[all-commits] [llvm/llvm-project] 571195: [mlir][Tensor] Add rewrites to extract slices thro...

[Chris via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 5711957875738c1318f89afd7bf4be388f85a087
      https://github.com/llvm/llvm-project/commit/5711957875738c1318f89afd7bf4be388f85a087
  Author: Christopher Bate <cbate at nvidia.com>
  Date:   2022-09-02 (Fri, 02 Sep 2022)

  Changed paths:
    M mlir/include/mlir/Dialect/Tensor/IR/TensorOps.td
    A mlir/include/mlir/Dialect/Tensor/Transforms/TransformUtils.h
    M mlir/include/mlir/Dialect/Utils/ReshapeOpsUtils.h
    M mlir/include/mlir/Interfaces/ViewLikeInterface.h
    M mlir/lib/Dialect/Tensor/IR/TensorOps.cpp
    M mlir/lib/Dialect/Tensor/Transforms/CMakeLists.txt
    A mlir/lib/Dialect/Tensor/Transforms/ExtractSliceFromReshape.cpp
    M mlir/lib/Dialect/Utils/CMakeLists.txt
    M mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp
    M mlir/lib/Interfaces/ViewLikeInterface.cpp
    A mlir/test/Dialect/Tensor/extract-slice-from-collapse-shape.mlir
    M mlir/test/lib/Dialect/Tensor/CMakeLists.txt
    M mlir/test/lib/Dialect/Tensor/TestTensorTransforms.cpp

  Log Message:
  -----------
  [mlir][Tensor] Add rewrites to extract slices through `tensor.collape_shape`

This change adds a set of utilities to replace the result of a
`tensor.collapse_shape -> tensor.extract_slice` chain with the
equivalent result formed by aggregating slices of the
`tensor.collapse_shape` source. In general, it is not possible to
commute `extract_slice` and `collapse_shape` if linearized dimensions
are sliced. The i-th dimension of the `tensor.collapse_shape`
result is a "linearized sliced dimension" if:

1) Reassociation indices of tensor.collapse_shape in the i'th position
   is greater than size 1 (multiple dimensions of the input are collapsed)
2) The i-th dimension is sliced by `tensor.extract_slice`.

We can work around this by stitching together the result of
`tensor.extract_slice` by iterating over any linearized sliced dimensions.
This is equivalent to "tiling" the linearized-and-sliced dimensions of
the `tensor.collapse_shape` operation in order to manifest the result
tile (the result of the `tensor.extract_slice`). The user of the
utilities must provide the mechanism to create the tiling (e.g. a loop).
In the tests, it is demonstrated how to apply the utilities using either
`scf.for` or `scf.foreach_thread`.

The below example illustrates the pattern using `scf.for`:

```
%0 = linalg.generic ... -> tensor<3x7x11x10xf32>
%1 = tensor.collapse_shape %0 [[0, 1, 2], [3]] : ... to tensor<341x10xf32>
%2 = tensor.extract_slice %1 [13, 0] [10, 10] [2, 1] : .... tensor<10x10xf32>
```

We can construct %2 by generating the following IR:

```
%dest = linalg.init_tensor() : tensor<10x10xf32>
%2 = scf.for %iv = %c0 to %c10 step %c1 iter_args(%arg0) -> tensor<10x10xf32> {
   // Step 1: Map this output idx (%iv) to a multi-index for the input (%3):
   %linear_index = affine.apply affine_map<(d0)[]->(d0*2 + 11)>(%iv)
   %3:3 = arith.delinearize_index %iv into (3, 7, 11)
   // Step 2: Extract the slice from the input
   %4 = tensor.extract_slice %0 [%3#0, %3#1, %3#2, 0] [1, 1, 1, 10] [1, 1, 1, 1] :
         tensor<3x7x11x10xf32> to tensor<1x1x1x10xf32>
   %5 = tensor.collapse_shape %4 [[0, 1, 2], [3]] :
         tensor<1x1x1x10xf32> into tensor<1x10xf32>
   // Step 3: Insert the slice into the destination
   %6 = tensor.insert_slice %5 into %arg0 [%iv, 0] [1, 10] [1, 1] :
         tensor<1x10xf32> into tensor<10x10xf32>
   scf.yield %6 : tensor<10x10xf32>
}
```

The pattern was discussed in the RFC here: https://discourse.llvm.org/t/rfc-tensor-extracting-slices-from-tensor-collapse-shape/64034

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D129699