[mlir][arith] Refine the verifier for arith.constant #87999
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
Disallows initialization of scalable vectors with an attribute of arbitrary values, e.g.: ```mlir %c = arith.constant dense<[0, 1]> : vector<[2] x i32> ``` Initialization using vector splats remains allowed (i.e. when all the init values are identical): ```mlir %c = arith.constant dense<[1, 1]> : vector<[2] x i32> ``` Note: This is a re-upload of llvm#86178
|
@llvm/pr-subscribers-mlir-arith Author: Andrzej Warzyński (banach-space) ChangesDisallows initialization of scalable vectors with an attribute of %c = arith.constant dense<[0, 1]> : vector<[2] x i32>Initialization using vector splats remains allowed (i.e. when all the %c = arith.constant dense<[1, 1]> : vector<[2] x i32>Note: This is a re-upload of #86178 Full diff: https://github.com/llvm/llvm-project/pull/87999.diff 3 Files Affected:
diff --git a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
index 1d68a4f7292b53..6f995b93bc3ecd 100644
--- a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
+++ b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
@@ -213,6 +213,15 @@ LogicalResult arith::ConstantOp::verify() {
return emitOpError(
"value must be an integer, float, or elements attribute");
}
+
+ // Note, we could relax this for vectors with 1 scalable dim, e.g.:
+ // * arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ // However, this would most likely require updating the lowerings to LLVM.
+ auto vecType = dyn_cast<VectorType>(type);
+ if (vecType && vecType.isScalable() && !isa<SplatElementsAttr>(getValue()))
+ return emitOpError(
+ "intializing scalable vectors with elements attribute is not supported"
+ " unless it's a vector splat");
return success();
}
diff --git a/mlir/test/Dialect/Arith/invalid.mlir b/mlir/test/Dialect/Arith/invalid.mlir
index 6d8ac0ada52be3..ada849220bb839 100644
--- a/mlir/test/Dialect/Arith/invalid.mlir
+++ b/mlir/test/Dialect/Arith/invalid.mlir
@@ -64,6 +64,24 @@ func.func @constant_out_of_range() {
// -----
+func.func @constant_invalid_scalable_1d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[0, 1]> : vector<[2] x i32>
+ return
+}
+
+// -----
+
+func.func @constant_invalid_scalable_2d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ return
+}
+
+// -----
+
func.func @constant_wrong_type() {
^bb:
%x = "arith.constant"(){value = 10.} : () -> f32 // expected-error {{'arith.constant' op failed to verify that all of {value, result} have same type}}
diff --git a/mlir/test/Dialect/Vector/linearize.mlir b/mlir/test/Dialect/Vector/linearize.mlir
index 212541c79565b6..22be78cd682057 100644
--- a/mlir/test/Dialect/Vector/linearize.mlir
+++ b/mlir/test/Dialect/Vector/linearize.mlir
@@ -153,14 +153,3 @@ func.func @test_0d_vector() -> vector<f32> {
// ALL: return %[[CST]]
return %0 : vector<f32>
}
-
-// -----
-
-func.func @test_scalable_no_linearize(%arg0: vector<2x[2]xf32>) -> vector<2x[2]xf32> {
- // expected-error@+1 {{failed to legalize operation 'arith.constant' that was explicitly marked illegal}}
- %0 = arith.constant dense<[[1., 1.], [3., 3.]]> : vector<2x[2]xf32>
- %1 = math.sin %arg0 : vector<2x[2]xf32>
- %2 = arith.addf %0, %1 : vector<2x[2]xf32>
-
- return %2 : vector<2x[2]xf32>
-}
|
|
@llvm/pr-subscribers-mlir Author: Andrzej Warzyński (banach-space) ChangesDisallows initialization of scalable vectors with an attribute of %c = arith.constant dense<[0, 1]> : vector<[2] x i32>Initialization using vector splats remains allowed (i.e. when all the %c = arith.constant dense<[1, 1]> : vector<[2] x i32>Note: This is a re-upload of #86178 Full diff: https://github.com/llvm/llvm-project/pull/87999.diff 3 Files Affected:
diff --git a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
index 1d68a4f7292b53..6f995b93bc3ecd 100644
--- a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
+++ b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
@@ -213,6 +213,15 @@ LogicalResult arith::ConstantOp::verify() {
return emitOpError(
"value must be an integer, float, or elements attribute");
}
+
+ // Note, we could relax this for vectors with 1 scalable dim, e.g.:
+ // * arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ // However, this would most likely require updating the lowerings to LLVM.
+ auto vecType = dyn_cast<VectorType>(type);
+ if (vecType && vecType.isScalable() && !isa<SplatElementsAttr>(getValue()))
+ return emitOpError(
+ "intializing scalable vectors with elements attribute is not supported"
+ " unless it's a vector splat");
return success();
}
diff --git a/mlir/test/Dialect/Arith/invalid.mlir b/mlir/test/Dialect/Arith/invalid.mlir
index 6d8ac0ada52be3..ada849220bb839 100644
--- a/mlir/test/Dialect/Arith/invalid.mlir
+++ b/mlir/test/Dialect/Arith/invalid.mlir
@@ -64,6 +64,24 @@ func.func @constant_out_of_range() {
// -----
+func.func @constant_invalid_scalable_1d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[0, 1]> : vector<[2] x i32>
+ return
+}
+
+// -----
+
+func.func @constant_invalid_scalable_2d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ return
+}
+
+// -----
+
func.func @constant_wrong_type() {
^bb:
%x = "arith.constant"(){value = 10.} : () -> f32 // expected-error {{'arith.constant' op failed to verify that all of {value, result} have same type}}
diff --git a/mlir/test/Dialect/Vector/linearize.mlir b/mlir/test/Dialect/Vector/linearize.mlir
index 212541c79565b6..22be78cd682057 100644
--- a/mlir/test/Dialect/Vector/linearize.mlir
+++ b/mlir/test/Dialect/Vector/linearize.mlir
@@ -153,14 +153,3 @@ func.func @test_0d_vector() -> vector<f32> {
// ALL: return %[[CST]]
return %0 : vector<f32>
}
-
-// -----
-
-func.func @test_scalable_no_linearize(%arg0: vector<2x[2]xf32>) -> vector<2x[2]xf32> {
- // expected-error@+1 {{failed to legalize operation 'arith.constant' that was explicitly marked illegal}}
- %0 = arith.constant dense<[[1., 1.], [3., 3.]]> : vector<2x[2]xf32>
- %1 = math.sin %arg0 : vector<2x[2]xf32>
- %2 = arith.addf %0, %1 : vector<2x[2]xf32>
-
- return %2 : vector<2x[2]xf32>
-}
|
|
@llvm/pr-subscribers-mlir-vector Author: Andrzej Warzyński (banach-space) ChangesDisallows initialization of scalable vectors with an attribute of %c = arith.constant dense<[0, 1]> : vector<[2] x i32>Initialization using vector splats remains allowed (i.e. when all the %c = arith.constant dense<[1, 1]> : vector<[2] x i32>Note: This is a re-upload of #86178 Full diff: https://github.com/llvm/llvm-project/pull/87999.diff 3 Files Affected:
diff --git a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
index 1d68a4f7292b53..6f995b93bc3ecd 100644
--- a/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
+++ b/mlir/lib/Dialect/Arith/IR/ArithOps.cpp
@@ -213,6 +213,15 @@ LogicalResult arith::ConstantOp::verify() {
return emitOpError(
"value must be an integer, float, or elements attribute");
}
+
+ // Note, we could relax this for vectors with 1 scalable dim, e.g.:
+ // * arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ // However, this would most likely require updating the lowerings to LLVM.
+ auto vecType = dyn_cast<VectorType>(type);
+ if (vecType && vecType.isScalable() && !isa<SplatElementsAttr>(getValue()))
+ return emitOpError(
+ "intializing scalable vectors with elements attribute is not supported"
+ " unless it's a vector splat");
return success();
}
diff --git a/mlir/test/Dialect/Arith/invalid.mlir b/mlir/test/Dialect/Arith/invalid.mlir
index 6d8ac0ada52be3..ada849220bb839 100644
--- a/mlir/test/Dialect/Arith/invalid.mlir
+++ b/mlir/test/Dialect/Arith/invalid.mlir
@@ -64,6 +64,24 @@ func.func @constant_out_of_range() {
// -----
+func.func @constant_invalid_scalable_1d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[0, 1]> : vector<[2] x i32>
+ return
+}
+
+// -----
+
+func.func @constant_invalid_scalable_2d_vec_initialization() {
+^bb0:
+ // expected-error@+1 {{'arith.constant' op intializing scalable vectors with elements attribute is not supported unless it's a vector splat}}
+ %c = arith.constant dense<[[3, 3], [1, 1]]> : vector<2 x [2] x i32>
+ return
+}
+
+// -----
+
func.func @constant_wrong_type() {
^bb:
%x = "arith.constant"(){value = 10.} : () -> f32 // expected-error {{'arith.constant' op failed to verify that all of {value, result} have same type}}
diff --git a/mlir/test/Dialect/Vector/linearize.mlir b/mlir/test/Dialect/Vector/linearize.mlir
index 212541c79565b6..22be78cd682057 100644
--- a/mlir/test/Dialect/Vector/linearize.mlir
+++ b/mlir/test/Dialect/Vector/linearize.mlir
@@ -153,14 +153,3 @@ func.func @test_0d_vector() -> vector<f32> {
// ALL: return %[[CST]]
return %0 : vector<f32>
}
-
-// -----
-
-func.func @test_scalable_no_linearize(%arg0: vector<2x[2]xf32>) -> vector<2x[2]xf32> {
- // expected-error@+1 {{failed to legalize operation 'arith.constant' that was explicitly marked illegal}}
- %0 = arith.constant dense<[[1., 1.], [3., 3.]]> : vector<2x[2]xf32>
- %1 = math.sin %arg0 : vector<2x[2]xf32>
- %2 = arith.addf %0, %1 : vector<2x[2]xf32>
-
- return %2 : vector<2x[2]xf32>
-}
|
banach-space
added a commit
to banach-space/llvm-project
that referenced
this pull request
Apr 8, 2024
This PR adds two small convenience Vector types:
* `ScalableVectorType` and `FixedWidthVectorType`.
The goal of these new types is two-fold:
* enable idiomatic checks like `isa<ScalableVectorType>(...)`,
* make the split into "Scalable" and "Fixed-wdith" vectors a bit more
explicit and more visible in the code-base.
Depends on llvm#87999
kuhar
approved these changes
Apr 8, 2024
banach-space
added a commit
to banach-space/llvm-project
that referenced
this pull request
Apr 8, 2024
This PR adds two small convenience Vector types:
* `ScalableVectorType` and `FixedWidthVectorType`.
The goal of these new types is two-fold:
* enable idiomatic checks like `isa<ScalableVectorType>(...)`,
* make the split into "Scalable" and "Fixed-wdith" vectors a bit more
explicit and more visible in the code-base.
Depends on llvm#87999
banach-space
added a commit
to banach-space/llvm-project
that referenced
this pull request
Apr 16, 2024
This PR adds two small convenience Vector types:
* `ScalableVectorType` and `FixedWidthVectorType`.
The goal of these new types is two-fold:
* enable idiomatic checks like `isa<ScalableVectorType>(...)`,
* make the split into "Scalable" and "Fixed-wdith" vectors a bit more
explicit and more visible in the code-base.
Depends on llvm#87999
banach-space
added a commit
to banach-space/llvm-project
that referenced
this pull request
Nov 29, 2024
This PR adds two small convenience Vector types:
* `ScalableVectorType` and `FixedWidthVectorType`.
The goal of these new types is two-fold:
* enable idiomatic checks like `isa<ScalableVectorType>(...)`,
* make the split into "Scalable" and "Fixed-wdith" vectors a bit more
explicit and more visible in the code-base.
Depends on llvm#87999
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.
Disallows initialization of scalable vectors with an attribute of
arbitrary values, e.g.:
Initialization using vector splats remains allowed (i.e. when all the
init values are identical):
Note: This is a re-upload of #86178