[mlir] Match before rewrite in BubbleUpPackOpThroughGenericOp #126946
+40
−12
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
Signed-off-by: Max Dawkins <[email protected]>
Member
|
@llvm/pr-subscribers-mlir-linalg Author: None (Max191) ChangesThe BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic. Full diff: https://github.com/llvm/llvm-project/pull/126946.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp b/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
index d79399b6588be..c906f3bdcc632 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
@@ -399,6 +399,18 @@ bubbleUpPackOpThroughGenericOp(RewriterBase &rewriter, tensor::PackOp packOp,
if (!genericOp->getResult(0).hasOneUse())
return failure();
+ // TODO: Add an option for allowing padding values. It could introduce
+ // undefined behavior if we unconditionally propagate pack op through all
+ // the ops. E.g., if the padding value is zero and there are division ops in
+ // a generic op. Some values of padding area could be NaN (0/0).
+ if (packOp.getPaddingValue())
+ return failure();
+
+ OpOperand *opOperand = genericOp.getDpsInitOperand(0);
+ auto packInfo = getPackingInfoFromOperand(opOperand, genericOp, packOp);
+ if (failed(packInfo))
+ return failure();
+
// We want to move the pack not the generic.
OpBuilder::InsertionGuard guard(rewriter);
rewriter.setInsertionPoint(genericOp);
@@ -422,18 +434,6 @@ bubbleUpPackOpThroughGenericOp(RewriterBase &rewriter, tensor::PackOp packOp,
return failure();
}
- // TODO: Add an option for allowing padding values. It could introduce
- // undefined behavior if we unconditionally propagate pack op through all
- // the ops. E.g., if the padding value is zero and there are division ops in
- // a generic op. Some values of padding area could be NaN (0/0).
- if (packOp.getPaddingValue())
- return failure();
-
- OpOperand *opOperand = genericOp.getDpsInitOperand(0);
- auto packInfo = getPackingInfoFromOperand(opOperand, genericOp, packOp);
- if (failed(packInfo))
- return failure();
-
// Rebuild the indexing map for the corresponding init operand.
auto [packedOutOperand, packedOutIndexingMap] =
getOrCreatePackedViewOfOperand(rewriter, genericOp.getLoc(), *packInfo,
diff --git a/mlir/test/Dialect/Linalg/data-layout-propagation.mlir b/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
index cb8064411bbae..b2b29b2b2fee2 100644
--- a/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
+++ b/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
@@ -46,6 +46,34 @@ func.func @dynamic_elem_pack(%arg0: tensor<?x?xf32>, %dest: tensor<?x?x8x2xf32>)
// -----
+#map0 = affine_map<(d0, d1) -> (d0, d1)>
+func.func @dynamic_elem_pack_padding_value(%arg0: tensor<?x?xf32>, %dest: tensor<?x?x8x2xf32>) -> tensor<?x?x8x2xf32>
+{
+ %c0 = arith.constant 0 : index
+ %c1 = arith.constant 1 : index
+ %cst = arith.constant 3.000000e+00 : f32
+ %0 = tensor.dim %arg0, %c0 : tensor<?x?xf32>
+ %1 = tensor.dim %arg0, %c1 : tensor<?x?xf32>
+ %2 = tensor.empty(%0, %1) : tensor<?x?xf32>
+ %3 = linalg.generic {indexing_maps = [#map0, #map0], iterator_types = ["parallel", "parallel"]}
+ ins(%arg0 : tensor<?x?xf32>)
+ outs(%2 : tensor<?x?xf32>) {
+ ^bb0(%arg3: f32, %arg4: f32):
+ %4 = arith.addf %arg3, %arg3 : f32
+ linalg.yield %4 : f32
+ } -> tensor<?x?xf32>
+ %4 = tensor.pack %3 padding_value(%cst : f32)
+ inner_dims_pos = [0, 1]
+ inner_tiles = [8, 2]
+ into %dest : tensor<?x?xf32> -> tensor<?x?x8x2xf32>
+ return %4 : tensor<?x?x8x2xf32>
+}
+// CHECK-LABEL: func.func @dynamic_elem_pack_padding_value
+// CHECK: %[[GENERIC:.+]] = linalg.generic
+// CHECK: tensor.pack %[[GENERIC]]
+
+// -----
+
#map0 = affine_map<(d0, d1) -> (d0, d1)>
func.func @elem_pack_transpose_inner_dims(%arg0: tensor<128x256xi32>, %dest: tensor<4x16x16x32xi32>) -> tensor<4x16x16x32xi32>{
%init = tensor.empty() : tensor<128x256xi32>
|
Member
|
@llvm/pr-subscribers-mlir Author: None (Max191) ChangesThe BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic. Full diff: https://github.com/llvm/llvm-project/pull/126946.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp b/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
index d79399b6588be..c906f3bdcc632 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/DataLayoutPropagation.cpp
@@ -399,6 +399,18 @@ bubbleUpPackOpThroughGenericOp(RewriterBase &rewriter, tensor::PackOp packOp,
if (!genericOp->getResult(0).hasOneUse())
return failure();
+ // TODO: Add an option for allowing padding values. It could introduce
+ // undefined behavior if we unconditionally propagate pack op through all
+ // the ops. E.g., if the padding value is zero and there are division ops in
+ // a generic op. Some values of padding area could be NaN (0/0).
+ if (packOp.getPaddingValue())
+ return failure();
+
+ OpOperand *opOperand = genericOp.getDpsInitOperand(0);
+ auto packInfo = getPackingInfoFromOperand(opOperand, genericOp, packOp);
+ if (failed(packInfo))
+ return failure();
+
// We want to move the pack not the generic.
OpBuilder::InsertionGuard guard(rewriter);
rewriter.setInsertionPoint(genericOp);
@@ -422,18 +434,6 @@ bubbleUpPackOpThroughGenericOp(RewriterBase &rewriter, tensor::PackOp packOp,
return failure();
}
- // TODO: Add an option for allowing padding values. It could introduce
- // undefined behavior if we unconditionally propagate pack op through all
- // the ops. E.g., if the padding value is zero and there are division ops in
- // a generic op. Some values of padding area could be NaN (0/0).
- if (packOp.getPaddingValue())
- return failure();
-
- OpOperand *opOperand = genericOp.getDpsInitOperand(0);
- auto packInfo = getPackingInfoFromOperand(opOperand, genericOp, packOp);
- if (failed(packInfo))
- return failure();
-
// Rebuild the indexing map for the corresponding init operand.
auto [packedOutOperand, packedOutIndexingMap] =
getOrCreatePackedViewOfOperand(rewriter, genericOp.getLoc(), *packInfo,
diff --git a/mlir/test/Dialect/Linalg/data-layout-propagation.mlir b/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
index cb8064411bbae..b2b29b2b2fee2 100644
--- a/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
+++ b/mlir/test/Dialect/Linalg/data-layout-propagation.mlir
@@ -46,6 +46,34 @@ func.func @dynamic_elem_pack(%arg0: tensor<?x?xf32>, %dest: tensor<?x?x8x2xf32>)
// -----
+#map0 = affine_map<(d0, d1) -> (d0, d1)>
+func.func @dynamic_elem_pack_padding_value(%arg0: tensor<?x?xf32>, %dest: tensor<?x?x8x2xf32>) -> tensor<?x?x8x2xf32>
+{
+ %c0 = arith.constant 0 : index
+ %c1 = arith.constant 1 : index
+ %cst = arith.constant 3.000000e+00 : f32
+ %0 = tensor.dim %arg0, %c0 : tensor<?x?xf32>
+ %1 = tensor.dim %arg0, %c1 : tensor<?x?xf32>
+ %2 = tensor.empty(%0, %1) : tensor<?x?xf32>
+ %3 = linalg.generic {indexing_maps = [#map0, #map0], iterator_types = ["parallel", "parallel"]}
+ ins(%arg0 : tensor<?x?xf32>)
+ outs(%2 : tensor<?x?xf32>) {
+ ^bb0(%arg3: f32, %arg4: f32):
+ %4 = arith.addf %arg3, %arg3 : f32
+ linalg.yield %4 : f32
+ } -> tensor<?x?xf32>
+ %4 = tensor.pack %3 padding_value(%cst : f32)
+ inner_dims_pos = [0, 1]
+ inner_tiles = [8, 2]
+ into %dest : tensor<?x?xf32> -> tensor<?x?x8x2xf32>
+ return %4 : tensor<?x?x8x2xf32>
+}
+// CHECK-LABEL: func.func @dynamic_elem_pack_padding_value
+// CHECK: %[[GENERIC:.+]] = linalg.generic
+// CHECK: tensor.pack %[[GENERIC]]
+
+// -----
+
#map0 = affine_map<(d0, d1) -> (d0, d1)>
func.func @elem_pack_transpose_inner_dims(%arg0: tensor<128x256xi32>, %dest: tensor<4x16x16x32xi32>) -> tensor<4x16x16x32xi32>{
%init = tensor.empty() : tensor<128x256xi32>
|
flovent
pushed a commit
to flovent/llvm-project
that referenced
this pull request
Feb 13, 2025
…26946) The BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic. Signed-off-by: Max Dawkins <[email protected]>
joaosaffran
pushed a commit
to joaosaffran/llvm-project
that referenced
this pull request
Feb 14, 2025
…26946) The BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic. Signed-off-by: Max Dawkins <[email protected]>
sivan-shani
pushed a commit
to sivan-shani/llvm-project
that referenced
this pull request
Feb 24, 2025
…26946) The BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic. Signed-off-by: Max Dawkins <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
The BubbleUpPackOpThroughGenericOp pattern had some unsafe rewrites happening before matching was fully complete, which causes the pattern rewriter to fail to converge. This PR fixes the bug by moving all matching logic to before the rewrite logic.