[mlir] Add ElementwiseMappable trait and apply it to std elementwise ops.

This patch adds an `ElementwiseMappable` trait as discussed in the RFC
here:
https://llvm.discourse.group/t/rfc-std-elementw

[mlir] Add ElementwiseMappable trait and apply it to std elementwise ops.

This patch adds an `ElementwiseMappable` trait as discussed in the RFC
here:
https://llvm.discourse.group/t/rfc-std-elementwise-ops-on-tensors/2113/23

This trait can power a number of transformations and analyses.
A subsequent patch adds a convert-elementwise-to-linalg pass exhibits
how this trait allows writing generic transformations.
See https://reviews.llvm.org/D90354 for that patch.

This trait slightly changes some verifier messages, but the diagnostics
are usually about as good. I fiddled with the ordering of the trait in
the .td file trait lists to minimize the changes here.

Differential Revision: https://reviews.llvm.org/D90731

show more ...

[mlir] Optimize Op definitions and registration to optimize for code size

This revision refactors the base Op/AbstractOperation classes to reduce the amount of generated code size when defining a ne

[mlir] Optimize Op definitions and registration to optimize for code size

This revision refactors the base Op/AbstractOperation classes to reduce the amount of generated code size when defining a new operation. The current scheme involves taking the address of functions defined directly on Op and Trait classes. This is problematic because even when these functions are empty/unused we still result in these functions being defined in the main executable. In this revision, we switch to using SFINAE and template type filtering to remove remove functions that are not needed/used. For example, if an operation does not define a custom `print` method we shouldn't define a templated `printAssembly` method for it. The same applies to parsing/folding/verification/etc. This dropped MLIR code size for a large downstream library by ~10%(~1 mb in an opt build).

Differential Revision: https://reviews.llvm.org/D90196

show more ...

[MLIR] Add idempotent trait folding

This trait simply adds a fold of f(f(x)) = f(x) when an operation is labelled as idempotent

Reviewed By: rriddle, andyly

Differential Revision: https://reviews.

[MLIR] Add idempotent trait folding

This trait simply adds a fold of f(f(x)) = f(x) when an operation is labelled as idempotent

Reviewed By: rriddle, andyly

Differential Revision: https://reviews.llvm.org/D89421

show more ...

Remove unused SideEffectInterfaces header

This change removes an unnecessary header introduced in
https://github.com/llvm/llvm-project/commit/c0b3abd19a3e.

Differential Revision: https://reviews.ll

Remove unused SideEffectInterfaces header

This change removes an unnecessary header introduced in
https://github.com/llvm/llvm-project/commit/c0b3abd19a3e.

Differential Revision: https://reviews.llvm.org/D89347

show more ...

[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait

This is the same diff as https://reviews.llvm.org/D88809/ except side effect
free check is r

[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait

This is the same diff as https://reviews.llvm.org/D88809/ except side effect
free check is removed for involution and a FIXME is added until the dependency
is resolved for shared builds. The old diff has more details on possible fixes.

Reviewed By: rriddle, andyly

Differential Revision: https://reviews.llvm.org/D89333

show more ...

[mlir] Fix bug in computing operation order

When attempting to compute a differential orderIndex we were calculating the
bailout condition correctly, but then an errant "+ 1" meant the orderIndex we

[mlir] Fix bug in computing operation order

When attempting to compute a differential orderIndex we were calculating the
bailout condition correctly, but then an errant "+ 1" meant the orderIndex we
created was invalid.

Added test.

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D89115

show more ...

Revert "[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait"

This reverts commit 1ceaffd95a6bdc4b7d2193e049bcd6b40ee9ff50.

The build is broken

Revert "[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait"

This reverts commit 1ceaffd95a6bdc4b7d2193e049bcd6b40ee9ff50.

The build is broken with -DBUILD_SHARED_LIBS=ON ; seems like a possible
layering issue to investigate:

tools/mlir/lib/IR/CMakeFiles/obj.MLIRIR.dir/Operation.cpp.o: In function `mlir::MemoryEffectOpInterface::hasNoEffect(mlir::Operation*)':
Operation.cpp:(.text._ZN4mlir23MemoryEffectOpInterface11hasNoEffectEPNS_9OperationE[_ZN4mlir23MemoryEffectOpInterface11hasNoEffectEPNS_9OperationE]+0x9c): undefined reference to `mlir::MemoryEffectOpInterface::getEffects(llvm::SmallVectorImpl<mlir::SideEffects::EffectInstance<mlir::MemoryEffects::Effect> >&)'

show more ...

[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait

This change allows folds to be done on a newly introduced involution trait rather than havin

[MLIR] Add a foldTrait() mechanism to allow traits to define folding and test it with an Involution trait

This change allows folds to be done on a newly introduced involution trait rather than having to manually rewrite this optimization for every instance of an involution

Reviewed By: rriddle, andyly, stephenneuendorffer

Differential Revision: https://reviews.llvm.org/D88809

show more ...

[MLIR] Change Operation::create() methods to use Value/Type/Block ranges.

- Introduce a new BlockRange class to represent range of blocks (constructible from
an ArrayRef<Block *> or a SuccessorRan

[MLIR] Change Operation::create() methods to use Value/Type/Block ranges.

- Introduce a new BlockRange class to represent range of blocks (constructible from
an ArrayRef<Block *> or a SuccessorRange);
- Change Operation::create() methods to use TypeRange for result types, ValueRange for
operands and BlockRange for successors.

Differential Revision: https://reviews.llvm.org/D86985

show more ...

Store an Identifier instead of a StringRef for the OperationName inside an AbstractOperation (NFC)

Instead of storing a StringRef, we keep an Identifier which otherwise requires a lock on the contex

Store an Identifier instead of a StringRef for the OperationName inside an AbstractOperation (NFC)

Instead of storing a StringRef, we keep an Identifier which otherwise requires a lock on the context to retrieve.
This will allow to get an Identifier for any registered Operation for "free".

Reviewed By: rriddle

Differential Revision: https://reviews.llvm.org/D86994

show more ...

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.

This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.

To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.

1) For passes, you need to override the method:

virtual void getDependentDialects(DialectRegistry &registry) const {}

and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.

2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.

3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:

mlir::DialectRegistry registry;
registry.insert<mlir::standalone::StandaloneDialect>();
registry.insert<mlir::StandardOpsDialect>();

Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:

mlir::registerAllDialects(registry);

4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()

Differential Revision: https://reviews.llvm.org/D85622

show more ...

Revert "Separate the Registration from Loading dialects in the Context"

This reverts commit d14cf45735b0d09d7d3caf0824779520dd20ef10.
The build is broken with GCC-5.

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.

This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.

To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.

1) For passes, you need to override the method:

virtual void getDependentDialects(DialectRegistry &registry) const {}

and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.

2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.

3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:

mlir::DialectRegistry registry;
registry.insert<mlir::standalone::StandaloneDialect>();
registry.insert<mlir::StandardOpsDialect>();

Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:

mlir::registerAllDialects(registry);

4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()

Differential Revision: https://reviews.llvm.org/D85622

show more ...

Revert "Separate the Registration from Loading dialects in the Context"

This reverts commit e1de2b75501e5eaf8777bd5248382a7c55a44fd6.
Broke a build bot.

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.

This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.

To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.

1) For passes, you need to override the method:

virtual void getDependentDialects(DialectRegistry &registry) const {}

and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.

2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.

3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:

mlir::DialectRegistry registry;
mlir::registerDialect<mlir::standalone::StandaloneDialect>();
mlir::registerDialect<mlir::StandardOpsDialect>();

Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:

mlir::registerAllDialects(registry);

4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()

show more ...

Fix method name to start with lower case to match style guide (NFC)

Revert "Separate the Registration from Loading dialects in the Context"

This reverts commit 20563933875a9396c8ace9c9770ecf6a988c4ea6.

Build is broken on a few bots

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects are only loaded explicitly on demand:
- the Parser is lazily loading Dialects in the context as it encounters them during parsing. This is the only purpose for registering dialects and not load them in the context.
- Passes are expected to declare the dialects they will create entity from (Operations, Attributes, or Types), and the PassManager is loading Dialects into the Context when starting a pipeline.

This changes simplifies the configuration of the registration: a compiler only need to load the dialect for the IR it will emit, and the optimizer is self-contained and load the required Dialects. For example in the Toy tutorial, the compiler only needs to load the Toy dialect in the Context, all the others (linalg, affine, std, LLVM, ...) are automatically loaded depending on the optimization pipeline enabled.

Differential Revision: https://reviews.llvm.org/D85622

show more ...

Revert "Separate the Registration from Loading dialects in the Context"

This was landed by accident, will reland with the right comments
addressed from the reviews.
Also revert dependent build fixes.

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects

Separate the Registration from Loading dialects in the Context

This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects are only loaded explicitly on demand:
- the Parser is lazily loading Dialects in the context as it encounters them during parsing. This is the only purpose for registering dialects and not load them in the context.
- Passes are expected to declare the dialects they will create entity from (Operations, Attributes, or Types), and the PassManager is loading Dialects into the Context when starting a pipeline.

This changes simplifies the configuration of the registration: a compiler only need to load the dialect for the IR it will emit, and the optimizer is self-contained and load the required Dialects. For example in the Toy tutorial, the compiler only needs to load the Toy dialect in the Context, all the others (linalg, affine, std, LLVM, ...) are automatically loaded depending on the optimization pipeline enabled.

show more ...

Remove DialectHooks and introduce a Dialect Interfaces instead

These hooks were introduced before the Interfaces mechanism was available.

DialectExtractElementHook is unused and entirely removed. T

Remove DialectHooks and introduce a Dialect Interfaces instead

These hooks were introduced before the Interfaces mechanism was available.

DialectExtractElementHook is unused and entirely removed. The
DialectConstantFoldHook is used a fallback in the
operation fold() method, and is replaced by a DialectInterface.
The DialectConstantDecodeHook is used for interpreting OpaqueAttribute
and should be revamped, but is replaced with an interface in 1:1 fashion
for now.

Differential Revision: https://reviews.llvm.org/D85595

show more ...

[MLIR] Add argument related API to Region

- Arguments of the first block of a region are considered region arguments.
- Add API on Region class to deal with these arguments directly instead of
usi

[MLIR] Add argument related API to Region

- Arguments of the first block of a region are considered region arguments.
- Add API on Region class to deal with these arguments directly instead of
using the front() block.
- Changed several instances of existing code that can use this API
- Fixes https://bugs.llvm.org/show_bug.cgi?id=46535

Differential Revision: https://reviews.llvm.org/D83599

show more ...

[mlir][NFC] Remove usernames and google bug numbers from TODO comments.

These were largely leftover from when MLIR was a google project, and don't really follow LLVM guidelines.

[MLIR] Add a NoRegionArguments trait

- This trait will verify that all regions attached to an Op have no arguments
- Fixes https://bugs.llvm.org/show_bug.cgi?id=46521 : Add trait NoRegionArguments

[MLIR] Add a NoRegionArguments trait

- This trait will verify that all regions attached to an Op have no arguments
- Fixes https://bugs.llvm.org/show_bug.cgi?id=46521 : Add trait NoRegionArguments

Differential Revision: https://reviews.llvm.org/D83016

show more ...

[MLIR] Add an Op util which returns its name with the dialect stripped.

Differential Revision: https://reviews.llvm.org/D81435