[Mlir-commits] [mlir] [mlir] Fix alignment for predicate (i1) vectors (PR #175961)
llvmlistbot at llvm.org
llvmlistbot at llvm.org
Wed Jan 14 06:56:14 PST 2026
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-mlir
Author: Andrzej Warzyński (banach-space)
<details>
<summary>Changes</summary>
Legal scalable predicate vectors (legal in the LLVM sense), e.g.
vector<[16]xi1> (or <vscale x 16 x i1>, using LLVM syntax) ought to have
alignment 2 rather than 16, see e.g. [1].
MLIR currently computes the vector “size in bits” as:
```cpp
vecType.getNumElements()
* dataLayout.getTypeSize(vecType.getElementType()) * 8
```
but `getTypeSize()` returns a size in *bytes* (rounded up from bits), so for
`i1` it returns 1. Multiplying by 8 converts that storage byte back to 8 bits
per element, which overestimates predicate vector sizes.
Instead, use:
```cpp
vecType.getNumElements()
* dataLayout.getTypeSizeInBits(vecType.getElementType())
```
For `vector<[16]xi1>` this changes:
* [before]: `16 * (1 byte * 8) = 128 bits`
* [after]: `16 * 1 bit = 16 bits`
This is a very small update that, based on the available tests, only
affects types like `vector<[16]xi1>`. It aligns MLIR with LLVM, making
sure that the corresponding alignment is 2 rather that 16. For context,
LLVM computes the alignment in this case via `getTypeStoreSize`, which
for `16 x i1` returns 2 bytes. Perhaps MLIR should follow similar path
in the future.
[1] https://developer.arm.com/documentation/ddi0602/2025-12/SVE-Instructions/LDR--predicate---Load-predicate-register-?lang=en
---
Full diff: https://github.com/llvm/llvm-project/pull/175961.diff
2 Files Affected:
- (modified) mlir/lib/Interfaces/DataLayoutInterfaces.cpp (+1-1)
- (modified) mlir/test/Interfaces/DataLayoutInterfaces/query.mlir (+6)
``````````diff
diff --git a/mlir/lib/Interfaces/DataLayoutInterfaces.cpp b/mlir/lib/Interfaces/DataLayoutInterfaces.cpp
index 782384999c70c..a6922ee5f4b5b 100644
--- a/mlir/lib/Interfaces/DataLayoutInterfaces.cpp
+++ b/mlir/lib/Interfaces/DataLayoutInterfaces.cpp
@@ -78,7 +78,7 @@ mlir::detail::getDefaultTypeSizeInBits(Type type, const DataLayout &dataLayout,
if (auto vecType = dyn_cast<VectorType>(type)) {
uint64_t baseSize = vecType.getNumElements() / vecType.getShape().back() *
llvm::PowerOf2Ceil(vecType.getShape().back()) *
- dataLayout.getTypeSize(vecType.getElementType()) * 8;
+ dataLayout.getTypeSizeInBits(vecType.getElementType());
return llvm::TypeSize::get(baseSize, vecType.isScalable());
}
diff --git a/mlir/test/Interfaces/DataLayoutInterfaces/query.mlir b/mlir/test/Interfaces/DataLayoutInterfaces/query.mlir
index 5df32555000ad..97ef8b2a8ae1c 100644
--- a/mlir/test/Interfaces/DataLayoutInterfaces/query.mlir
+++ b/mlir/test/Interfaces/DataLayoutInterfaces/query.mlir
@@ -44,6 +44,12 @@ func.func @no_layout_builtin() {
// CHECK: preferred = 16
// CHECK: size = {minimal_size = 16 : index, scalable}
"test.data_layout_query"() : () -> vector<[4]xi32>
+ // CHECK: alignment = 2
+ // CHECK: bitsize = {minimal_size = 16 : index, scalable}
+ // CHECK: index = 0
+ // CHECK: preferred = 2
+ // CHECK: size = {minimal_size = 2 : index, scalable}
+ "test.data_layout_query"() : () -> vector<[16]xi1>
return
}
``````````
</details>
https://github.com/llvm/llvm-project/pull/175961
More information about the Mlir-commits
mailing list