their producers, etc
 Cstd::bool_constant
 Ctvm::runtime::memory::Buffer
 Ctvm::tir::BufferAxisHash
 Ctvm::ir::CalleeCollector
 Ctvm::runtime::profiling::CallFrame
 Ctvm::arith::CanonicalSimplifierCanonical-form based simplifier
 Ctvm::arith::ConstIntBoundAnalyzerAnalyzer to get constant integer bound over expression
 Ctvm::arith::ConstraintContextConstraint context
 Ctvm::runtime::NDArray::ContainerBaseThe container base structure contains all the fields except for the Object header
 Ctvm::runtime::DataTypeRuntime primitive data type
 Ctvm::runtime::DeviceAPITVM Runtime Device API, abstracts the device specific interface for memory management
 Ctvm::relax::DFPatternFunctor< FType >A dynamical functor that dispatches on in the first DFPattern argument
 Ctvm::relay::DFPatternFunctor< FType >A dynamical functor that dispatches on in the first DFPattern argument
 Ctvm::relax::DFPatternFunctor< R(const DFPattern &n, Args...)>
 Ctvm::relay::DFPatternFunctor< R(const DFPattern &n, Args...)>
 Ctvm::relax::DFPatternFunctor< void(const DFPattern &)>
 Ctvm::DiagnosticBuilderA wrapper around std::stringstream to build a diagnostic
 Ctvm::relax::BlockBuilder::DisableOperatorSpecificNormalizationForTVMScriptA marker struct to disable FNormalize
 Ctvm::runtime::DiscoChannelA bi-directional channel for controler-worker communication. This channel is primarily used to transfer control messages but not data
 Ctvm::runtime::DiscoWorkerA worker in Disco. It takes a channel to communication with the controler. The worker can be run in a separate thread or process as long as the channel supports bi-directional communication in-between
 Ctvm::topi::EinsumEquation
 CError
 Ctvm::relay::ErrorBuilderA wrapper around std::stringstream to build error. include/tvm/ir/type.h Can be consumed by CompileError to construct an error
 Ctvm::relay::ErrorReporterAn abstraction around how errors are stored and reported. Designed to be opaque to users, so we can support a robust and simpler error reporting mode, as well as a more complex mode
 Ctvm::relay::ExecutorRegEntryHelper structure to register Executors
 Ctvm::tir::ExprDeepEqualCompare two expressions recursively and check if they are equal to each other without var remapping
 Ctvm::relax::ExprFunctor< FType >A dynamical functor that dispatches on in the first Expr argument. You can use this as a more powerful Visitor, since it allows you to define function signatures of Visit Function
 Ctvm::relay::ExprFunctor< FType >A dynamical functor that dispatches on in the first Expr argument. You can use this as a more powerful Visitor, since it allows you to define function signatures of Visit Function
 Ctvm::tir::ExprFunctor< FType >A dynamical functor that dispatches on in the first Expr argument. You can use this as a more powerful Visitor, since it allows you to define function signatures of Visit Function
 Ctvm::relax::ExprFunctor< Expr(const Expr &)>
 Ctvm::tir::ExprFunctor< PrimExpr(const PrimExpr &)>
 Ctvm::relax::ExprFunctor< R(const Expr &n, Args...)>
 Ctvm::relay::ExprFunctor< R(const Expr &n, Args...)>
 Ctvm::tir::ExprFunctor< R(const PrimExpr &n, Args...)>
 Ctvm::relax::ExprFunctor< void(const Expr &)>
 Ctvm::relay::ExprFunctor< void(const Expr &n)>
 Ctvm::tir::ExprFunctor< void(const PrimExpr &)>
 Ctvm::relay::ExprRewriterA non-iterating Expression Rewriter
 Ctvm::runtime::PackedFuncObj::Extractor< TPackedFuncSubObj >Internal struct for extracting the callable method from callable type
 Cstd::false_type
 Ctvm::relay::FeatureSetA finite set of Feature
 Ctvm::runtime::relax_vm::NDArrayCacheMetadata::FileRecordEach shard of NDArray cache, which by default, is named as "params_shard_x.bin"
 Ctvm::tir::SeqStmt::FlattenerHelper class to flatten sequence of arguments into Array
 CGlobalVar
 Ctvm::transform::GlobalVarReplacer
 Ctvm::tir::usmp::algo::GreedyBaseThis is the base class for Greedy Algorithms where the sorting is specialized in the extended classes based on the greedy criteria
 Ctvm::runtime::SimpleObjAllocator::Handler< T >
 Ctvm::SEqualReducer::HandlerInternal handler that defines custom behaviors.
 Ctvm::SHashReducer::HandlerInternal handler that defines custom behaviors
 Cdmlc::serializer::Handler< DLDataType >
 Cdmlc::serializer::Handler< DLDevice >
 Ctvm::detail::ImplSEqualReduce< T, bool >
 Ctvm::detail::ImplSEqualReduce< T, true >
 Ctvm::detail::ImplSHashReduce< T, bool >
 Ctvm::detail::ImplSHashReduce< T, true >
 Ctvm::detail::ImplVisitAttrs< T, bool >
 Ctvm::detail::ImplVisitAttrs< T, true >
 Ctvm::runtime::InplaceArrayBase< ArrayType, ElemType >Base template for classes with array like memory layout
 Ctvm::runtime::InplaceArrayBase< ADTObj, ObjectRef >
 Ctvm::runtime::InplaceArrayBase< ArrayNode, ObjectRef >
 Ctvm::runtime::InplaceArrayBase< SmallMapNode, MapNode::KVType >
 Ctvm::runtime::relax_vm::InstructionA single virtual machine instruction
 Ctvm::runtime::vm::InstructionA single virtual machine instruction
 Ctvm::tir::InstructionKindRegEntryAn entry in the registry of InstructionKind
 Ctvm::arith::IntSetAnalyzerInteger set analyzer
 Ctvm::script::printer::IRDocsifierFunctor< R, Args >Dynamic dispatch functor based on ObjectPath
 Ctvm::runtime::IterAdapter< Converter, TIter >Iterator adapter that adapts TIter to return another type
 Ctvm::runtime::Map< K, V, typename, typename >::iteratorIterator of the hash map
 Ctvm::runtime::MapNode::iterator
 Ctvm::support::Span< T, W >::iterator_base< W1 >
 Ctvm::auto_scheduler::AttachMapNode::IterKeyHash
 Ctvm::tir::LayoutAxis
 Ctvm::support::LinearCongruentialEngineThis linear congruential engine is a drop-in replacement for std::minstd_rand. It strictly corresponds to std::minstd_rand and is designed to be platform-independent
 Ctvm::tir::MemCpyDetailsHelper struct for return value of IdentifyMemCpy
 Ctvm::runtime::memory::MemoryManager
 Ctvm::arith::ModularSetAnalyzerAnalyzer to get modular information over expression
 Ctvm::script::ir_builder::details::Namer
 Ctvm::runtime::relax_vm::NDArrayCacheMetadataMetadata for NDArray cache, which by default, is named as "ndarray-cache.json"
 Ctvm::NDArrayContainerTrait
 Ctvm::NodeFunctor< FType >A dynamically dispatched functor on the type of the first argument
 Ctvm::NodeFunctor< R(const ObjectRef &n, Args...)>
 Ctvm::runtime::NullOptTypeHelper to represent nullptr for optional
 Ctvm::runtime::NVTXScopedRangeA class to create a NVTX range. No-op if TVM is not built against NVTX
 Ctvm::runtime::ObjAllocatorBase< Derived >Base class of object allocators that implements make. Use curiously recurring template pattern
 Ctvm::runtime::ObjAllocatorBase< SimpleObjAllocator >
 Ctvm::runtime::ObjectBase class of all object containers
 Ctvm::runtime::ObjectEqualString-aware ObjectRef hash functor
 Ctvm::runtime::ObjectHashString-aware ObjectRef equal functor
 Ctvm::runtime::ObjectPtr< T >A custom smart pointer for Object
 Ctvm::runtime::ObjectPtr< tvm::AttrFieldInfoNode >
 Ctvm::runtime::ObjectPtr< tvm::runtime::Object >
 Ctvm::runtime::ObjectPtr< tvm::runtime::relax_vm::Executable >
 Ctvm::runtime::ObjectPtr< tvm::runtime::vm::Executable >
 Ctvm::runtime::ObjectPtrEqualObjectRef equal functor
 Ctvm::runtime::ObjectPtrHashObjectRef hash functor
 Ctvm::runtime::ObjectRefBase class of all object reference
 Ctvm::runtime::ObjectTypeChecker< T >Type traits for runtime type check during FFI conversion
 Ctvm::runtime::ObjectTypeChecker< Array< T > >
 Ctvm::runtime::ObjectTypeChecker< Map< K, V > >
 Ctvm::runtime::ObjectTypeChecker< Variant< FirstVariant, RemainingVariants... > >
 Ctvm::runtime::ObjectTypeChecker< Variant< OnlyVariant > >
 Ctvm::OpRegEntryHelper structure to register operators
 Ctvm::runtime::PackedFuncValueConverter< TObjectRef >Type trait to specify special value conversion rules from TVMArgValue and TVMRetValue
 Ctvm::runtime::PackedFuncValueConverter< Array< T > >
 Ctvm::runtime::PackedFuncValueConverter< Map< T, U > >
 Ctvm::runtime::PackedFuncValueConverter< Optional< T > >
 Ctvm::runtime::PackedFuncValueConverter< PrimExpr >
 Ctvm::runtime::PackedFuncValueConverter< runtime::Int >
 Ctvm::runtime::PackedFuncValueConverter< tvm::Bool >
 Ctvm::runtime::PackedFuncValueConverter< tvm::FloatImm >
 Ctvm::runtime::PackedFuncValueConverter< tvm::Integer >
 Ctvm::runtime::PackedFuncValueConverter< tvm::IntImm >
 Ctvm::runtime::PackedFuncValueConverter< tvm::tir::StringImm >
 Ctvm::runtime::PackedFuncValueConverter< Variant< VariantTypes... > >
 Ctvm::runtime::PackedFuncValueConverter<::tvm::runtime::String >
 Ctvm::relax::PairConsConstraint of a DFPattern edge (producer -> consumer) in graph-level matching
 Ctvm::runtime::relax_vm::NDArrayCacheMetadata::FileRecord::ParamRecordMetadata of each parameter
 Ctvm::relay::PatternFunctor< FType >A dynamical functor on ADT patterns that dispatches on its first argument. You can use this as a more powerful visitor, since it allows you to define the types of further arguments to VisitPattern
 Ctvm::relay::PatternFunctor< Pattern(const Pattern &)>
 Ctvm::relay::PatternFunctor< R(const Pattern &n, Args...)>
 Ctvm::relay::PatternFunctor< void(const Pattern &n)>
 Ctvm::runtime::profiling::Profiler
 Ctvm::ReflectionVTableVirtual function table to support IR/AST node reflection
 Ctvm::ReflectionVTable::RegistryRegistry of a reflection table
 Ctvm::runtime::RegistryRegistry for global function
 Ctvm::ReprLegacyPrinterLegacy behavior of ReprPrinter
 Ctvm::ReprPrinterA printer class to print the AST/IR nodes
 Ctvm::runtime::ReverseIterAdapter< Converter, TIter >Iterator adapter that adapts TIter to return another type
 Ctvm::arith::RewriteSimplifierRewrite-rule based simplifier
 Ctvm::relay::RuntimeRegEntryHelper structure to register Runtimes
 Ctvm::te::ScheduleContextContext helper to collect debug information of Schedule
 Ctvm::meta_schedule::ScopedTimer
 Ctvm::auto_scheduler::SearchPolicyKeyAttribute keys of ops used for SearchPolicy
 Ctvm::detail::SelectSEqualReduce< T, TraitName, bool >
 Ctvm::detail::SelectSEqualReduce< T, TraitName, false >
 Ctvm::detail::SelectSHashReduce< T, TraitName, bool >
 Ctvm::detail::SelectSHashReduce< T, TraitName, false >
 Ctvm::detail::SelectVisitAttrs< T, TraitName, bool >
 Ctvm::detail::SelectVisitAttrs< T, TraitName, false >
 Ctvm::SEqualReducerA Reducer class to reduce the structural equality result of two objects
 Ctvm::SHashReducerA Reducer class to reduce the structural hash value
 Ctvm::SIBuilderSource Information Builder, SIBuilder provides helper APIs for filling spans, particularly useful for one-to-many, many-to-one and many-to-many IR transformations
 Ctvm::runtime::SignaturePrinter< TSignature >
 Ctvm::te::Tensor::SliceData structure to represent a slice that fixes first k coordinates. This is used to enable syntax sugar of Tensor[x][y][z] to get the element
 Ctvm::support::Span< T, W >A partial implementation of the C++20 std::span
 Ctvm::auto_scheduler::StageAttributesStage-level attributes
 Ctvm::tir::StmtFunctor< FType >Same as ExprFunctor except it is applied on statements
 Ctvm::tir::StmtFunctor< R(const Stmt &n, Args... args)>
 Ctvm::tir::StmtFunctor< Stmt(const Stmt &)>
 Ctvm::tir::StmtFunctor< void(const Stmt &)>
 Ctvm::relax::StructInfoFunctor< FStructInfo >
 Ctvm::relax::StructInfoFunctor< R(const StructInfo &n, Args...)>
 Ctvm::relax::StructInfoFunctor< StructInfo(const StructInfo &n)>
 Ctvm::relax::StructInfoFunctor< void(const StructInfo &n)>
 Ctvm::TargetKindRegEntryHelper structure to register TargetKind
 Ctvm::TargetTagRegEntry
 Ctvm::te::TensorDomTemporary data structure to store union of bounds of each axis of Tensor
 Ctvm::runtime::threading::ThreadGroupA platform-agnostic abstraction for managing a collection of thread pool threads
 Ctvm::runtime::ThreadLocalDiscoWorkerA threadlocal wrapper of DiscoWorker
 Ctvm::arith::TransitiveComparisonAnalyzerUsing previously specified knowns, compare the expressions provided
 Cstd::true_type
 Ctvm::runtime::TVMArgsArguments into TVM functions
 Ctvm::runtime::TVMArgsSetter
 Ctvm::runtime::TVMArgsSetterApply< i, T >
 CTVMByteArrayByte array type used to pass in byte array When kTVMBytes is used as data type
 CTVMConstantInfoDescribes one constant argument to run_model
 CTVMMetadataTop-level metadata structure. Holds all other metadata types
 Ctvm::runtime::TVMMovableArgValueWithContext_Internal auxiliary struct for TypedPackedFunc to indicate a movable argument with additional context information (function name and argument index) for better error reporting
 CTVMParallelGroupEnvEnvironment for TVM parallel task
 Ctvm::runtime::TVMPODValue_Internal base class to handle conversion to POD values
 Ctvm::TVMScriptPrinterLegacy behavior of ReprPrinter
 CTVMTensorInfoDescribes one tensor argument to run_model. NOTE: while TIR allows for other types of arguments, such as scalars, the AOT run_model function does not currently accept these. Therefore it's not possible to express those in this metadata. A future patch may modify this
 CTVMValueUnion type of values being passed through API and function calls
 Ctvm::TypedEnvFunc< FType >Please refer to TypedEnvFunc<R(Args..)>
 Ctvm::TypedEnvFunc< bool(const Array< Type > &args, int num_inputs, const Attrs &attrs, const TypeReporter &reporter)>
 Ctvm::runtime::TypedPackedFunc< FType >Please refer to TypedPackedFunc<R(Args..)>
 Ctvm::runtime::TypedPackedFunc< Array< ObjectRef >(ObjectRef json_attrs)>
 Ctvm::runtime::TypedPackedFunc< Array< ObjectRef >(Schedule sch, const Array< ObjectRef > &inputs, const Array< ObjectRef > &attrs, const Optional< ObjectRef > &decision)>
 Ctvm::runtime::TypedPackedFunc< Array< RunnerResult >(int)>
 Ctvm::runtime::TypedPackedFunc< Array< te::Tensor >(const Attrs &attrs, const Array< te::Tensor > &inputs, const Type &out_type)>
 Ctvm::runtime::TypedPackedFunc< Array< TuningRecord >()>
 Ctvm::runtime::TypedPackedFunc< Array< TuningRecord >(const Workload &, int)>
 Ctvm::runtime::TypedPackedFunc< Array< tvm::runtime::NDArray >(const TuneContext &context, const Array< MeasureCandidate > &candidates)>
 Ctvm::runtime::TypedPackedFunc< bool()>
 Ctvm::runtime::TypedPackedFunc< bool(const IRModule &)>
 Ctvm::runtime::TypedPackedFunc< int()>
 Ctvm::runtime::TypedPackedFunc< int64_t()>
 Ctvm::runtime::TypedPackedFunc< ObjectRef(Array< ObjectRef > attrs)>
 Ctvm::runtime::TypedPackedFunc< Optional< IRModule >(const IRModule &, const Target &, const String &)>
 Ctvm::runtime::TypedPackedFunc< Optional< tir::Schedule >(const IRModule &, const Target &, const String &)>
 Ctvm::runtime::TypedPackedFunc< Optional< TuningRecord >(const IRModule &, const Target &, const String &)>
 Ctvm::runtime::TypedPackedFunc< R(Args...)>A PackedFunc wrapper to provide typed function signature. It is backed by a PackedFunc internally
 Ctvm::runtime::TypedPackedFunc< RunnerResult()>
 Ctvm::runtime::TypedPackedFunc< String()>
 Ctvm::runtime::TypedPackedFunc< String(const Array< ObjectRef > &inputs, const Array< ObjectRef > &attrs, const Optional< ObjectRef > &decision, const Array< String > &outputs)>
 Ctvm::runtime::TypedPackedFunc< TargetJSON(TargetJSON)>
 Ctvm::runtime::TypedPackedFunc< te::Schedule(const Attrs &attrs, const Array< te::Tensor > &outs, const Target &target)>
 Ctvm::runtime::TypedPackedFunc< tvm::tir::Stmt(Array< tvm::tir::Var > loop_vars, Array< Range > loop_extents, tvm::tir::Stmt loop_body)>
 Ctvm::runtime::TypedPackedFunc< void()>
 Ctvm::runtime::TypedPackedFunc< void(Array< TuneContext > tasks, Array< FloatImm > task_weights, int max_trials_global, int max_trials_per_task, int num_trials_per_iter, Builder builder, Runner runner, Array< MeasureCallback > measure_callbacks, Optional< Database > database, Optional< CostModel > cost_model)>
 Ctvm::runtime::TypedPackedFunc< void(const TaskScheduler &task_scheduler, int task_id, const Array< MeasureCandidate > &measure_candidates, const Array< BuilderResult > &builds, const Array< RunnerResult > &results)>
 Ctvm::runtime::TypedPackedFunc< void(const TuneContext &, const Array< MeasureCandidate > &, const Array< RunnerResult > &)>
 Ctvm::runtime::TypedPackedFunc< void(const TuneContext &, const Array< MeasureCandidate > &, void *p_addr)>
 Ctvm::runtime::TypedPackedFunc< void(const TuningRecord &)>
 Ctvm::runtime::TypedPackedFunc< void(size_t, void *)>
 Ctvm::runtime::TypedPackedFunc< void(String)>
 Ctvm::runtime::TypedPackedFunc< void(tvm::DiagnosticContext ctx)>
 Ctvm::runtime::TypedPackedFunc< Workload(const IRModule &)>
 Ctvm::TypeFunctor< FType >
 Ctvm::TypeFunctor< R(const Type &n, Args...)>
 Ctvm::TypeFunctor< Type(const Type &n)>
 Ctvm::TypeFunctor< void(const Type &n)>
 Ctvm::runtime::TypeIndexNamespace for the list of type index
 Ctvm::detail::TypeName< T >Helper struct to get the type name known to tvm
 Ctvm::detail::TypeName< bool >
 Ctvm::detail::TypeName< DataType >
 Ctvm::detail::TypeName< double >
 Ctvm::detail::TypeName< int >
 Ctvm::detail::TypeName< int64_t >
 Ctvm::detail::TypeName< uint64_t >
 Ctvm::detail::TypeName< void * >
 Ctvm::relay::v_infoA struct to keep info of traversed expr in ExpandDataflow function
 Ctvm::runtime::Array< T, typename >::ValueConverter
 Ctvm::detail::ValueTypeInfoMaker< ValueType, IsArray, IsMap >
 Ctvm::script::printer::IRDocsifierNode::VariableInfoInformation about a variable, including its optional name and its doc creator
 Ctvm::relax::VarUsageInfoA utility struct returned by CollectVarUsage
 Ctvm::VirtualDeviceCacheA cache of VirtualDevices. This can be used:
 Ctvm::runtime::vm::VMFrameA representation of a stack frame
 Ctvm::runtime::relax_vm::VMFuncInfoInformation entry in executable function table
 Ctvm::runtime::vm::VMFunctionA representation of a Relay function in the VM
 Ctvm::With< ContextType >RAII wrapper function to enter and exit a context object similar to python's with syntax
 Ctvm::runtime::WorkerZeroDataA special communication channel between controler and worker-0, assuming they are always collocated in the same process
 Ctvm::meta_schedule::WorkloadEqualThe equality check for Workload
 Ctvm::relax::WorkloadEqualThe equality check for Workload
 Ctvm::meta_schedule::WorkloadHashThe hash method for Workload
 Ctvm::runtime::WrappedPythonObjectA class that wraps a Python object and preserves its ownership