# MLIR

Multi-Level IR Compiler Framework

# Operation Definition Specification (ODS)

In addition to specializing the mlir::Op C++ template, MLIR also supports defining operations and data types in a table-driven manner. This is achieved via TableGen, which is both a generic language and its tooling to maintain records of domain-specific information. Facts regarding an operation are specified concisely into a TableGen record, which will be expanded into an equivalent mlir::Op C++ template specialization at compiler build time.

This manual explains in detail all the available mechanisms for defining operations in such a table-driven manner. It aims to be a specification instead of a tutorial. Please refer to Quickstart tutorial to adding MLIR graph rewrite for the latter.

In addition to detailing each mechanism, this manual also tries to capture best practices. They are rendered as quoted bullet points.

## Motivation ¶

MLIR allows pluggable dialects, and dialects contain, among others, a list of operations. This open and extensible ecosystem leads to the “stringly” type IR problem, e.g., repetitive string comparisons during optimization and analysis passes, unintuitive accessor methods (e.g., generic/error prone getOperand(3) vs self-documenting getStride()) with more generic return types, verbose and generic constructors without default arguments, verbose textual IR dumps, and so on. Furthermore, operation verification is:

1. best case: a central string-to-verification-function map,
2. middle case: duplication of verification across the code base, or
3. worst case: no verification functions.

The fix is to support defining ops in a table-driven manner. Then for each dialect, we can have a central place that contains everything you need to know about each op, including its constraints, custom assembly form, etc. This description is also used to generate helper functions and classes to allow building, verification, parsing, printing, analysis, and many more.

## Benefits ¶

Compared to the C++ template, this table-driven approach has several benefits including but not limited to:

• Single source of truth: We strive to encode all facts regarding an operation into the record, so that readers don’t need to jump among code snippets to fully understand an operation.
• Removing boilerplate: We can automatically generate operand/attribute/result getter methods, operation build methods, operation verify methods, and many more utilities from the record. This greatly reduces the boilerplate needed for defining a new op.
• Facilitating auto-generation: The usage of these operation information records are by no means limited to op definition itself. We can use them to drive the auto-generation of many other components, like computation graph serialization.

## TableGen Syntax ¶

We use TableGen as the language for specifying operation information. TableGen itself just provides syntax for writing records; the syntax and constructs allowed in a TableGen file (typically with the filename suffix .td) can be found here.

• TableGen class is similar to C++ class; it can be templated and subclassed.
• TableGen def is similar to C++ object; it can be declared by specializing a TableGen class (e.g., def MyDef : MyClass<...>;) or completely independently (e.g., def MyDef;). It cannot be further templated or subclassed.
• TableGen dag is a dedicated type for directed acyclic graph of elements. A dag has one operator and zero or more arguments. Its syntax is (operator arg0, arg1, argN). The operator can be any TableGen def; an argument can be anything, including dag itself. We can have names attached to both the operator and the arguments like (MyOp:$op_name MyArg:$arg_name).

Please see the language reference to learn about all the types and expressions supported by TableGen.

## Operation Definition ¶

MLIR defines several common constructs to help operation definition and provide their semantics via a special TableGen backend: OpDefinitionsGen. These constructs are defined in OpBase.td. The main ones are:

• The Op class: It is the main construct for defining operations. All facts regarding the operation are specified when specializing this class, with the help of the following constructs.
• The Dialect class: Operations belonging to one logical group are placed in the same dialect. The Dialect class contains dialect-level information.
• The OpTrait class hierarchy: They are used to specify special properties and constraints of the operation, including whether the operation has side effect or whether its output has the same shape as the input.
• The ins/outs marker: These are two special markers builtin to the OpDefinitionsGen backend. They lead to the definitions of operands/attributes and results respectively.
• The TypeConstraint class hierarchy: They are used to specify the constraints over operands or results. A notable subclass hierarchy is Type, which stands for constraints for common C++ types.
• The AttrConstraint class hierarchy: They are used to specify the constraints over attributes. A notable subclass hierarchy is Attr, which stands for constraints for attributes whose values are of common types.

An operation is defined by specializing the Op class with concrete contents for all the fields it requires. For example, tf.AvgPool is defined as

def TF_AvgPoolOp : TF_Op<"AvgPool", [NoSideEffect]> {
let summary = "Performs average pooling on the input.";

let description = [{
Each entry in output is the mean of the corresponding size ksize
window in value.
}];

let arguments = (ins
TF_FpTensor:$value, ConfinedAttr<I64ArrayAttr, [ArrayMinCount<4>]>:$ksize,
ConfinedAttr<I64ArrayAttr, [ArrayMinCount<4>]>:$strides, TF_AnyStrAttrOf<["SAME", "VALID"]>:$padding,
DefaultValuedAttr<TF_ConvertDataFormatAttr, "NHWC">:$data_format ); let results = (outs TF_FpTensor:$output
);

TF_DerivedOperandTypeAttr T = TF_DerivedOperandTypeAttr<0>;
}


In the following we describe all the fields needed. Please see the definition of the Op class for the complete list of fields supported.

### Operation name ¶

The operation name is a unique identifier for the operation within MLIR, e.g., tf.Add for addition operation in the TensorFlow dialect. This is the equivalent of the mnemonic in assembly language. It is used for parsing and printing in the textual format. It is also used for pattern matching in graph rewrites.

The full operation name is composed of the dialect name and the op name, with the former provided via the dialect and the latter provided as the second template parameter to the Op class.

### Operation documentation ¶

This includes both a one-line summary and a longer human-readable description. They will be used to drive automatic generation of dialect documentation. They need to be provided in the operation’s definition body:

let summary = "...";

let description = [{
...
}];


description should be written in Markdown syntax.

Placing the documentation at the beginning is recommended since it helps in understanding the operation.

• Place documentation at the beginning of the operation definition
• The summary should be short and concise. It should be a one-liner without trailing punctuation. Put expanded explanation in description.

### Operation arguments ¶

There are two kinds of arguments: operands and attributes. Operands are runtime values produced by other ops; while attributes are compile-time known constant values, including two categories:

1. Natural attributes: these attributes affect the behavior of the operations (e.g., padding for convolution);

2. Derived attributes: these attributes are not needed to define the operation but are instead derived from information of the operation. E.g., the output shape of type. This is mostly used for convenience interface generation or interaction with other frameworks/translation.

All derived attributes should be materializable as an Attribute. That is, even though they are not materialized, it should be possible to store as an attribute.

Both operands and attributes are specified inside the dag-typed arguments, led by ins:

let arguments = (ins
<type-constraint>:$<operand-name>, ... <attr-constraint>:$<attr-name>,
...
);


Here <type-constraint> is a TableGen def from the TypeConstraint class hierarchy. Similarly, <attr-constraint> is a TableGen def from the AttrConstraint class hierarchy. See Constraints for more information.

There is no requirements on the relative order of operands and attributes; they can mix freely. The relative order of operands themselves matters. From each named argument a named getter will be generated that returns the argument with the return type (in the case of attributes the return type will be constructed from the storage type, while for operands it will be Value). Each attribute’s raw value (e.g., as stored) can also be accessed via generated <name>Attr getters for use in transformation passes where the more user-friendly return type is less suitable.

All the arguments should be named to:

• provide documentation,
• drive auto-generation of getter methods, and
• provide a handle to reference for other places like constraints.

To declare a variadic operand, wrap the TypeConstraint for the operand with Variadic<...>.

Normally operations have no variadic operands or just one variadic operand. For the latter case, it is easy to deduce which dynamic operands are for the static variadic operand definition. However, if an operation has more than one variable length operands (either optional or variadic), it would be impossible to attribute dynamic operands to the corresponding static variadic operand definitions without further information from the operation. Therefore, either the SameVariadicOperandSize or AttrSizedOperandSegments trait is needed to indicate that all variable length operands have the same number of dynamic values.

To declare a variadic operand that has a variadic number of sub-ranges, wrap the TypeConstraint for the operand with VariadicOfVariadic<..., "<segment-attribute-name>">.

The second field of the VariadicOfVariadic is the name of an I32ElementsAttr argument that contains the sizes of the variadic sub-ranges. This attribute will be used when determining the size of sub-ranges, or when updating the size of sub-ranges.

#### Optional operands ¶

To declare an optional operand, wrap the TypeConstraint for the operand with Optional<...>.

Normally operations have no optional operands or just one optional operand. For the latter case, it is easy to deduce which dynamic operands are for the static operand definition. However, if an operation has more than one variable length operands (either optional or variadic), it would be impossible to attribute dynamic operands to the corresponding static variadic operand definitions without further information from the operation. Therefore, either the SameVariadicOperandSize or AttrSizedOperandSegments trait is needed to indicate that all variable length operands have the same number of dynamic values.

#### Optional attributes ¶

To declare an optional attribute, wrap the AttrConstraint for the attribute with OptionalAttr<...>.

#### Attributes with default values ¶

To declare an attribute with a default value, wrap the AttrConstraint for the attribute with DefaultValuedAttr<..., "...">.

The second parameter to DefaultValuedAttr should be a string containing the C++ default value. For example, a float default value should be specified as like "0.5f", and an integer array default value should be specified as like "{1, 2, 3}".

#### Confining attributes ¶

ConfinedAttr is provided as a general mechanism to help modelling further constraints on attributes beyond the ones brought by value types. You can use ConfinedAttr to compose complex constraints out of more primitive ones. For example, a 32-bit integer attribute whose minimum value must be 10 can be expressed as ConfinedAttr<I32Attr, [IntMinValue<10>]>.

Right now, the following primitive constraints are supported:

• IntMinValue<N>: Specifying an integer attribute to be greater than or equal to N
• IntMaxValue<N>: Specifying an integer attribute to be less than or equal to N
• ArrayMinCount<N>: Specifying an array attribute to have at least N elements
• IntArrayNthElemEq<I, N>: Specifying an integer array attribute’s I-th element to be equal to N
• IntArrayNthElemMinValue<I, N>: Specifying an integer array attribute’s I-th element to be greater than or equal to N

TODO: Design and implement more primitive constraints

### Operation regions ¶

The regions of an operation are specified inside of the dag-typed regions, led by region:

let regions = (region
<region-constraint>:$<region-name>, ... );  #### Variadic regions ¶ Similar to the Variadic class used for variadic operands and results, VariadicRegion<...> can be used for regions. Variadic regions can currently only be specified as the last region in the regions list. ### Operation results ¶ Similar to operands, results are specified inside the dag-typed results, led by outs: let results = (outs <type-constraint>:$<result-name>,
...
);


Similar to variadic operands, Variadic<...> can also be used for results. And similarly, SameVariadicResultSize for multiple variadic results in the same operation.

### Operation successors ¶

For terminator operations, the successors are specified inside of the dag-typed successors, led by successor:

let successors = (successor
<successor-constraint>:$<successor-name>, ... );  #### Variadic successors ¶ Similar to the Variadic class used for variadic operands and results, VariadicSuccessor<...> can be used for successors. Variadic successors can currently only be specified as the last successor in the successor list. ### Operation traits and constraints ¶ Traits are operation properties that affect syntax or semantics. MLIR C++ models various traits in the mlir::OpTrait namespace. Both operation traits, interfaces, and constraints involving multiple operands/attributes/results are provided as the third template parameter to the Op class. They should be deriving from the OpTrait class. See Constraints for more information. ### Builder methods ¶ For each operation, there are a few builders automatically generated based on the arguments and returns types. For example, given the following op definition: def MyOp : ... { let arguments = (ins I32:$i32_operand,
F32:$f32_operand, ..., I32Attr:$i32_attr,
F32Attr:$f32_attr, ... ); let results = (outs I32:$i32_result,
F32:$f32_result, ... ); }  The following builders are generated: // All result-types/operands/attributes have one aggregate parameter. static void build(OpBuilder &odsBuilder, OperationState &odsState, TypeRange resultTypes, ValueRange operands, ArrayRef<NamedAttribute> attributes); // Each result-type/operand/attribute has a separate parameter. The parameters // for attributes are of mlir::Attribute types. static void build(OpBuilder &odsBuilder, OperationState &odsState, Type i32_result, Type f32_result, ..., Value i32_operand, Value f32_operand, ..., IntegerAttr i32_attr, FloatAttr f32_attr, ...); // Each result-type/operand/attribute has a separate parameter. The parameters // for attributes are raw values unwrapped with mlir::Attribute instances. // (Note that this builder will not always be generated. See the following // explanation for more details.) static void build(OpBuilder &odsBuilder, OperationState &odsState, Type i32_result, Type f32_result, ..., Value i32_operand, Value f32_operand, ..., APInt i32_attr, StringRef f32_attr, ...); // Each operand/attribute has a separate parameter but result type is aggregate. static void build(OpBuilder &odsBuilder, OperationState &odsState, TypeRange resultTypes, Value i32_operand, Value f32_operand, ..., IntegerAttr i32_attr, FloatAttr f32_attr, ...); // All operands/attributes have aggregate parameters. // Generated if return type can be inferred. static void build(OpBuilder &odsBuilder, OperationState &odsState, ValueRange operands, ArrayRef<NamedAttribute> attributes); // (And manually specified builders depending on the specific op.)  The first form provides basic uniformity so that we can create ops using the same form regardless of the exact op. This is particularly useful for implementing declarative pattern rewrites. The second and third forms are good for use in manually written code, given that they provide better guarantee via signatures. The third form will be generated if any of the op’s attribute has different Attr.returnType from Attr.storageType and we know how to build an attribute from an unwrapped value (i.e., Attr.constBuilderCall is defined.) Additionally, for the third form, if an attribute appearing later in the arguments list has a default value, the default value will be supplied in the declaration. This works for BoolAttr, StrAttr, EnumAttr for now and the list can grow in the future. So if possible, the default-valued attribute should be placed at the end of the arguments list to leverage this feature. (This behavior is essentially due to C++ function parameter default value placement restrictions.) Otherwise, the builder of the third form will still be generated but default values for the attributes not at the end of the arguments list will not be supplied in the builder’s signature. ODS will generate a builder that doesn’t require the return type specified if • Op implements InferTypeOpInterface interface; • All return types are either buildable types or are the same as a given operand (e.g., AllTypesMatch constraint between operand and result); And there may potentially exist other builders depending on the specific op; please refer to the generated C++ file for the complete list. #### Custom builder methods ¶ However, if the above cases cannot satisfy all needs, you can define additional convenience build methods in the builders field as follows. def MyOp : Op<"my_op", []> { let arguments = (ins F32Attr:$attr);

let builders = [
OpBuilder<(ins "float":$val)> ]; }  The builders field is a list of custom builders that are added to the Op class. In this example, we provide a convenience builder that takes a floating point value instead of an attribute. The ins prefix is common to many function declarations in ODS, which use a TableGen dag. What follows is a comma-separated list of types (quoted string) and names prefixed with the $ sign. This will generate the declaration of a builder method that looks like:

class MyOp : /*...*/ {
/*...*/
static void build(::mlir::OpBuilder &builder, ::mlir::OperationState &state,
float val);
};


Note that the method has two additional leading arguments. These arguments are useful to construct the operation. In particular, the method must populate state with attributes, operands, regions and result types of the operation to be constructed. builder can be used to construct any IR objects that belong to the Op, such as types or nested operations. Since the type and name are generated as is in the C++ code, they should be valid C++ constructs for a type (in the namespace of the Op) and an identifier (e.g., class is not a valid identifier).

Implementations of the builder can be provided directly in ODS, using TableGen code block as follows.

def MyOp : Op<"my_op", []> {
let arguments = (ins F32Attr:$attr); let builders = [ OpBuilder<(ins "float":$val), [{
$_state.addAttribute("attr",$_builder.getF32FloatAttr(val));
}]>
];
}


The equivalents of builder and state arguments are available as $_builder and $_state special variables. The named arguments listed in the ins part are available directly, e.g. val. The body of the builder will be generated by substituting special variables and should otherwise be valid C++. While there is no limitation on the code size, we encourage one to define only short builders inline in ODS and put definitions of longer builders in C++ files.

Finally, if some arguments need a default value, they can be defined using CArg to wrap the type and this value as follows.

def MyOp : Op<"my_op", []> {
let arguments = (ins F32Attr:$attr); let builders = [ OpBuilder<(ins CArg<"float", "0.5f">:$val), [{
$_state.addAttribute("attr",$_builder.getF32FloatAttr(val));
}]>
];
}


The generated code will use default value in the declaration, but not in the definition, as required by C++.

/// Header file.
class MyOp : /*...*/ {
/*...*/
static void build(::mlir::OpBuilder &builder, ::mlir::OperationState &state,
float val = 0.5f);
};

/// Source file.
MyOp::build(::mlir::OpBuilder &builder, ::mlir::OperationState &state,
float val) {
}


Deprecated: OpBuilder class allows one to specify the custom builder signature as a raw string, without separating parameters into different dag arguments. It also supports leading parameters of OpBuilder & and OperationState & types, which will be used instead of the autogenerated ones if present.

### Custom parser and printer methods ¶

Functions to parse and print the operation’s custom assembly form.

### Custom verifier code ¶

Verification code will be automatically generated for constraints specified on various entities of the op. To perform additional verification, you can use

let hasVerifier = 1;
let hasRegionVerifier = 1;


This will generate LogicalResult verify()/LogicalResult verifyRegions() method declarations on the op class that can be defined with any additional verification constraints. For verificaiton which needs to access the nested operations, you should use hasRegionVerifier to ensure that it won’t access any ill-formed operation. Except that, The other verifications can be implemented with hasVerifier. Check the next section for the execution order of these verification methods.

#### Verification Ordering ¶

The verification of an operation involves several steps,

1. StructuralOpTrait will be verified first, they can be run independently.
2. verifyInvariants which is constructed by ODS, it verifies the type, attributes, .etc.
3. Other Traits/Interfaces that have marked their verifier as verifyTrait or verifyWithRegions=0.
4. Custom verifier which is defined in the op and has been marked hasVerifier=1

If an operation has regions, then it may have the second phase,

1. Traits/Interfaces that have marked their verifier as verifyRegionTrait or verifyWithRegions=1. This implies the verifier needs to access the operations in its regions.
2. Custom verifier which is defined in the op and has been marked hasRegionVerifier=1

Note that the second phase will be run after the operations in the region are verified. Verifiers further down the order can rely on certain invariants being verified by a previous verifier and do not need to re-verify them.

#### Emitting diagnostics in custom verifiers ¶

Custom verifiers should avoid printing operations using custom operation printers, because they require the printed operation (and sometimes its parent operation) to be verified first. In particular, when emitting diagnostics, custom verifiers should use the Error severity level, which prints operations in generic form by default, and avoid using lower severity levels (Note, Remark, Warning).

### Declarative Assembly Format ¶

The custom assembly form of the operation may be specified in a declarative string that matches the operations operands, attributes, etc. With the ability to express additional information that needs to be parsed to build the operation:

def CallOp : Std_Op<"call", ...> {
let arguments = (ins FlatSymbolRefAttr:$callee, Variadic<AnyType>:$args);

let assemblyFormat = [{

#### Literals ¶

A literal is either a keyword or punctuation surrounded by .

The following are the set of valid punctuation:

:, ,, =, <, >, (, ), {, }, [, ], ->, ?, +, *

The following are valid whitespace punctuation:

\n,

The \n literal emits a newline an indents to the start of the operation. An example is shown below:

let assemblyFormat = [{
{ \n     this_is_on_a_newline \n } attr-dict
}];

%results = my.operation {
this_is_on_a_newline
}


An empty literal  may be used to remove a space that is inserted implicitly after certain literal elements, such as )/]/etc. For example, “]” may result in an output of ] it is not the last element in the format. “] ” would trim the trailing space in this situation.

#### Variables ¶

A variable is an entity that has been registered on the operation itself, i.e. an argument(attribute or operand), region, result, successor, etc. In the CallOp example above, the variables would be $callee and $args.

Attribute variables are printed with their respective value type, unless that value type is buildable. In those cases, the type of the attribute is elided.

#### Custom Directives ¶

The declarative assembly format specification allows for handling a large majority of the common cases when formatting an operation. For the operations that require or desire specifying parts of the operation in a form not supported by the declarative syntax, custom directives may be specified. A custom directive essentially allows for users to use C++ for printing and parsing subsections of an otherwise declaratively specified format. Looking at the specification of a custom directive above:

custom-directive ::= custom < UserDirective > ( Params )


A custom directive has two main parts: The UserDirective and the Params. A custom directive is transformed into a call to a print* and a parse* method when generating the C++ code for the format. The UserDirective is an identifier used as a suffix to these two calls, i.e., custom<MyDirective>(...) would result in calls to parseMyDirective and printMyDirective within the parser and printer respectively. Params may be any combination of variables (i.e. Attribute, Operand, Successor, etc.), type directives, attr-dict, and strings of C++ code. The type directives must refer to a variable, but that variable need not also be a parameter to the custom directive.

The arguments to the parse<UserDirective> method are firstly a reference to the OpAsmParser(OpAsmParser &), and secondly a set of output parameters corresponding to the parameters specified in the format. The mapping of declarative parameter to parse method argument is detailed below:

• Attribute Variables
• Single: <Attribute-Storage-Type>(e.g. Attribute) &
• Optional: <Attribute-Storage-Type>(e.g. Attribute) &
• Operand Variables
• Single: OpAsmParser::UnresolvedOperand &
• Optional: Optional<OpAsmParser::UnresolvedOperand> &
• Variadic: SmallVectorImpl<OpAsmParser::UnresolvedOperand> &
• VariadicOfVariadic: SmallVectorImpl<SmallVector<OpAsmParser::UnresolvedOperand>> &
• Ref Directives
• A reference directive is passed to the parser using the same mapping as the input operand. For example, a single region would be passed as a Region &.
• Region Variables
• Single: Region &
• Variadic: SmallVectorImpl<std::unique_ptr<Region>> &
• Successor Variables
• Single: Block *&
• Variadic: SmallVectorImpl<Block *> &
• Type Directives
• Single: Type &
• Optional: Type &
• Variadic: SmallVectorImpl<Type> &
• VariadicOfVariadic: SmallVectorImpl<SmallVector<Type>> &
• attr-dict Directive: NamedAttrList &

When a variable is optional, the value should only be specified if the variable is present. Otherwise, the value should remain None or null.

The arguments to the print<UserDirective> method is firstly a reference to the OpAsmPrinter(OpAsmPrinter &), second the op (e.g. FooOp op which can be Operation *op alternatively), and finally a set of output parameters corresponding to the parameters specified in the format. The mapping of declarative parameter to print method argument is detailed below:

• Attribute Variables
• Single: <Attribute-Storage-Type>(e.g. Attribute)
• Optional: <Attribute-Storage-Type>(e.g. Attribute)
• Operand Variables
• Single: Value
• Optional: Value
• Variadic: OperandRange
• VariadicOfVariadic: OperandRangeRange
• Ref Directives
• A reference directive is passed to the printer using the same mapping as the input operand. For example, a single region would be passed as a Region &.
• Region Variables
• Single: Region &
• Variadic: MutableArrayRef<Region>
• Successor Variables
• Single: Block *
• Variadic: SuccessorRange
• Type Directives
• Single: Type
• Optional: Type
• Variadic: TypeRange
• VariadicOfVariadic: TypeRangeRange
• attr-dict Directive: DictionaryAttr

When a variable is optional, the provided value may be null. When a variable is referenced in a custom directive parameter using ref, it is passed in by value. Referenced variables to print<UserDirective> are passed as the same as bound variables, but referenced variables to parse<UserDirective> are passed like to the printer.

A custom directive can take a string of C++ code as a parameter. The code is pasted verbatim in the calls to the custom parser and printers, with the substitutions $_builder and $_ctxt. String literals can be used to parameterize custom directives.

#### Optional Groups ¶

In certain situations operations may have “optional” information, e.g. attributes or an empty set of variadic operands. In these situations a section of the assembly format can be marked as optional based on the presence of this information. An optional group is defined as follows:

optional-group: ( then-elements ) (: ( else-elements ))? ?


The elements of an optional group have the following requirements:

• The first element of then-elements must either be a attribute, literal, operand, or region.
• This is because the first element must be optionally parsable.
• Exactly one argument variable or type directive within either then-elements or else-elements must be marked as the anchor of the group.
• The anchor is the element whose presence controls which elements should be printed/parsed.
• An element is marked as the anchor by adding a trailing ^.
• The first element is not required to be the anchor of the group.
• When a non-variadic region anchors a group, the detector for printing the group is if the region is empty.
• Literals, variables, custom directives, and type directives are the only valid elements within the group.
• Any attribute variable may be used, but only optional attributes can be marked as the anchor.
• Only variadic or optional results and operand arguments and can be used.
• All region variables can be used. When a non-variable length region is used, if the group is not present the region is empty.

An example of an operation with an optional group is func.return, which has a variadic number of operands.

def ReturnOp : ... {

### Generated C++ code ¶

OpDefinitionsGen processes the op definition spec file and generates two files containing the corresponding C++ code: one for declarations, the other for definitions. The former is generated via the -gen-op-decls command-line option, while the latter is via the -gen-op-defs option.

The definition file contains all the op method definitions, which can be included and enabled by defining GET_OP_CLASSES. For each operation, OpDefinitionsGen generates an operation class and an operand adaptor class. Besides, it also contains a comma-separated list of all defined ops, which can be included and enabled by defining GET_OP_LIST.

#### Class name and namespaces ¶

For each operation, its generated C++ class name is the symbol defed with TableGen with dialect prefix removed. The first _ serves as the delimiter. For example, for def TF_AddOp, the C++ class name would be AddOp. We remove the TF prefix because it is for scoping ops; other dialects may as well define their own AddOps.

The namespaces of the generated C++ class will come from the dialect’s cppNamespace field. For example, if a dialect’s cppNamespace is A::B, then an op of that dialect will be placed in namespace A { namespace B { ... } }. If a dialect does not specify a cppNamespace, we then use the dialect’s name as the namespace.

This means the qualified name of the generated C++ class does not necessarily match exactly with the operation name as explained in Operation name. This is to allow flexible naming to satisfy coding style requirements.

For each operation, we automatically generate an operand adaptor. This class solves the problem of accessing operands provided as a list of Values without using “magic” constants. The operand adaptor takes a reference to an array of Value and provides methods with the same names as those in the operation class to access them. For example, for a binary arithmetic operation, it may provide .lhs() to access the first operand and .rhs() to access the second operand.

The operand adaptor class lives in the same namespace as the operation class, and has the name of the operation followed by Adaptor as well as an alias Adaptor inside the op class.

Operand adaptors can be used in function templates that also process operations:

template <typename BinaryOpTy>
std::pair<Value, Value> zip(BinaryOpTy &&op) {
return std::make_pair(op.lhs(), op.rhs());;
}

void process(AddOp op, ArrayRef<Value> newOperands) {
zip(op);
/*...*/
}


## Constraints ¶

Constraint is a core concept in table-driven operation definition: operation verification and graph operation matching are all based on satisfying constraints. So both the operation definition and rewrite rules specification significantly involve writing constraints. We have the Constraint class in OpBase.td as the common base class for all constraints.

An operation’s constraint can cover different range; it may

• Only concern a single attribute (e.g. being a 32-bit integer greater than 5),
• Multiple operands and results (e.g., the 1st result’s shape must be the same as the 1st operand), or
• Intrinsic to the operation itself (e.g., having no side effect).

We call them as single-entity constraint, multi-entity constraint, and traits, respectively.

### Single-entity constraint ¶

Constraints scoped to a single operand, attribute, or result are specified at the entity’s declaration place as described in Operation arguments and Operation results.

To help modelling constraints of common types, a set of TypeConstraints are created; they are the Type subclass hierarchy. It includes F32 for the constraints of being a float, TensorOf<[F32]> for the constraints of being a float tensor, and so on.

Similarly, a set of AttrConstraints are created for helping modelling constraints of common attribute kinds. They are the Attr subclass hierarchy. It includes F32Attr for the constraints of being a float attribute, F32ArrayAttr for the constraints of being a float array attribute, and so on.

### Multi-entity constraint ¶

Constraints involving more than one operand/attribute/result are quite common on operations, like the element type and shape relation between operands and results. These constraints should be specified as the Op class template parameter as described in Operation traits and constraints.

Multi-entity constraints are modeled as PredOpTrait (a subclass of OpTrait) in OpBase.td.A bunch of constraint primitives are provided to help specification. See OpBase.td for the complete list.

### Trait ¶

Traits are intrinsic properties of the operation like having side effect or not, commutative or not, whether is a terminator, etc. These constraints should be specified as the Op class template parameter as described in Operation traits and constraints.

Traits are modeled as NativeOpTrait (a subclass of OpTrait) in OpBase.td. They are backed and will be translated into the corresponding C++ mlir::OpTrait classes.

### How to specify new constraint ¶

To write a constraint, you need to provide its predicates and give it a descriptive name. Predicates, modeled with the Pred class, are the workhorse for composing constraints. The predicate for a constraint is typically built up in a nested manner, using the two categories of predicates:

1. CPred: the primitive leaf predicate.
2. Compound predicate: a predicate composed from child predicates using predicate combiners (conjunction: And, disjunction: Or, negation: Neg, substitution: SubstLeaves, concatenation: Concat).

CPred is the basis for composing more complex predicates. It is the “atom” predicate from the perspective of TableGen and the “interface” between TableGen and C++. What is inside is already C++ code, which will be treated as opaque strings with special placeholders to be substituted.

You can put any C++ code that returns a boolean value inside a CPred, including evaluating expressions, calling functions, calling class methods, and so on.

To help interaction with the C++ environment, there are a few special placeholders provided to refer to entities in the context where this predicate is used. They serve as “hooks” to the enclosing environment. This includes $_builder, $_op, and $_self: • $_builder will be replaced by a mlir::Builder instance so that you can access common build methods.
• $_op will be replaced by the current operation so that you can access information of the current operation. • $_self will be replaced with the entity this predicate is attached to. E.g., BoolAttr is an attribute constraint that wraps a CPred<"$_self.isa<BoolAttr>()">. Then for BoolAttr:$attr,$_self will be replaced by $attr. For type constraints, it’s a little bit special since we want the constraints on each type definition reads naturally and we want to attach type constraints directly to an operand/result, $_self will be replaced by the operand/result’s type. E.g., for F32 in F32:$operand, its $_self will be expanded as operand(...).getType(). TODO: Reconsider the leading symbol for special placeholders. Eventually we want to allow referencing operand/result $-names; such $-names can start with underscore. For example, to write an attribute attr is an IntegerAttr, in C++ you can just call attr.isa<IntegerAttr>(). The code can be wrapped in a CPred as $_self.isa<IntegerAttr>(), with $_self as the special placeholder to be replaced by the current attribute attr at expansion time. For more complicated predicates, you can wrap it in a single CPred, or you can use predicate combiners to combine them. For example, to write the constraint that an attribute attr is a 32-bit or 64-bit integer, you can write it as And<[ CPred<"$_self.isa<IntegerAttr>()">,
Or<[
CPred<"$_self.cast<IntegerAttr>().getType().isInteger(32)">, CPred<"$_self.cast<IntegerAttr>().getType().isInteger(64)">
]>
]>


(Note that the above is just to show with a familiar example how you can use CPred and predicate combiners to write complicated predicates. For integer attributes specifically, OpBase.td already defines I32Attr and I64Attr. So you can actually reuse them to write it as Or<[I32Attr.predicate, I64Attr.predicate]>.)

TODO: Build up a library of reusable primitive constraints

If the predicate is very complex to write with CPred together with predicate combiners, you can also write it as a normal C++ function and use the CPred as a way to “invoke” the function. For example, to verify an attribute attr has some property, you can write a C++ function like

bool HasSomeProperty(Attribute attr) { ... }


and then define the op as:

def HasSomeProperty : AttrConstraint<CPred<"HasSomeProperty($_self)">, "has some property">; def MyOp : Op<...> { let arguments = (ins ... HasSomeProperty:$attr
);
}


As to whether we should define the predicate using a single CPred wrapping the whole expression, multiple CPreds with predicate combiners, or a single CPred “invoking” a function, there are no clear-cut criteria. Defining using CPred and predicate combiners is preferable since it exposes more information (instead hiding all the logic behind a C++ function) into the op definition spec so that it can potentially drive more auto-generation cases. But it will require a nice library of common predicates as the building blocks to avoid the duplication, which is being worked on right now.

## Attribute Definition ¶

An attribute is a compile-time known constant of an operation.

ODS provides attribute wrappers over C++ attribute classes. There are a few common C++ attribute classes defined in MLIR’s core IR library and one is free to define dialect-specific attribute classes. ODS allows one to use these attributes in TableGen to define operations, potentially with more fine-grained constraints. For example, StrAttr directly maps to StringAttr; F32Attr/F64Attr requires the FloatAttr to additionally be of a certain bitwidth.

ODS attributes are defined as having a storage type (corresponding to a backing mlir::Attribute that stores the attribute), a return type (corresponding to the C++ return type of the generated helper getters) as well as a method to convert between the internal storage and the helper method.

### Attribute decorators ¶

There are a few important attribute adapters/decorators/modifiers that can be applied to ODS attributes to specify common additional properties like optionality, default values, etc.:

• DefaultValuedAttr: specifies the default value for an attribute.
• OptionalAttr: specifies an attribute as optional.
• ConfinedAttr: adapts an attribute with further constraints.

### Enum attributes ¶

Some attributes can only take values from a predefined enum, e.g., the comparison kind of a comparison op. To define such attributes, ODS provides several mechanisms: IntEnumAttr, and BitEnumAttr.

• IntEnumAttr: each enum case is an integer, the attribute is stored as a IntegerAttr in the op.
• BitEnumAttr: each enum case is a either the empty case, a single bit, or a group of single bits, and the attribute is stored as a IntegerAttr in the op.

All these *EnumAttr attributes require fully specifying all of the allowed cases via their corresponding *EnumAttrCase. With this, ODS is able to generate additional verification to only accept allowed cases. To facilitate the interaction between *EnumAttrs and their C++ consumers, the EnumsGen TableGen backend can generate a few common utilities: a C++ enum class, llvm::DenseMapInfo for the enum class, conversion functions from/to strings. This is controlled via the -gen-enum-decls and -gen-enum-defs command-line options of mlir-tblgen.

For example, given the following EnumAttr:

def Case15: I32EnumAttrCase<"Case15", 15>;
def Case20: I32EnumAttrCase<"Case20", 20>;

def MyIntEnum: I32EnumAttr<"MyIntEnum", "An example int enum",
[Case15, Case20]> {
let cppNamespace = "Outer::Inner";
let stringToSymbolFnName = "ConvertToEnum";
let symbolToStringFnName = "ConvertToString";
}


The following will be generated via mlir-tblgen -gen-enum-decls:

namespace Outer {
namespace Inner {
// An example int enum
enum class MyIntEnum : uint32_t {
Case15 = 15,
Case20 = 20,
};

llvm::Optional<MyIntEnum> symbolizeMyIntEnum(uint32_t);
llvm::StringRef ConvertToString(MyIntEnum);
llvm::Optional<MyIntEnum> ConvertToEnum(llvm::StringRef);
inline constexpr unsigned getMaxEnumValForMyIntEnum() {
return 20;
}

} // namespace Inner
} // namespace Outer

namespace llvm {
template<> struct DenseMapInfo<Outer::Inner::MyIntEnum> {
using StorageInfo = llvm::DenseMapInfo<uint32_t>;

static inline Outer::Inner::MyIntEnum getEmptyKey() {
return static_cast<Outer::Inner::MyIntEnum>(StorageInfo::getEmptyKey());
}

static inline Outer::Inner::MyIntEnum getTombstoneKey() {
return static_cast<Outer::Inner::MyIntEnum>(StorageInfo::getTombstoneKey());
}

static unsigned getHashValue(const Outer::Inner::MyIntEnum &val) {
return StorageInfo::getHashValue(static_cast<uint32_t>(val));
}

static bool isEqual(const Outer::Inner::MyIntEnum &lhs, const Outer::Inner::MyIntEnum &rhs) {
return lhs == rhs;
}
};
}


The following will be generated via mlir-tblgen -gen-enum-defs:

namespace Outer {
namespace Inner {
llvm::StringRef ConvertToString(MyIntEnum val) {
switch (val) {
case MyIntEnum::Case15: return "Case15";
case MyIntEnum::Case20: return "Case20";
}
return "";
}

llvm::Optional<MyIntEnum> ConvertToEnum(llvm::StringRef str) {
return llvm::StringSwitch<llvm::Optional<MyIntEnum>>(str)
.Case("Case15", MyIntEnum::Case15)
.Case("Case20", MyIntEnum::Case20)
.Default(llvm::None);
}
llvm::Optional<MyIntEnum> symbolizeMyIntEnum(uint32_t value) {
switch (value) {
case 15: return MyIntEnum::Case15;
case 20: return MyIntEnum::Case20;
default: return llvm::None;
}
}

} // namespace Inner
} // namespace Outer


Similarly for the following BitEnumAttr definition:

def None: I32BitEnumAttrCaseNone<"None">;
def Bit0: I32BitEnumAttrCaseBit<"Bit0", 0, "tagged">;
def Bit1: I32BitEnumAttrCaseBit<"Bit1", 1>;
def Bit2: I32BitEnumAttrCaseBit<"Bit2", 2>;
def Bit3: I32BitEnumAttrCaseBit<"Bit3", 3>;

def MyBitEnum: BitEnumAttr<"MyBitEnum", "An example bit enum",
[None, Bit0, Bit1, Bit2, Bit3]>;


We can have:

// An example bit enum
enum class MyBitEnum : uint32_t {
None = 0,
Bit0 = 1,
Bit1 = 2,
Bit2 = 4,
Bit3 = 8,
};

llvm::Optional<MyBitEnum> symbolizeMyBitEnum(uint32_t);
std::string stringifyMyBitEnum(MyBitEnum);
llvm::Optional<MyBitEnum> symbolizeMyBitEnum(llvm::StringRef);

inline constexpr MyBitEnum operator|(MyBitEnum a, MyBitEnum b) {
return static_cast<MyBitEnum>(static_cast<uint32_t>(a) | static_cast<uint32_t>(b));
}
inline constexpr MyBitEnum operator&(MyBitEnum a, MyBitEnum b) {
return static_cast<MyBitEnum>(static_cast<uint32_t>(a) & static_cast<uint32_t>(b));
}
inline constexpr MyBitEnum operator^(MyBitEnum a, MyBitEnum b) {
return static_cast<MyBitEnum>(static_cast<uint32_t>(a) ^ static_cast<uint32_t>(b));
}
inline constexpr MyBitEnum operator~(MyBitEnum bits) {
// Ensure only bits that can be present in the enum are set
return static_cast<MyBitEnum>(~static_cast<uint32_t>(bits) & static_cast<uint32_t>(15u));
}
inline constexpr bool bitEnumContainsAll(MyBitEnum bits, MyBitEnum bit) {
return (bits & bit) == bit;
}
inline constexpr bool bitEnumContainsAny(MyBitEnum bits, MyBitEnum bit) {
return (static_cast<uint32_t>(bits) & static_cast<uint32_t>(bit)) != 0;
}
inline constexpr MyBitEnum bitEnumClear(MyBitEnum bits, MyBitEnum bit) {
return bits & ~bit;
}

inline std::string stringifyEnum(MyBitEnum enumValue) {
return stringifyMyBitEnum(enumValue);
}

template <typename EnumType>
::llvm::Optional<EnumType> symbolizeEnum(::llvm::StringRef);

template <>
inline ::llvm::Optional<MyBitEnum> symbolizeEnum<MyBitEnum>(::llvm::StringRef str) {
return symbolizeMyBitEnum(str);
}

namespace llvm {
template<> struct DenseMapInfo<::MyBitEnum> {
using StorageInfo = llvm::DenseMapInfo<uint32_t>;

static inline ::MyBitEnum getEmptyKey() {
return static_cast<::MyBitEnum>(StorageInfo::getEmptyKey());
}

static inline ::MyBitEnum getTombstoneKey() {
return static_cast<::MyBitEnum>(StorageInfo::getTombstoneKey());
}

static unsigned getHashValue(const ::MyBitEnum &val) {
return StorageInfo::getHashValue(static_cast<uint32_t>(val));
}

static bool isEqual(const ::MyBitEnum &lhs, const ::MyBitEnum &rhs) {
return lhs == rhs;
}
};

std::string stringifyMyBitEnum(MyBitEnum symbol) {
auto val = static_cast<uint32_t>(symbol);
assert(15u == (15u | val) && "invalid bits set in bit enum");
// Special case for all bits unset.
if (val == 0) return "None";
llvm::SmallVector<llvm::StringRef, 2> strs;
if (1u == (1u & val)) { strs.push_back("tagged"); }
if (2u == (2u & val)) { strs.push_back("Bit1"); }
if (4u == (4u & val)) { strs.push_back("Bit2"); }
if (8u == (8u & val)) { strs.push_back("Bit3"); }

return llvm::join(strs, "|");
}

llvm::Optional<MyBitEnum> symbolizeMyBitEnum(llvm::StringRef str) {
// Special case for all bits unset.
if (str == "None") return MyBitEnum::None;

llvm::SmallVector<llvm::StringRef, 2> symbols;
str.split(symbols, "|");

uint32_t val = 0;
for (auto symbol : symbols) {
auto bit = llvm::StringSwitch<llvm::Optional<uint32_t>>(symbol)
.Case("tagged", 1)
.Case("Bit1", 2)
.Case("Bit2", 4)
.Case("Bit3", 8)
.Default(llvm::None);
if (bit) { val |= *bit; } else { return llvm::None; }
}
return static_cast<MyBitEnum>(val);
}

llvm::Optional<MyBitEnum> symbolizeMyBitEnum(uint32_t value) {
// Special case for all bits unset.
if (value == 0) return MyBitEnum::None;

if (value & ~static_cast<uint32_t>(15u)) return llvm::None;
return static_cast<MyBitEnum>(value);
}


## Debugging Tips ¶

### Run mlir-tblgen to see the generated content ¶

TableGen syntax sometimes can be obscure; reading the generated content can be a very helpful way to understand and debug issues. To build mlir-tblgen, run cmake --build . --target mlir-tblgen in your build directory and find the mlir-tblgen binary in the bin/ subdirectory. All the supported generators can be found via mlir-tblgen --help. For example, --gen-op-decls and --gen-op-defs as explained in Generated C++ code.

To see the generated code, invoke mlir-tblgen with a specific generator by providing include paths via -I. For example,

# To see op C++ class declaration
mlir-tblgen --gen-op-decls -I /path/to/mlir/include /path/to/input/td/file
# To see op C++ class definition
mlir-tblgen --gen-op-defs -I /path/to/mlir/include /path/to/input/td/file
# To see op documentation
mlir-tblgen --gen-dialect-doc -I /path/to/mlir/include /path/to/input/td/file

# To see op interface C++ class declaration
mlir-tblgen --gen-op-interface-decls -I /path/to/mlir/include /path/to/input/td/file
# To see op interface C++ class definition
mlir-tblgen --gen-op-interface-defs -I /path/to/mlir/include /path/to/input/td/file
# To see op interface documentation
mlir-tblgen --gen-op-interface-doc -I /path/to/mlir/include /path/to/input/td/file


## Appendix ¶

### Reporting deprecation ¶

Classes/defs can be marked as deprecated by using the Deprecate helper class, e.g.,

def OpTraitA : NativeOpTrait<"OpTraitA">, Deprecated<"use bar instead">;


would result in marking OpTraitA as deprecated and mlir-tblgen can emit a warning (default) or error (depending on -on-deprecated flag) to make deprecated state known.

### Requirements and existing mechanisms analysis ¶

The op description should be as declarative as possible to allow a wide range of tools to work with them and query methods generated from them. In particular this means specifying traits, constraints and shape inference information in a way that is easily analyzable (e.g., avoid opaque calls to C++ functions where possible).

We considered the approaches of several contemporary systems and focused on requirements that were desirable:

• Ops registered using a registry separate from C++ code.

• Unknown ops are allowed in MLIR, so ops need not be registered. The ability of the compiler to optimize those ops or graphs containing those ops is constrained but correct.
• The current proposal does not include a runtime op description, but it does not preclude such description, it can be added later.
• The op registry is essential for generating C++ classes that make manipulating ops, verifying correct construction etc. in C++ easier by providing a typed representation and accessors.
• The op registry will be defined in TableGen and be used to generate C++ classes and utility functions (builder/verifier/parser/printer).

• TableGen is a modelling specification language used by LLVM’s backends and fits in well with trait-based modelling. This is an implementation decision and there are alternative ways of doing this. But the specification language is good for the requirements of modelling the traits (as seen from usage in LLVM processor backend modelling) and easy to extend, so a practical choice. If another good option comes up, we will consider it.
• MLIR allows both defined and undefined ops.

• Defined ops should have fixed semantics and could have a corresponding reference implementation defined.
• Dialects are under full control of the dialect owner and normally live with the framework of the dialect.
• The op’s traits (e.g., commutative) are modelled along with the op in the registry.

• The op’s operand/return type constraints are modelled along with the op in the registry (see Shape inference discussion below), this allows (e.g.) optimized concise syntax in textual dumps.

• Behavior of the op is documented along with the op with a summary and a description. The description is written in markdown and extracted for inclusion in the generated LangRef section of the dialect.

• The generic assembly form of printing and parsing is available as normal, but a custom parser and printer can either be specified or automatically generated from an optional string representation showing the mapping of the “assembly” string to operands/type.

• Parser-level remappings (e.g., eq to enum) will be supported as part of the parser generation.
• Matching patterns are specified separately from the op description.

• Contrasted with LLVM there is no “base” set of ops that every backend needs to be aware of. Instead there are many different dialects and the transformations/legalizations between these dialects form a graph of transformations.
• Reference implementation may be provided along with the op definition.

• The reference implementation may be in terms of either standard ops or other reference implementations.

TODO: document expectation if the dependent op’s definition changes.