`getDestinationOperands` was almost a duplicate of `DestinationStyleOpInterface::getOutputOperands`. Now that the interface has been moved to mlir/Interfaces, it is no longer needed.
Differential Revision: https://reviews.llvm.org/D136240
This is the second (and final) step of making "destination style" usable without depending on the Linalg dialect. (The first step was D135129.)
This change allows us to provide default bufferization implementations for all destination-style ops. It also allows us to simplify `TilingInterface`. (E.g., `getDestinationOperands` can be removed.)
Differential Revision: https://reviews.llvm.org/D136179
This patch implements the vectorization of tensor.extract for the
basic 1-d lookup case. It only vectorizes the tensor.extract to a
vector.gather when the op extracts value from an 1-d tensor.
Related discussion: https://github.com/iree-org/iree/issues/9198
Reviewed By: dcaballe
Differential Revision: https://reviews.llvm.org/D133786
`getIteratorTypesArray` should be used instead. It's a better substitute for all the current usages of the interface.
The current `ArrayAttr iterator_types()` has a few problems:
* It creates an assumption operation has iterators types as an attribute, but it's not always the case. Sometime iterator types can be inferred from other attribute, or they're just static.
* ArrayAttr is an obscure contained and required extracting values in the client code.
* Makes it hard to migrate iterator types from strings to enums ([RFC](https://discourse.llvm.org/t/rfc-enumattr-for-iterator-types-in-linalg/64535/9)).
Concrete ops, like `linalg.generic` will still have iterator types as an attribute if needed.
As a side effect, this change helps a bit with migration to prefixed accessors.
Differential Revision: https://reviews.llvm.org/D135765
Context: https://discourse.llvm.org/t/psa-retire-linalg-filter-based-patterns/63785
Uses of `LinalgTilingPattern::returningMatchAndRewrite` are replaced by a top-level `tileWithLinalgTilingOptions` function that is marked obsolete and serves
as a temporary means to transition away from `LinalgTilingOptions`-based tiling.
LinalgTilingOptions supports too many options that have been orthogonalized with the use of the transform dialect.
Additionally, the revision introduces a `transform.structured.tile_to_scf_for` structured transform operation that is needed to properly tile `tensor.pad`
via the TilingInterface. Uses of `transform.structured.tile` will be deprecated and replaced by this new op.
This will achieve the deprecation of `linalg::tileLinalgOp`.
Context: https://discourse.llvm.org/t/psa-retire-tileandfuselinalgops-method/63850
In the process of transitioning, tests that were performing tile and distribute on tensors are retired: transformations should be orthogonalized better in the future.
In particular, tiling to specific loop types and tileAndDistribute behavior are not available via the transform ops.
The behavior is still available as part of the `tileWithLinalgTilingOptions` method to allow downstream clients to transition without breakages but is meant to be retired soon.
As more tests are ported to the transform dialect, it became necessary to introduce a test-transform-dialect-erase-schedule-pass to discard the transform specification
once applied so that e2e lowering and execution is possible.
Lastly, a number of redundant tests that were testing composition of patterns are retired as they are available with a better mechanism via the transform dialect.
Differential Revision: https://reviews.llvm.org/D135573
Context: https://discourse.llvm.org/t/psa-retire-linalg-filter-based-patterns/63785
In the process, also retire `tileConsumerAndFuseProducers` that is now replaced by `tileConsumerAndFuseProducerGreedilyUsingSCFForOp`.
Context: https://discourse.llvm.org/t/psa-retire-tileandfuselinalgops-method/63850
When performing this replacement, a change of behavior appeared: the older `tileConsumerAndFuseProducers` would split the parallel
and non-parallel dimensions automatically and perform a first level of tile-and-fuse on parallel dimensions only and then introduce a
second level of tiling-only on the reduction dimensions. The newer `tileConsumerAndFuseProducerGreedilyUsingSCFForOp` on the other hand
does not perform this breakdown. As a consequence, the transform specification is evolved to produce the same output.
Additionally, replace some uses of `unsigned` by `int64_t` where possible without pulling in larger interface changes (left for a future PR).
Context: https://www.youtube.com/watch?v=Puio5dly9N8
Lastly, tests that were performing tile and fuse and distribute on tensors are retired: the generated IR mixing scf.for, tensors and
distributed processor ids was racy at best ..
Differential Revision: https://reviews.llvm.org/D135559
This patch exposes the method to check if an op can be vectorized or
not for downstream uses. Also adds a check to mark elementwise operations
that have non-vectorizable ops (like `tensor.extract`) as non vectorizable.
Reviewed By: nicolasvasilache, dcaballe, ThomasRaoux
Differential Revision: https://reviews.llvm.org/D135201
tensor.empty/linalg.init_tensor produces an uninititalized tensor that can be used as a destination operand for destination-style ops (ops that implement `DestinationStyleOpInterface`).
This change makes it possible to implement `TilingInterface` for non-destination-style ops without depending on the Linalg dialect.
RFC: https://discourse.llvm.org/t/rfc-add-tensor-from-shape-operation/65101
Differential Revision: https://reviews.llvm.org/D135129
This is much more explicit, and prevents annoying conflicts with op
specific accessors (which may have a different contract). This is similar
to the past rename of getType -> getFunctionType,
Fixes#58030
Differential Revision: https://reviews.llvm.org/D135007
Previously, splitReduction transformation added the split parallel dimension
*before* the reduction dimension, leading to tiling for reduction. This
commit creates an option to create the parallel dimension *after* the
reduction dimension, allowing us to transform the op into vertical reduction
with SIMD parallelism.
Reviewed By: ThomasRaoux, dcaballe
Differential Revision: https://reviews.llvm.org/D134764
Summary:
Most of the code that gets `iterator_types` from LinalgInterface is forced to
extract values from an `Attribute`. As a result, the usage pattern looks like
this:
```
SmallVector<StringRef> iterators = llvm::to_vector<4>(linalgOp.iterator_types().getAsValueRange<StringAttr>());
```
It also forces all operations that implement LinalgOp interface to have
`iterator_types` attribute even when the information can be easily infered from
other parameters.
In perfect future, `getIteratorTypeArray` should be the only method to get
iterator types from the interface. The default implementation can rely on
`iterator_types` attribute though.
The name `getIteratorTypeArray` was picked to be consistent with existing
`getIndexingMapsArray`.
This patch add a few sample usages. More cleanups will follow.
Differential Revision: https://reviews.llvm.org/D134729
Improve hoisting logic to support cases where the read being hoisted
comes from a transfer_write with disjoint indices.
Differential Revision: https://reviews.llvm.org/D134624
This patch fixes:
mlir/lib/Dialect/Linalg/Transforms/TilingInterfaceImpl.cpp:99:16:
error: 'getValue' is deprecated: Use value
instead. [-Werror,-Wdeprecated-declarations]
Summary:
Use the new enum in TilingIterface and verify that `iterator_type` attribute in
LinalgOp interface is compatible with the enum values. Later IteratorType enum
will be used in LinalgInterface to replace the current `iterator_type` attribute
array of string.
Existing enums in Linalg are moved into a separate td file and tablegen build
target. This is necessary, have one I32EnumAttr in a shared space that generated
enum class definition and EnumAttrs is dialect-specific location. Otherwise
there might be a conflict that I32EnumAttr generates enum definitions in
multiple places.
Differential Revision: https://reviews.llvm.org/D134634
This commit adds a pattern to swap
```
tensor.extract_slice(linalg.fill(%cst, %init))
```
into
```
linalg.fill(%cst, tensor.extract_slice(%init))
```
when the linalg.fill op have no other users.
This helps to reduce the fill footprint.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D134102
Custom attributes can be set on the operation. It prevents them to be
removed when doing named ops conversion.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D133892
This revision retires the LinalgCodegenStrategy vectorization pattern. Please see the context: https://discourse.llvm.org/t/psa-retire-linalg-filter-based-patterns/63785.
This revision improves the transform dialect's VectorizeOp in different ways below:
- Adds LinalgDialect as a dependent dialect. When `transform.structured.vectorize` vectorizes `tensor.pad`, it generates `linalg.init_tensor`. In this case, linalg dialect must be registered.
- Inserts CopyVectorizationPattern in order to vectorize `memref.copy`.
- Creates two attributes: `disable_multi_reduction_to_contract_patterns` and `disable_transfer_permutation_map_lowering_patterns`. They are limiting the power of vectorization and are currently intended for testing purposes.
It also removes some of the "CHECK: vector.transfer_write" in the vectorization.mlir test. They are redundant writes, at the end of the code there is a rewrite to the same place. Transform dialect no longer generates them.
Depends on D133684 that retires the LinalgCodegenStrategy vectorization pass.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D133699
We retire linalg's strategy vectorize pass. Our goal is to use transform dialect instead of passes.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D133684
This patch fixes three warnings of the form:
mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp:1436:5: error:
default label in switch which covers all enumeration values
[-Werror,-Wcovered-switch-default]
Windows build requires brackets on switch-cases that initializes
variables.
Reviewed By: hanchung
Differential Revision: https://reviews.llvm.org/D133889
Most computer vision torch models uses nchw/ncw convolution. In a previous patch we added decomposition conv2dNchw to conv1dNcw. To enhance the performance on torch models we add this vectorization pattern for conv1dNcw which would consquently also improve the performance on conv2dNchw.
On IREE + Intel Xeon 8360 + Resnet50, we were able to get ~7x speed up ~880ms to 126ms.
Reviewed By: nicolasvasilache, hanchung
Differential Revision: https://reviews.llvm.org/D133675
This is the first step in replacing interator_type from strings with enums in Vector and Linalg dialect. This change adds IteratorTypeAttr and uses it in ContractionOp.
To avoid breaking all the tests, print/parse code has conversion between string and enum for now.
There is a shared code in StructuredOpsUtils.h that expects iterator types to be strings. To break this dependancy, this change forks helper function `isParallelIterator` and `isReductionIterator` to utils in both dialects and adds `getIteratorTypeNames()` to support backward compatibility with StructuredGenerator.
In the later changes, I plan to add a similar enum attribute to Linalg.
Differential Revision: https://reviews.llvm.org/D133696
The tiling interface generates the order of parallel_insert_slice incorrectly when there are multiple destionation operands. This revision fixes that and adds a test for it. It depends on D132937
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D133204
This revision relaxes constraint of tiling when there are multiple destination operands. It also adds a test.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D132937
