[mlir][tensor] Simplify pad-like tensor pack and unpack #92388
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Extend existing tensor patterns to simplify pad-like tensor pack/unpack into expand/collapse shape operations.
|
@llvm/pr-subscribers-mlir Author: Adam Siemieniuk (adam-smnk) ChangesExtend existing tensor patterns to simplify pad-like tensor pack/unpack into expand/collapse shape operations. Full diff: https://github.com/llvm/llvm-project/pull/92388.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp b/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
index ebcb34e9ef024..5d6e3ec9756af 100644
--- a/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
+++ b/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
@@ -91,7 +91,8 @@ struct SimplifyPackToExpandShape : public OpRewritePattern<PackOp> {
RankedTensorType sourceType = packOp.getSourceType();
if (failed(isPackOnInnerMostDim(rewriter, packOp)) &&
failed(isPackOn1D(rewriter, packOp, sourceType.getShape(),
- packOp.getStaticTiles()))) {
+ packOp.getStaticTiles())) &&
+ !packOp.isLikePad()) {
return failure();
}
@@ -152,7 +153,8 @@ struct SimplifyUnPackToCollapseShape : public OpRewritePattern<UnPackOp> {
RankedTensorType destType = unpackOp.getDestType();
if (failed(isUnpackOnInnerMostDim(rewriter, unpackOp)) &&
failed(isPackOn1D(rewriter, unpackOp, destType.getShape(),
- unpackOp.getStaticTiles()))) {
+ unpackOp.getStaticTiles())) &&
+ !unpackOp.isLikeUnPad()) {
return failure();
}
diff --git a/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir b/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
index 5a2eade0ecccf..0ad93863dc501 100644
--- a/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
+++ b/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
@@ -266,3 +266,51 @@ func.func @unpack_16x1x1x2_to_32x1(%arg0 : tensor<16x1x1x2xf32>) -> tensor<32x1x
: tensor<16x1x1x2xf32> -> tensor<32x1xf32>
return %unpack : tensor<32x1xf32>
}
+
+// -----
+
+// CHECK-LABEL: func.func @pad_like_pack(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<32x64xf32>)
+// CHECK: %[[EXPANDED:.+]] = tensor.expand_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] output_shape [1, 1, 32, 64] : tensor<32x64xf32> into tensor<1x1x32x64xf32>
+// CHECK: return %[[EXPANDED]] : tensor<1x1x32x64xf32>
+func.func @pad_like_pack(%arg0: tensor<32x64xf32>) -> tensor<1x1x32x64xf32> {
+ %empty = tensor.empty() : tensor<1x1x32x64xf32>
+ %0 = tensor.pack %arg0 inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<32x64xf32> -> tensor<1x1x32x64xf32>
+ return %0 : tensor<1x1x32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @pad_like_pack_with_outer_dims_perm(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<32x64xf32>)
+// CHECK: %[[EXPANDED:.+]] = tensor.expand_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] output_shape [1, 1, 32, 64] : tensor<32x64xf32> into tensor<1x1x32x64xf32>
+// CHECK: return %[[EXPANDED]] : tensor<1x1x32x64xf32>
+func.func @pad_like_pack_with_outer_dims_perm(%arg0: tensor<32x64xf32>) -> tensor<1x1x32x64xf32> {
+ %empty = tensor.empty() : tensor<1x1x32x64xf32>
+ %0 = tensor.pack %arg0 outer_dims_perm = [1, 0] inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<32x64xf32> -> tensor<1x1x32x64xf32>
+ return %0 : tensor<1x1x32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @unpad_like_unpack(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<1x1x32x64xf32>)
+// CHECK: %[[COLLAPSED:.+]] = tensor.collapse_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] : tensor<1x1x32x64xf32> into tensor<32x64xf32>
+// CHECK: return %[[COLLAPSED]]
+func.func @unpad_like_unpack(%arg0: tensor<1x1x32x64xf32>) -> tensor<32x64xf32> {
+ %empty = tensor.empty() : tensor<32x64xf32>
+ %0 = tensor.unpack %arg0 inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<1x1x32x64xf32> -> tensor<32x64xf32>
+ return %0 : tensor<32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @unpad_like_unpack_with_outer_dims_perm(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<1x1x32x64xf32>)
+// CHECK: %[[COLLAPSED:.+]] = tensor.collapse_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] : tensor<1x1x32x64xf32> into tensor<32x64xf32>
+// CHECK: return %[[COLLAPSED]]
+func.func @unpad_like_unpack_with_outer_dims_perm(%arg0: tensor<1x1x32x64xf32>) -> tensor<32x64xf32> {
+ %empty = tensor.empty() : tensor<32x64xf32>
+ %0 = tensor.unpack %arg0 outer_dims_perm = [1, 0] inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<1x1x32x64xf32> -> tensor<32x64xf32>
+ return %0 : tensor<32x64xf32>
+}
|
|
@llvm/pr-subscribers-mlir-tensor Author: Adam Siemieniuk (adam-smnk) ChangesExtend existing tensor patterns to simplify pad-like tensor pack/unpack into expand/collapse shape operations. Full diff: https://github.com/llvm/llvm-project/pull/92388.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp b/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
index ebcb34e9ef024..5d6e3ec9756af 100644
--- a/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
+++ b/mlir/lib/Dialect/Tensor/Transforms/PackAndUnpackPatterns.cpp
@@ -91,7 +91,8 @@ struct SimplifyPackToExpandShape : public OpRewritePattern<PackOp> {
RankedTensorType sourceType = packOp.getSourceType();
if (failed(isPackOnInnerMostDim(rewriter, packOp)) &&
failed(isPackOn1D(rewriter, packOp, sourceType.getShape(),
- packOp.getStaticTiles()))) {
+ packOp.getStaticTiles())) &&
+ !packOp.isLikePad()) {
return failure();
}
@@ -152,7 +153,8 @@ struct SimplifyUnPackToCollapseShape : public OpRewritePattern<UnPackOp> {
RankedTensorType destType = unpackOp.getDestType();
if (failed(isUnpackOnInnerMostDim(rewriter, unpackOp)) &&
failed(isPackOn1D(rewriter, unpackOp, destType.getShape(),
- unpackOp.getStaticTiles()))) {
+ unpackOp.getStaticTiles())) &&
+ !unpackOp.isLikeUnPad()) {
return failure();
}
diff --git a/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir b/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
index 5a2eade0ecccf..0ad93863dc501 100644
--- a/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
+++ b/mlir/test/Dialect/Tensor/simplify-pack-unpack.mlir
@@ -266,3 +266,51 @@ func.func @unpack_16x1x1x2_to_32x1(%arg0 : tensor<16x1x1x2xf32>) -> tensor<32x1x
: tensor<16x1x1x2xf32> -> tensor<32x1xf32>
return %unpack : tensor<32x1xf32>
}
+
+// -----
+
+// CHECK-LABEL: func.func @pad_like_pack(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<32x64xf32>)
+// CHECK: %[[EXPANDED:.+]] = tensor.expand_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] output_shape [1, 1, 32, 64] : tensor<32x64xf32> into tensor<1x1x32x64xf32>
+// CHECK: return %[[EXPANDED]] : tensor<1x1x32x64xf32>
+func.func @pad_like_pack(%arg0: tensor<32x64xf32>) -> tensor<1x1x32x64xf32> {
+ %empty = tensor.empty() : tensor<1x1x32x64xf32>
+ %0 = tensor.pack %arg0 inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<32x64xf32> -> tensor<1x1x32x64xf32>
+ return %0 : tensor<1x1x32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @pad_like_pack_with_outer_dims_perm(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<32x64xf32>)
+// CHECK: %[[EXPANDED:.+]] = tensor.expand_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] output_shape [1, 1, 32, 64] : tensor<32x64xf32> into tensor<1x1x32x64xf32>
+// CHECK: return %[[EXPANDED]] : tensor<1x1x32x64xf32>
+func.func @pad_like_pack_with_outer_dims_perm(%arg0: tensor<32x64xf32>) -> tensor<1x1x32x64xf32> {
+ %empty = tensor.empty() : tensor<1x1x32x64xf32>
+ %0 = tensor.pack %arg0 outer_dims_perm = [1, 0] inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<32x64xf32> -> tensor<1x1x32x64xf32>
+ return %0 : tensor<1x1x32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @unpad_like_unpack(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<1x1x32x64xf32>)
+// CHECK: %[[COLLAPSED:.+]] = tensor.collapse_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] : tensor<1x1x32x64xf32> into tensor<32x64xf32>
+// CHECK: return %[[COLLAPSED]]
+func.func @unpad_like_unpack(%arg0: tensor<1x1x32x64xf32>) -> tensor<32x64xf32> {
+ %empty = tensor.empty() : tensor<32x64xf32>
+ %0 = tensor.unpack %arg0 inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<1x1x32x64xf32> -> tensor<32x64xf32>
+ return %0 : tensor<32x64xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func.func @unpad_like_unpack_with_outer_dims_perm(
+// CHECK-SAME: %[[ARG0:.+]]: tensor<1x1x32x64xf32>)
+// CHECK: %[[COLLAPSED:.+]] = tensor.collapse_shape %[[ARG0]] {{\[}}[0, 1, 2], [3]] : tensor<1x1x32x64xf32> into tensor<32x64xf32>
+// CHECK: return %[[COLLAPSED]]
+func.func @unpad_like_unpack_with_outer_dims_perm(%arg0: tensor<1x1x32x64xf32>) -> tensor<32x64xf32> {
+ %empty = tensor.empty() : tensor<32x64xf32>
+ %0 = tensor.unpack %arg0 outer_dims_perm = [1, 0] inner_dims_pos = [0, 1] inner_tiles = [32, 64] into %empty : tensor<1x1x32x64xf32> -> tensor<32x64xf32>
+ return %0 : tensor<32x64xf32>
+}
|
|
@chelini Any comments? |
chelini
reviewed
May 22, 2024
chelini
approved these changes
May 23, 2024
hanhanW
approved these changes
May 23, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Extend existing tensor patterns to simplify pad-like tensor pack/unpack into expand/collapse shape operations.