[mlir][tosa] Don't fold mul with zero lhs/rhs if resulting type is dynamic#153420
Merged
[mlir][tosa] Don't fold mul with zero lhs/rhs if resulting type is dynamic#153420
Conversation
Member
|
@llvm/pr-subscribers-mlir Author: Sayan Saha (sahas3) ChangesCanonicalizing the following IR: resulted in a crash from the folder for Full diff: https://github.com/llvm/llvm-project/pull/153420.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp b/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
index e3cba38871909..fce61f27ca3ea 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
@@ -1120,13 +1120,14 @@ OpFoldResult MulOp::fold(FoldAdaptor adaptor) {
}
if (rhsTy == resultTy) {
- if (isSplatZero(resultETy, lhsAttr))
+ if (isSplatZero(resultETy, lhsAttr) && resultTy.hasStaticShape())
+ // constant values can only be resized if resulting type is static
return lhsAttr.resizeSplat(resultTy);
if (isSplatOne(resultETy, lhsAttr, shift))
return rhs;
}
if (lhsTy == resultTy) {
- if (isSplatZero(resultETy, rhsAttr))
+ if (isSplatZero(resultETy, rhsAttr) && resultTy.hasStaticShape())
return rhsAttr.resizeSplat(resultTy);
if (isSplatOne(resultETy, rhsAttr, shift))
return lhs;
diff --git a/mlir/test/Dialect/Tosa/canonicalize.mlir b/mlir/test/Dialect/Tosa/canonicalize.mlir
index 5150ee36e9e5e..930bb9fe96811 100644
--- a/mlir/test/Dialect/Tosa/canonicalize.mlir
+++ b/mlir/test/Dialect/Tosa/canonicalize.mlir
@@ -565,6 +565,33 @@ func.func @mul_zero_broadcast(%arg0: tensor<2x3xf32>) -> (tensor<2x3xf32>, tenso
// -----
+// CHECK-LABEL: @mul_zero_dynamic_nofold
+// CHECK-SAME: %[[ARG0:.*]]: tensor<?x17xf32>) -> tensor<?x17xf32> {
+// CHECK: %[[ZERO:.*]] = "tosa.const"() <{values = dense<0.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+// CHECK: %[[SHIFT:.*]] = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+// CHECK: %[[MUL:.*]] = tosa.mul %[[ARG0]], %[[ZERO]], %[[SHIFT]] : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+// CHECK: return %[[MUL]]
+func.func @mul_zero_dynamic_nofold(%arg0: tensor<?x17xf32>) -> tensor<?x17xf32> {
+ %0 = "tosa.const"() <{values = dense<0.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+ %1 = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+ %2 = tosa.mul %arg0, %0, %1 : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+ return %2 : tensor<?x17xf32>
+}
+
+// -----
+
+// CHECK-LABEL: @mul_one_dynamic_fold
+// CHECK-SAME: %[[ARG0:.*]]: tensor<?x17xf32>) -> tensor<?x17xf32> {
+// CHECK: return %[[ARG0]]
+func.func @mul_one_dynamic_fold(%arg0: tensor<?x17xf32>) -> tensor<?x17xf32> {
+ %0 = "tosa.const"() <{values = dense<1.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+ %1 = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+ %2 = tosa.mul %arg0, %0, %1 : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+ return %2 : tensor<?x17xf32>
+}
+
+// -----
+
// CHECK-LABEL: @select_same_value
func.func @select_same_value(%arg0: tensor<2x3xi1>, %arg1: tensor<2x3xi32>) -> tensor<2x3xi32> {
%0 = tosa.select %arg0, %arg1, %arg1 : (tensor<2x3xi1>, tensor<2x3xi32>, tensor<2x3xi32>) -> tensor<2x3xi32>
|
Member
|
@llvm/pr-subscribers-mlir-tosa Author: Sayan Saha (sahas3) ChangesCanonicalizing the following IR: resulted in a crash from the folder for Full diff: https://github.com/llvm/llvm-project/pull/153420.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp b/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
index e3cba38871909..fce61f27ca3ea 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp
@@ -1120,13 +1120,14 @@ OpFoldResult MulOp::fold(FoldAdaptor adaptor) {
}
if (rhsTy == resultTy) {
- if (isSplatZero(resultETy, lhsAttr))
+ if (isSplatZero(resultETy, lhsAttr) && resultTy.hasStaticShape())
+ // constant values can only be resized if resulting type is static
return lhsAttr.resizeSplat(resultTy);
if (isSplatOne(resultETy, lhsAttr, shift))
return rhs;
}
if (lhsTy == resultTy) {
- if (isSplatZero(resultETy, rhsAttr))
+ if (isSplatZero(resultETy, rhsAttr) && resultTy.hasStaticShape())
return rhsAttr.resizeSplat(resultTy);
if (isSplatOne(resultETy, rhsAttr, shift))
return lhs;
diff --git a/mlir/test/Dialect/Tosa/canonicalize.mlir b/mlir/test/Dialect/Tosa/canonicalize.mlir
index 5150ee36e9e5e..930bb9fe96811 100644
--- a/mlir/test/Dialect/Tosa/canonicalize.mlir
+++ b/mlir/test/Dialect/Tosa/canonicalize.mlir
@@ -565,6 +565,33 @@ func.func @mul_zero_broadcast(%arg0: tensor<2x3xf32>) -> (tensor<2x3xf32>, tenso
// -----
+// CHECK-LABEL: @mul_zero_dynamic_nofold
+// CHECK-SAME: %[[ARG0:.*]]: tensor<?x17xf32>) -> tensor<?x17xf32> {
+// CHECK: %[[ZERO:.*]] = "tosa.const"() <{values = dense<0.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+// CHECK: %[[SHIFT:.*]] = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+// CHECK: %[[MUL:.*]] = tosa.mul %[[ARG0]], %[[ZERO]], %[[SHIFT]] : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+// CHECK: return %[[MUL]]
+func.func @mul_zero_dynamic_nofold(%arg0: tensor<?x17xf32>) -> tensor<?x17xf32> {
+ %0 = "tosa.const"() <{values = dense<0.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+ %1 = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+ %2 = tosa.mul %arg0, %0, %1 : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+ return %2 : tensor<?x17xf32>
+}
+
+// -----
+
+// CHECK-LABEL: @mul_one_dynamic_fold
+// CHECK-SAME: %[[ARG0:.*]]: tensor<?x17xf32>) -> tensor<?x17xf32> {
+// CHECK: return %[[ARG0]]
+func.func @mul_one_dynamic_fold(%arg0: tensor<?x17xf32>) -> tensor<?x17xf32> {
+ %0 = "tosa.const"() <{values = dense<1.000000e+00> : tensor<1x1xf32>}> : () -> tensor<1x1xf32>
+ %1 = "tosa.const"() <{values = dense<0> : tensor<1xi8>}> : () -> tensor<1xi8>
+ %2 = tosa.mul %arg0, %0, %1 : (tensor<?x17xf32>, tensor<1x1xf32>, tensor<1xi8>) -> tensor<?x17xf32>
+ return %2 : tensor<?x17xf32>
+}
+
+// -----
+
// CHECK-LABEL: @select_same_value
func.func @select_same_value(%arg0: tensor<2x3xi1>, %arg1: tensor<2x3xi32>) -> tensor<2x3xi32> {
%0 = tosa.select %arg0, %arg1, %arg1 : (tensor<2x3xi1>, tensor<2x3xi32>, tensor<2x3xi32>) -> tensor<2x3xi32>
|
lhutton1
approved these changes
Aug 13, 2025
sjarus
approved these changes
Aug 13, 2025
Contributor
sjarus
left a comment
There was a problem hiding this comment.
Dynamic shape handling is nice, thanks!
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
Canonicalizing the following IR:
resulted in a crash
from the folder for
tosa::mulsince the zero value was being reshaped to?x17size which isn't supported. AFAIK,tosa.constrequires all dimensions to be static. So in this case, the fix is to not to fold the op.