[mlir][tensor] Fix getReassociationForCollapse for tensor/scalar re…
#144118
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
…shapes Commit 6e5a142 changed the behavior of the function when computing reassociations between tensors (consisting of unit/dynamic dimensions) and scalars/0d vectors. The IR representation for such reshapes actually expects an empty reassociation, like so: ``` func.func @example(%arg0 : tensor<?x?x?xf32>) -> tensor<f32> { %0 = tensor.collapse_shape %arg0 [] : tensor<?x?x?xf32> into tensor<f32> } ``` Restore the original behavior - the routine should resort to reporting failures when compile time-known non-unit dimensions are part of the attempted reassociation. Signed-off-by: Artem Gindinson <[email protected]>
IanWood1
reviewed
Jun 13, 2025
Comment on lines
305
to
313
| if (numTargetDims == 0) { | ||
| ReassociationIndices allSourceIndices; | ||
| allSourceIndices.reserve(numSourceDims); | ||
| for (unsigned sourceDimIdx = 0; sourceDimIdx < numSourceDims; | ||
| ++sourceDimIdx) { | ||
| int64_t sourceSize = sourceShape[sourceDimIdx]; | ||
| // All source dimensions must be unit or dynamic. | ||
| if (sourceSize != 1 && sourceSize != ShapedType::kDynamic) | ||
| return std::nullopt; | ||
| allSourceIndices.push_back(sourceDimIdx); | ||
| } | ||
| return SmallVector<ReassociationIndices>{allSourceIndices}; | ||
| return SmallVector<ReassociationIndices>{}; | ||
| } |
There was a problem hiding this comment.
It might be cleaner to use an llvm::all_of/any_of but that's a matter of preference.
There was a problem hiding this comment.
Ah sorry, missed the comment. Good call, any_of would be preferable :D
|
@llvm/pr-subscribers-mlir Author: Artem Gindinson (AGindinson) Changes…shapes Commit 6e5a142 changed the behavior of the function when computing reassociations between tensors (consisting of unit/dynamic dimensions) and scalars/0d vectors. The IR representation for such reshapes actually expects an empty reassociation, like so: Restore the original behavior - the routine should resort to reporting failures when compile time-known non-unit dimensions are part of the attempted reassociation. Full diff: https://github.com/llvm/llvm-project/pull/144118.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp b/mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp
index 3b1fdb69e8ef1..aa566c0086a2f 100644
--- a/mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp
+++ b/mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp
@@ -299,19 +299,17 @@ mlir::getReassociationIndicesForCollapse(ArrayRef<int64_t> sourceShape,
// this utility).
if (numSourceDims <= numTargetDims)
return std::nullopt;
- // Early handling for scalar target types.
+ // Early handling for scalar target types. We should report an invalid
+ // reassociation for non-unit static dimensions - no chance to collapse these
+ // into a scalar.
if (numTargetDims == 0) {
- ReassociationIndices allSourceIndices;
- allSourceIndices.reserve(numSourceDims);
for (unsigned sourceDimIdx = 0; sourceDimIdx < numSourceDims;
++sourceDimIdx) {
int64_t sourceSize = sourceShape[sourceDimIdx];
- // All source dimensions must be unit or dynamic.
if (sourceSize != 1 && sourceSize != ShapedType::kDynamic)
return std::nullopt;
- allSourceIndices.push_back(sourceDimIdx);
}
- return SmallVector<ReassociationIndices>{allSourceIndices};
+ return SmallVector<ReassociationIndices>{};
}
// Collect source ranges by iterating over the target shape left-to-right.
diff --git a/mlir/unittests/Dialect/Utils/ReshapeOpsUtilsTest.cpp b/mlir/unittests/Dialect/Utils/ReshapeOpsUtilsTest.cpp
index db1a87a4de2d5..05f97e875e2dc 100644
--- a/mlir/unittests/Dialect/Utils/ReshapeOpsUtilsTest.cpp
+++ b/mlir/unittests/Dialect/Utils/ReshapeOpsUtilsTest.cpp
@@ -23,16 +23,16 @@ makeOptionalIndices(std::initializer_list<ReassociationIndices> list) {
TEST(ReassociationIndicesForCollapse, ScalarTest) {
EXPECT_EQ(getReassociationIndicesForCollapse({1}, {}),
- makeOptionalIndices({{0}}));
+ makeOptionalIndices({}));
EXPECT_EQ(getReassociationIndicesForCollapse({1, 1}, {}),
- makeOptionalIndices({{0, 1}}));
+ makeOptionalIndices({}));
EXPECT_EQ(getReassociationIndicesForCollapse({ShapedType::kDynamic}, {}),
- makeOptionalIndices({{0}}));
+ makeOptionalIndices({}));
EXPECT_EQ(getReassociationIndicesForCollapse({1, ShapedType::kDynamic,
ShapedType::kDynamic, 1,
ShapedType::kDynamic},
{}),
- makeOptionalIndices({{0, 1, 2, 3, 4}}));
+ makeOptionalIndices({}));
}
TEST(ReassociationIndicesForCollapse, ScalarTestFailure) {
|
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/157/builds/30749 Here is the relevant piece of the build log for the reference |
|
At a glance, the failure feels completely unrelated. |
tomtor
pushed a commit
to tomtor/llvm-project
that referenced
this pull request
Jun 14, 2025
llvm#144118) …shapes Commit 6e5a142 changed the behavior of the function when computing reassociations between tensors (consisting of unit/dynamic dimensions) and scalars/0d vectors. The IR representation for such reshapes actually expects an empty reassociation, like so: ``` func.func @example(%arg0 : tensor<?x?x?xf32>) -> tensor<f32> { %0 = tensor.collapse_shape %arg0 [] : tensor<?x?x?xf32> into tensor<f32> } ``` Restore the original behavior - the routine should resort to reporting failures when compile time-known non-unit dimensions are part of the attempted reassociation. Signed-off-by: Artem Gindinson <[email protected]>
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.
…shapes
Commit 6e5a142 changed the behavior of the function when computing reassociations between tensors (consisting of unit/dynamic dimensions) and scalars/0d vectors. The IR representation for such reshapes actually expects an empty reassociation, like so:
Restore the original behavior - the routine should resort to reporting failures when compile time-known non-unit dimensions are part of the attempted reassociation.