mlir.dialects._ml_program_ops_gen

Attributes

_ods_ir

Classes

_Dialect
FuncOp	This simple function container represents callables in an ML program where
GlobalLoadConstOp	Loads a constant (immutable) value from a global directly by symbol.
GlobalLoadGraphOp	Performs a non-atomic, non-volatile, non-synchronized load from a global
GlobalLoadOp	Performs a non-atomic, non-volatile, non-synchronized load from a global
GlobalOp	Declares a named global variable (or constant).
GlobalStoreGraphOp	Performs a non-atomic, non-volatile, non-synchronized store to a mutable
GlobalStoreOp	Performs a non-atomic, non-volatile, non-synchronized store to a mutable
OutputOp	The output operation terminates a subgraph by yielding values
ReturnOp	The return operation terminates a func function by yielding values
SubgraphOp	This simple function container represents callables in an ML program where
TokenOp	Token values are used to chain side effecting ops in a graph so as to

Functions

func(→ FuncOp)
global_load_const(→ _ods_ir)
global_load_graph(→ _ods_ir)
global_load(→ _ods_ir)
global_(→ GlobalOp)
global_store_graph(→ _ods_ir)
global_store(→ GlobalStoreOp)
output(→ OutputOp)
return_(→ ReturnOp)
subgraph(→ SubgraphOp)
token(→ _ods_ir)

Module Contents

mlir.dialects._ml_program_ops_gen._ods_ir

class mlir.dialects._ml_program_ops_gen._Dialect(descriptor: object)

Bases: _ods_ir

DIALECT_NAMESPACE = 'ml_program'

class mlir.dialects._ml_program_ops_gen.FuncOp(sym_name, function_type, *, arg_attrs=None, res_attrs=None, sym_visibility=None, loc=None, ip=None)

Bases: _ods_ir

This simple function container represents callables in an ML program where the body is an SSACFG region. It must be terminated by a return op which yields values with the same arity and types as the FunctionType results of the containing func.

This op is a Symbol but does not introduce a new SymbolTable. As such, it cannot represent nested symbols.

Example:

ml_program.func private @some_extern(i32) -> i32
ml_program.func @compute(%arg0 : i32) -> i32 {
  ml_program.return %arg0 : i32
}

OPERATION_NAME = 'ml_program.func'

_ODS_REGIONS = (1, True)

sym_name() → _ods_ir

function_type() → _ods_ir

arg_attrs() → _ods_ir | None

res_attrs() → _ods_ir | None

sym_visibility() → _ods_ir | None

body() → _ods_ir

mlir.dialects._ml_program_ops_gen.func(sym_name, function_type, *, arg_attrs=None, res_attrs=None, sym_visibility=None, loc=None, ip=None) → FuncOp

class mlir.dialects._ml_program_ops_gen.GlobalLoadConstOp(result, global_, *, loc=None, ip=None)

Bases: _ods_ir

Loads a constant (immutable) value from a global directly by symbol.

This op is only legal for globals that are not mutable and exists because such a load can be considered to have no side effects.

Example:

%0 = ml_program.global_load_const @foobar : tensor<?xi32>

OPERATION_NAME = 'ml_program.global_load_const'

_ODS_REGIONS = (0, True)

global_() → _ods_ir

result() → _ods_ir

Shortcut to get an op result if it has only one (throws an error otherwise).

mlir.dialects._ml_program_ops_gen.global_load_const(result, global_, *, loc=None, ip=None) → _ods_ir

class mlir.dialects._ml_program_ops_gen.GlobalLoadGraphOp(result, produceToken, global_, consumeTokens, *, loc=None, ip=None)

Bases: _ods_ir

Performs a non-atomic, non-volatile, non-synchronized load from a global that may be mutable.

It is fully expected that these constraints are not suitable for all situations, and alternative ops should be defined and used for more advanced cases.

This op is side effecting and may not be valid to use in graph regions without additional consideration to evaluation order constraints.

Example:

%0, %cstr = ml_program.global_load_graph @foobar
  ordering (%token -> !ml_program.token) : tensor<?xi32>

OPERATION_NAME = 'ml_program.global_load_graph'

_ODS_REGIONS = (0, True)

consumeTokens() → _ods_ir

global_() → _ods_ir

result() → _ods_ir

Shortcut to get an op result if it has only one (throws an error otherwise).

produceToken() → _ods_ir

mlir.dialects._ml_program_ops_gen.global_load_graph(result, produce_token, global_, consume_tokens, *, loc=None, ip=None) → _ods_ir

class mlir.dialects._ml_program_ops_gen.GlobalLoadOp(result, global_, *, loc=None, ip=None)

Bases: _ods_ir

Performs a non-atomic, non-volatile, non-synchronized load from a global that may be mutable.

It is fully expected that these constraints are not suitable for all situations, and alternative ops should be defined and used for more advanced cases.

This op is side effecting and may not be valid to use in graph regions without additional consideration to evaluation order constraints. See global_load_graph for op which allows for explicit ordering constraints.

Example:

%0 = ml_program.global_load @foobar : tensor<?xi32>

OPERATION_NAME = 'ml_program.global_load'

_ODS_REGIONS = (0, True)

global_() → _ods_ir

result() → _ods_ir

Shortcut to get an op result if it has only one (throws an error otherwise).

mlir.dialects._ml_program_ops_gen.global_load(result, global_, *, loc=None, ip=None) → _ods_ir

class mlir.dialects._ml_program_ops_gen.GlobalOp(sym_name, type_, *, is_mutable=None, value=None, sym_visibility=None, loc=None, ip=None)

Bases: _ods_ir

Declares a named global variable (or constant).

A global contains a value of a specified type which can be accessed at runtime via appropriate load/store operations. It can be mutable or constant, optionally taking an initial value or declared as extern (in which case, the initial value is found in external storage by symbol name).

Generally, the type of the global and the type of the initial value will be the same. However, for type hierarchies which can have a more generalized bounding type that can be assigned from a narrow type, this is allowed (but not verified).

Examples:

// Constant global.
ml_program.global @foobar(dense<4> : tensor<4xi32>) : tensor<?xi32>

// Constant with external linkage.
ml_program.global mutable @foobar(#ml_program.extern<tensor<4xi32>>)
  : tensor<?xi32>

// Mutable global with an undefined initial value.
ml_program.global mutable @foobar : tensor<?xi32>

OPERATION_NAME = 'ml_program.global'

_ODS_REGIONS = (0, True)

sym_name() → _ods_ir

type_() → _ods_ir

is_mutable() → bool

value() → _ods_ir | None

sym_visibility() → _ods_ir | None

mlir.dialects._ml_program_ops_gen.global_(sym_name, type_, *, is_mutable=None, value=None, sym_visibility=None, loc=None, ip=None) → GlobalOp

class mlir.dialects._ml_program_ops_gen.GlobalStoreGraphOp(produceToken, global_, value, consumeTokens, *, loc=None, ip=None)

Bases: _ods_ir

Performs a non-atomic, non-volatile, non-synchronized store to a mutable global.

It is fully expected that these constraints are not suitable for all situations, and alternative ops should be defined and used for more advanced cases.

This op is side effecting and may not be valid to use in graph regions without additional consideration to evaluation order constraints.

Example:

%token = ml_program.global_store @foobar = %0 : tensor<?xi32>
  ordering (%in_token -> !ml_program.token) : tensor<?xi32>

OPERATION_NAME = 'ml_program.global_store_graph'

_ODS_REGIONS = (0, True)

value() → _ods_ir

consumeTokens() → _ods_ir

global_() → _ods_ir

produceToken() → _ods_ir

mlir.dialects._ml_program_ops_gen.global_store_graph(produce_token, global_, value, consume_tokens, *, loc=None, ip=None) → _ods_ir

class mlir.dialects._ml_program_ops_gen.GlobalStoreOp(global_, value, *, loc=None, ip=None)

Bases: _ods_ir

Performs a non-atomic, non-volatile, non-synchronized store to a mutable global.

It is fully expected that these constraints are not suitable for all situations, and alternative ops should be defined and used for more advanced cases.

This op is side effecting and may not be valid to use in graph regions without additional consideration to evaluation order constraints. See global_store_graph for op which allows for explicit ordering constraints.

Example:

ml_program.global_store @foobar = %0 : tensor<?xi32>

OPERATION_NAME = 'ml_program.global_store'

_ODS_REGIONS = (0, True)

value() → _ods_ir

global_() → _ods_ir

mlir.dialects._ml_program_ops_gen.global_store(global_, value, *, loc=None, ip=None) → GlobalStoreOp

class mlir.dialects._ml_program_ops_gen.OutputOp(operands_, *, loc=None, ip=None)

Bases: _ods_ir

The output operation terminates a subgraph by yielding values to the caller. The operation takes variable number of operands and produces no results. The operand number and types must match the signature of the function that contains the operation.

OPERATION_NAME = 'ml_program.output'

_ODS_REGIONS = (0, True)

operands_() → _ods_ir

mlir.dialects._ml_program_ops_gen.output(operands_, *, loc=None, ip=None) → OutputOp

class mlir.dialects._ml_program_ops_gen.ReturnOp(operands_, *, loc=None, ip=None)

Bases: _ods_ir

The return operation terminates a func function by yielding values to the caller. The operation takes variable number of operands and produces no results. The operand number and types must match the signature of the function that contains the operation.

OPERATION_NAME = 'ml_program.return'

_ODS_REGIONS = (0, True)

operands_() → _ods_ir

mlir.dialects._ml_program_ops_gen.return_(operands_, *, loc=None, ip=None) → ReturnOp

class mlir.dialects._ml_program_ops_gen.SubgraphOp(sym_name, function_type, *, arg_attrs=None, res_attrs=None, sym_visibility=None, loc=None, ip=None)

Bases: _ods_ir

This simple function container represents callables in an ML program where the body is a Graph region containing a single block. It must be terminated by an output op which yields values with the same arity and types as the FunctionType results of the containing subgraph.

This op is a Symbol but does not introduce a new SymbolTable. As such, it cannot represented nested symbols.

Example:

ml_program.subgraph private @some_extern(i32) -> i32
ml_program.subgraph @compute(%arg0 : i32) -> i32 {
  ml_program.output %arg0 : i32
}

OPERATION_NAME = 'ml_program.subgraph'

_ODS_REGIONS = (1, True)

sym_name() → _ods_ir

function_type() → _ods_ir

arg_attrs() → _ods_ir | None

res_attrs() → _ods_ir | None

sym_visibility() → _ods_ir | None

body() → _ods_ir

mlir.dialects._ml_program_ops_gen.subgraph(sym_name, function_type, *, arg_attrs=None, res_attrs=None, sym_visibility=None, loc=None, ip=None) → SubgraphOp

class mlir.dialects._ml_program_ops_gen.TokenOp(token, *, loc=None, ip=None)

Bases: _ods_ir

Token values are used to chain side effecting ops in a graph so as to establish an execution order. This op produces a token.

OPERATION_NAME = 'ml_program.token'

_ODS_REGIONS = (0, True)

token() → _ods_ir

mlir.dialects._ml_program_ops_gen.token(token, *, loc=None, ip=None) → _ods_ir