Skip to content

[mlir][vector] Replace vector.splat with vector.broadcast in some tests #152230

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 3 commits into from
Aug 7, 2025
Merged

[mlir][vector] Replace vector.splat with vector.broadcast in some tests #152230

merged 3 commits into from
Aug 7, 2025

Conversation

newling
Copy link
Contributor

@newling newling commented Aug 6, 2025 •
edited
Loading

Splat is deprecated, and being prepared for removal in a future release.
https://discourse.llvm.org/t/rfc-mlir-vector-deprecate-then-remove-vector-splat/87143/5

The command I used, catches almost every splat op:

perl -i -pe 's/vector\.splat\s+(\S+)\s*:\s*vector<((?:\[?\d+\]?x)*)\s*([^>]+)>/vector.broadcast $1 : $3 to vector<$2$3>/g' filename

@newling newling marked this pull request as ready for review August 6, 2025 15:51
@llvmbot
Copy link
Member

llvmbot commented Aug 6, 2025 •
edited
Loading

@llvm/pr-subscribers-mlir-arith
@llvm/pr-subscribers-mlir-vector
@llvm/pr-subscribers-mlir-sve
@llvm/pr-subscribers-mlir

@llvm/pr-subscribers-mlir-sme

Author: James Newling (newling)

Changes

Splat is deprecated, and being prepared for removal in a future release.
https://discourse.llvm.org/t/rfc-mlir-vector-deprecate-then-remove-vector-splat/87143/5

The command I used, catches almost every splat op:

perl -i -pe 's/vector\.splat\s+(\S+)\s*:\s*vector&lt;((?:\[?\d+\]?x)*)\s*([^&gt;]+)&gt;/vector.broadcast $1 : $3 to vector&lt;$2$3&gt;/g' filename

Patch is 26.60 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/152230.diff

16 Files Affected:

  • (modified) mlir/test/Dialect/Arith/canonicalize.mlir (+7-7)
  • (modified) mlir/test/Dialect/Vector/vector-sink.mlir (+15-15)
  • (modified) mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir (+4-4)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f32.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f64.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/ArmSME/transfer-write-2d.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/contraction.mlir (+1-1)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/scalable-interleave.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/interleave.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/outerproduct-f32.mlir (+3-3)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/outerproduct-i64.mlir (+3-3)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/transfer-read-1d.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/transfer-read-2d.mlir (+2-2)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/transfer-read-3d.mlir (+1-1)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/transfer-read.mlir (+1-1)
  • (modified) mlir/test/Integration/Dialect/Vector/CPU/transfer-write.mlir (+4-4)
diff --git a/mlir/test/Dialect/Arith/canonicalize.mlir b/mlir/test/Dialect/Arith/canonicalize.mlir
index 3d5a46d13e59d..78f67821da138 100644
--- a/mlir/test/Dialect/Arith/canonicalize.mlir
+++ b/mlir/test/Dialect/Arith/canonicalize.mlir
@@ -654,7 +654,7 @@ func.func @signExtendConstant() -> i16 {
 //       CHECK:   return %[[cres]]
 func.func @signExtendConstantSplat() -> vector<4xi16> {
   %c-2 = arith.constant -2 : i8
-  %splat = vector.splat %c-2 : vector<4xi8>
+  %splat = vector.broadcast %c-2 : i8 to vector<4xi8>
   %ext = arith.extsi %splat : vector<4xi8> to vector<4xi16>
   return %ext : vector<4xi16>
 }
@@ -682,7 +682,7 @@ func.func @unsignedExtendConstant() -> i16 {
 //       CHECK:   return %[[cres]]
 func.func @unsignedExtendConstantSplat() -> vector<4xi16> {
   %c2 = arith.constant 2 : i8
-  %splat = vector.splat %c2 : vector<4xi8>
+  %splat = vector.broadcast %c2 : i8 to vector<4xi8>
   %ext = arith.extui %splat : vector<4xi8> to vector<4xi16>
   return %ext : vector<4xi16>
 }
@@ -866,7 +866,7 @@ func.func @truncExtsiVector(%arg0: vector<2xi32>) -> vector<2xi16> {
 //       CHECK:   return %[[cres]]
 func.func @truncConstantSplat() -> vector<4xi8> {
   %c-2 = arith.constant -2 : i16
-  %splat = vector.splat %c-2 : vector<4xi16>
+  %splat = vector.broadcast %c-2 : i16 to vector<4xi16>
   %trunc = arith.trunci %splat : vector<4xi16> to vector<4xi8>
   return %trunc : vector<4xi8>
 }
@@ -2334,7 +2334,7 @@ func.func @constant_FPtoUI_splat() -> vector<4xi32> {
   // CHECK: %[[C0:.+]] = arith.constant dense<2> : vector<4xi32>
   // CHECK: return %[[C0]]
   %c0 = arith.constant 2.0 : f32
-  %splat = vector.splat %c0 : vector<4xf32>
+  %splat = vector.broadcast %c0 : f32 to vector<4xf32>
   %res = arith.fptoui %splat : vector<4xf32> to vector<4xi32>
   return %res : vector<4xi32>
 }
@@ -2374,7 +2374,7 @@ func.func @constant_FPtoSI_splat() -> vector<4xi32> {
   // CHECK: %[[C0:.+]] = arith.constant dense<-2> : vector<4xi32>
   // CHECK: return %[[C0]]
   %c0 = arith.constant -2.0 : f32
-  %splat = vector.splat %c0 : vector<4xf32>
+  %splat = vector.broadcast %c0 : f32 to vector<4xf32>
   %res = arith.fptosi %splat : vector<4xf32> to vector<4xi32>
   return %res : vector<4xi32>
 }
@@ -2413,7 +2413,7 @@ func.func @constant_SItoFP_splat() -> vector<4xf32> {
   // CHECK: %[[C0:.+]] = arith.constant dense<2.000000e+00> : vector<4xf32>
   // CHECK: return %[[C0]]
   %c0 = arith.constant 2 : i32
-  %splat = vector.splat %c0 : vector<4xi32>
+  %splat = vector.broadcast %c0 : i32 to vector<4xi32>
   %res = arith.sitofp %splat : vector<4xi32> to vector<4xf32>
   return %res : vector<4xf32>
 }
@@ -2442,7 +2442,7 @@ func.func @constant_UItoFP_splat() -> vector<4xf32> {
   // CHECK: %[[C0:.+]] = arith.constant dense<2.000000e+00> : vector<4xf32>
   // CHECK: return %[[C0]]
   %c0 = arith.constant 2 : i32
-  %splat = vector.splat %c0 : vector<4xi32>
+  %splat = vector.broadcast %c0 : i32 to vector<4xi32>
   %res = arith.uitofp %splat : vector<4xi32> to vector<4xf32>
   return %res : vector<4xf32>
 }
diff --git a/mlir/test/Dialect/Vector/vector-sink.mlir b/mlir/test/Dialect/Vector/vector-sink.mlir
index ef881ba05a416..577b06df42929 100644
--- a/mlir/test/Dialect/Vector/vector-sink.mlir
+++ b/mlir/test/Dialect/Vector/vector-sink.mlir
@@ -40,7 +40,7 @@ func.func @broadcast_scalar_with_bcast_scalable(%arg1: index, %arg2: index) -> v
 // CHECK:           %[[BCAST:.*]] = vector.broadcast %[[ADD]] : index to vector<1x4xindex>
 // CHECK:           return %[[BCAST]] : vector<1x4xindex>
 func.func @broadcast_scalar_with_bcast_and_splat(%arg1: index, %arg2: index) -> vector<1x4xindex> {
-  %0 = vector.splat %arg1 : vector<1x4xindex>
+  %0 = vector.broadcast %arg1 : index to vector<1x4xindex>
   %1 = vector.broadcast %arg2 : index to vector<1x4xindex>
   %2 = arith.addi %0, %1 : vector<1x4xindex>
   return %2 : vector<1x4xindex>
@@ -53,7 +53,7 @@ func.func @broadcast_scalar_with_bcast_and_splat(%arg1: index, %arg2: index) ->
 // CHECK:           %[[BCAST:.*]] = vector.broadcast %[[ADD]] : index to vector<1x[4]xindex>
 // CHECK:           return %[[BCAST]] : vector<1x[4]xindex>
 func.func @broadcast_scalar_with_bcast_and_splat_scalable(%arg1: index, %arg2: index) -> vector<1x[4]xindex> {
-  %0 = vector.splat %arg1 : vector<1x[4]xindex>
+  %0 = vector.broadcast %arg1 : index to vector<1x[4]xindex>
   %1 = vector.broadcast %arg2 : index to vector<1x[4]xindex>
   %2 = arith.addi %0, %1 : vector<1x[4]xindex>
   return %2 : vector<1x[4]xindex>
@@ -94,12 +94,12 @@ func.func @broadcast_vector_scalable(%arg1: vector<[4]xf32>, %arg2: vector<[4]xf
 // CHECK-LABEL:   func.func @broadcast_scalar_and_vec(
 // CHECK-SAME:       %[[ARG1:.*]]: index,
 // CHECK-SAME:       %[[ARG2:.*]]: vector<4xindex>) -> vector<1x4xindex> {
-// CHECK:            %[[SPLAT:.*]] = vector.splat %[[ARG1]] : vector<1x4xindex>
+// CHECK:            %[[SPLAT:.*]] = vector.broadcast %[[ARG1]] : index to vector<1x4xindex>
 // CHECK:            %[[BCAST:.*]] = vector.broadcast %[[ARG2]] : vector<4xindex> to vector<1x4xindex>
 // CHECK:            %[[ADD:.*]] = arith.addi %[[SPLAT]], %[[BCAST]] : vector<1x4xindex>
 // CHECK:            return %[[ADD]] : vector<1x4xindex>
 func.func @broadcast_scalar_and_vec(%arg1: index, %arg2: vector<4xindex>) -> vector<1x4xindex> {
-  %0 = vector.splat %arg1 : vector<1x4xindex>
+  %0 = vector.broadcast %arg1 : index to vector<1x4xindex>
   %1 = vector.broadcast %arg2 : vector<4xindex> to vector<1x4xindex>
   %2 = arith.addi %0, %1 : vector<1x4xindex>
   return %2 : vector<1x4xindex>
@@ -108,12 +108,12 @@ func.func @broadcast_scalar_and_vec(%arg1: index, %arg2: vector<4xindex>) -> vec
 // CHECK-LABEL:   func.func @broadcast_scalar_and_vec_scalable(
 // CHECK-SAME:       %[[ARG1:.*]]: index,
 // CHECK-SAME:       %[[ARG2:.*]]: vector<[4]xindex>) -> vector<1x[4]xindex> {
-// CHECK:            %[[SPLAT:.*]] = vector.splat %[[ARG1]] : vector<1x[4]xindex>
+// CHECK:            %[[SPLAT:.*]] = vector.broadcast %[[ARG1]] : index to vector<1x[4]xindex>
 // CHECK:            %[[BCAST:.*]] = vector.broadcast %[[ARG2]] : vector<[4]xindex> to vector<1x[4]xindex>
 // CHECK:            %[[ADD:.*]] = arith.addi %[[SPLAT]], %[[BCAST]] : vector<1x[4]xindex>
 // CHECK:            return %[[ADD]] : vector<1x[4]xindex>
 func.func @broadcast_scalar_and_vec_scalable(%arg1: index, %arg2: vector<[4]xindex>) -> vector<1x[4]xindex> {
-  %0 = vector.splat %arg1 : vector<1x[4]xindex>
+  %0 = vector.broadcast %arg1 : index to vector<1x[4]xindex>
   %1 = vector.broadcast %arg2 : vector<[4]xindex> to vector<1x[4]xindex>
   %2 = arith.addi %0, %1 : vector<1x[4]xindex>
   return %2 : vector<1x[4]xindex>
@@ -787,7 +787,7 @@ func.func @negative_extract_load_scalable(%arg0: memref<?xf32>, %arg1: index) ->
 //  CHECK-SAME:   (%[[ARG0:.*]]: memref<?xf32>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: f32)
 func.func @store_splat(%arg0: memref<?xf32>, %arg1: index, %arg2: f32) {
 // CHECK:   memref.store %[[ARG2]], %[[ARG0]][%[[ARG1]]] : memref<?xf32>
-  %0 = vector.splat %arg2 : vector<1xf32>
+  %0 = vector.broadcast %arg2 : f32 to vector<1xf32>
   vector.store %0, %arg0[%arg1] : memref<?xf32>, vector<1xf32>
   return
 }
@@ -813,9 +813,9 @@ func.func @store_broadcast_1d_to_2d(%arg0: memref<?x?xf32>, %arg1: index, %arg2:
 // CHECK-LABEL: @negative_store_scalable
 //  CHECK-SAME:   (%[[ARG0:.*]]: memref<?xf32>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: f32)
 func.func @negative_store_scalable(%arg0: memref<?xf32>, %arg1: index, %arg2: f32) {
-// CHECK:   %[[RES:.*]] = vector.splat %[[ARG2]] : vector<[1]xf32>
+// CHECK:   %[[RES:.*]] = vector.broadcast %[[ARG2]] : f32 to vector<[1]xf32>
 // CHECK:   vector.store %[[RES]], %[[ARG0]][%[[ARG1]]] : memref<?xf32>, vector<[1]xf32>
-  %0 = vector.splat %arg2 : vector<[1]xf32>
+  %0 = vector.broadcast %arg2 : f32 to vector<[1]xf32>
   vector.store %0, %arg0[%arg1] : memref<?xf32>, vector<[1]xf32>
   return
 }
@@ -823,9 +823,9 @@ func.func @negative_store_scalable(%arg0: memref<?xf32>, %arg1: index, %arg2: f3
 // CHECK-LABEL: @negative_store_memref_of_vec
 //  CHECK-SAME:   (%[[ARG0:.*]]: memref<?xvector<1xf32>>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: f32)
 func.func @negative_store_memref_of_vec(%arg0: memref<?xvector<1xf32>>, %arg1: index, %arg2: f32) {
-// CHECK:   %[[RES:.*]] = vector.splat %[[ARG2]] : vector<1xf32>
+// CHECK:   %[[RES:.*]] = vector.broadcast %[[ARG2]] : f32 to vector<1xf32>
 // CHECK:   vector.store %[[RES]], %[[ARG0]][%[[ARG1]]] : memref<?xvector<1xf32>>, vector<1xf32>
-  %0 = vector.splat %arg2 : vector<1xf32>
+  %0 = vector.broadcast %arg2 : f32 to vector<1xf32>
   vector.store %0, %arg0[%arg1] : memref<?xvector<1xf32>>, vector<1xf32>
   return
 }
@@ -833,9 +833,9 @@ func.func @negative_store_memref_of_vec(%arg0: memref<?xvector<1xf32>>, %arg1: i
 // CHECK-LABEL: @negative_store_more_than_one_element
 //  CHECK-SAME:   (%[[ARG0:.*]]: memref<?xf32>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: f32)
 func.func @negative_store_more_than_one_element(%arg0: memref<?xf32>, %arg1: index, %arg2: f32) {
-// CHECK:   %[[RES:.*]] = vector.splat %[[ARG2]] : vector<4xf32>
+// CHECK:   %[[RES:.*]] = vector.broadcast %[[ARG2]] : f32 to vector<4xf32>
 // CHECK:   vector.store %[[RES]], %[[ARG0]][%[[ARG1]]] : memref<?xf32>, vector<4xf32>
-  %0 = vector.splat %arg2 : vector<4xf32>
+  %0 = vector.broadcast %arg2 : f32 to vector<4xf32>
   vector.store %0, %arg0[%arg1] : memref<?xf32>, vector<4xf32>
   return
 }
@@ -843,10 +843,10 @@ func.func @negative_store_more_than_one_element(%arg0: memref<?xf32>, %arg1: ind
 // CHECK-LABEL: @negative_store_no_single_use
 //  CHECK-SAME:   (%[[ARG0:.*]]: memref<?xf32>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: f32)
 func.func @negative_store_no_single_use(%arg0: memref<?xf32>, %arg1: index, %arg2: f32) -> vector<1xf32> {
-// CHECK:   %[[RES:.*]] = vector.splat %[[ARG2]] : vector<1xf32>
+// CHECK:   %[[RES:.*]] = vector.broadcast %[[ARG2]] : f32 to vector<1xf32>
 // CHECK:   vector.store %[[RES]], %[[ARG0]][%[[ARG1]]] : memref<?xf32>, vector<1xf32>
 // CHECK:   return %[[RES:.*]] : vector<1xf32>
-  %0 = vector.splat %arg2 : vector<1xf32>
+  %0 = vector.broadcast %arg2 : f32 to vector<1xf32>
   vector.store %0, %arg0[%arg1] : memref<?xf32>, vector<1xf32>
   return %0 : vector<1xf32>
 }
diff --git a/mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir b/mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir
index 1b54d54ffbd9f..45afbffc1be48 100644
--- a/mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir
+++ b/mlir/test/Dialect/Vector/vector-transfer-to-vector-load-store.mlir
@@ -285,19 +285,19 @@ func.func @transfer_read_permutations(%mem_0 : memref<?x?xf32>, %mem_1 : memref<
   %c0 = arith.constant 0 : index
 
 // CHECK: %[[MASK0:.*]] = vector.broadcast %{{.*}} : i1 to vector<14x7xi1>
-  %mask0 = vector.splat %m : vector<14x7xi1>
+  %mask0 = vector.broadcast %m : i1 to vector<14x7xi1>
   %0 = vector.transfer_read %mem_1[%c0, %c0, %c0, %c0], %cst, %mask0 {in_bounds = [true, false, true, true], permutation_map = #map0} : memref<?x?x?x?xf32>, vector<7x14x8x16xf32>
 // CHECK: vector.transfer_read {{.*}} %[[MASK0]] {in_bounds = [false, true, true, true], permutation_map = #[[$MAP0]]} : memref<?x?x?x?xf32>, vector<14x7x8x16xf32>
 // CHECK: vector.transpose %{{.*}}, [1, 0, 2, 3] : vector<14x7x8x16xf32> to vector<7x14x8x16xf32>
 
 // CHECK: %[[MASK1:.*]] = vector.broadcast %{{.*}} : i1 to vector<16x14xi1>
-  %mask1 = vector.splat %m : vector<16x14xi1>
+  %mask1 = vector.broadcast %m : i1 to vector<16x14xi1>
   %1 = vector.transfer_read %mem_1[%c0, %c0, %c0, %c0], %cst, %mask1 {in_bounds = [true, false, true, false], permutation_map = #map1} : memref<?x?x?x?xf32>, vector<7x14x8x16xf32>
 // CHECK: vector.transfer_read {{.*}} %[[MASK1]] {in_bounds = [false, false, true, true], permutation_map = #[[$MAP0]]} : memref<?x?x?x?xf32>, vector<16x14x7x8xf32>
 // CHECK: vector.transpose %{{.*}}, [2, 1, 3, 0] : vector<16x14x7x8xf32> to vector<7x14x8x16xf32>
 
 // CHECK: %[[MASK3:.*]] = vector.broadcast %{{.*}} : i1 to vector<14x7xi1>
-  %mask2 = vector.splat %m : vector<14x7xi1>
+  %mask2 = vector.broadcast %m : i1 to vector<14x7xi1>
   %2 = vector.transfer_read %mem_1[%c0, %c0, %c0, %c0], %cst, %mask2 {in_bounds = [true, false, true, true], permutation_map = #map2} : memref<?x?x?x?xf32>, vector<7x14x8x16xf32>
 // CHECK: vector.transfer_read {{.*}} %[[MASK3]] {in_bounds = [false, true, true], permutation_map = #[[$MAP1]]} : memref<?x?x?x?xf32>, vector<14x16x7xf32>
 // CHECK: vector.broadcast %{{.*}} : vector<14x16x7xf32> to vector<8x14x16x7xf32>
@@ -337,7 +337,7 @@ func.func @transfer_write_permutations_tensor_masked(
   %c0 = arith.constant 0 : index
 
   // CHECK: %[[MASK:.*]] = vector.broadcast %[[M]] : i1 to vector<16x14x7x8xi1>
-  %mask0 = vector.splat %m : vector<16x14x7x8xi1>
+  %mask0 = vector.broadcast %m : i1 to vector<16x14x7x8xi1>
   %res = vector.transfer_write %vec, %dst[%c0, %c0, %c0, %c0], %mask0 {in_bounds = [true, false, false, true], permutation_map = affine_map<(d0, d1, d2, d3) -> (d2, d1, d3, d0)>} : vector<7x14x8x16xf32>, tensor<?x?x?x?xf32>
   // CHECK: %[[NEW_VEC0:.*]] = vector.transpose %{{.*}} [3, 1, 0, 2] : vector<7x14x8x16xf32> to vector<16x14x7x8xf32>
   // CHECK: %[[NEW_RES0:.*]] = vector.transfer_write %[[NEW_VEC0]], %[[DST]][%c0, %c0, %c0, %c0], %[[MASK]] {in_bounds = [true, false, true, false]} : vector<16x14x7x8xf32>, tensor<?x?x?x?xf32>
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f32.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f32.mlir
index 0ee016627440f..219367a41d51a 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f32.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f32.mlir
@@ -46,7 +46,7 @@ func.func @test_outerproduct_with_accumulator_4x4xf32() {
   %c0 = arith.constant 0 : index
   %f10 = arith.constant 10.0 : f32
 
-  %acc = vector.splat %f10 : vector<[4]x[4]xf32>
+  %acc = vector.broadcast %f10 : f32 to vector<[4]x[4]xf32>
   %vector_i32 = llvm.intr.stepvector : vector<[4]xi32>
   %vector = arith.sitofp %vector_i32 : vector<[4]xi32> to vector<[4]xf32>
   %tile = vector.outerproduct %vector, %vector, %acc : vector<[4]xf32>, vector<[4]xf32>
@@ -103,7 +103,7 @@ func.func @test_masked_outerproduct_with_accumulator_4x4xf32() {
   %ones = arith.constant dense<1> : vector<[4]xi32>
   %f10 = arith.constant 10.0 : f32
 
-  %acc = vector.splat %f10 : vector<[4]x[4]xf32>
+  %acc = vector.broadcast %f10 : f32 to vector<[4]x[4]xf32>
   %step_vector = llvm.intr.stepvector : vector<[4]xi32>
   %vector_i32 = arith.addi %step_vector, %ones : vector<[4]xi32>
   %vector = arith.sitofp %vector_i32 : vector<[4]xi32> to vector<[4]xf32>
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f64.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f64.mlir
index 8e812108c6055..059f24adbe721 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f64.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/outerproduct-f64.mlir
@@ -52,7 +52,7 @@ func.func @test_outerproduct_with_accumulator_2x2xf64() {
   %ones = arith.constant dense<1> : vector<[2]xi32>
   %f10 = arith.constant 10.0 : f64
 
-  %acc = vector.splat %f10 : vector<[2]x[2]xf64>
+  %acc = vector.broadcast %f10 : f64 to vector<[2]x[2]xf64>
   %step_vector = llvm.intr.stepvector : vector<[2]xi32>
   %vector_i32 = arith.addi %step_vector, %ones : vector<[2]xi32>
   %vector = arith.sitofp %vector_i32 : vector<[2]xi32> to vector<[2]xf64>
@@ -108,7 +108,7 @@ func.func @test_masked_outerproduct_with_accumulator_2x2xf64() {
   %ones = arith.constant dense<1> : vector<[2]xi32>
   %f10 = arith.constant 10.0 : f64
 
-  %acc = vector.splat %f10 : vector<[2]x[2]xf64>
+  %acc = vector.broadcast %f10 : f64 to vector<[2]x[2]xf64>
   %step_vector = llvm.intr.stepvector : vector<[2]xi32>
   %vector_i32 = arith.addi %step_vector, %ones : vector<[2]xi32>
   %vector = arith.sitofp %vector_i32 : vector<[2]xi32> to vector<[2]xf64>
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/transfer-write-2d.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/transfer-write-2d.mlir
index c3bf379cde617..bf6900ca810c2 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/transfer-write-2d.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSME/transfer-write-2d.mlir
@@ -10,7 +10,7 @@
 // Vector store.
 func.func @transfer_write_2d(%A : memref<?x?xf32>, %base1: index, %base2: index) {
   %c0 = arith.constant 0.0 : f32
-  %zero = vector.splat %c0 : vector<[4]x[4]xf32>
+  %zero = vector.broadcast %c0 : f32 to vector<[4]x[4]xf32>
   vector.transfer_write %zero, %A[%base1, %base2] {in_bounds=[true, true]} :
     vector<[4]x[4]xf32>, memref<?x?xf32>
   return
@@ -22,7 +22,7 @@ func.func @transfer_write_2d_mask(%A : memref<?x?xf32>, %base1: index, %base2: i
   %c2 = arith.constant 2 : index
   %c3 = arith.constant 3 : index
   %mask = vector.create_mask %c2, %c3 : vector<[4]x[4]xi1>
-  %zero = vector.splat %c0 : vector<[4]x[4]xf32>
+  %zero = vector.broadcast %c0 : f32 to vector<[4]x[4]xf32>
   vector.transfer_write %zero, %A[%base1, %base2], %mask {in_bounds=[true, true]} :
     vector<[4]x[4]xf32>, memref<?x?xf32>
   return
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/contraction.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/contraction.mlir
index c9904328763db..192f2911a3ee2 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/contraction.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/contraction.mlir
@@ -106,7 +106,7 @@ func.func @matvec_i32() {
   //    val = (123 * 314) * 4 * vscale
   // so ...
   %vscale = vector.vscale
-  %vscale_v = vector.splat %vscale : vector<3xindex>
+  %vscale_v = vector.broadcast %vscale : index to vector<3xindex>
   %vscale_i32 = arith.index_cast %vscale_v : vector<3xindex> to vector<3xi32>
   %mv1_div = arith.divui %mv1, %vscale_i32 : vector<3xi32>
   // ... val / vscale = 123 * 314 * 4 = 154488
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/scalable-interleave.mlir b/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/scalable-interleave.mlir
index d3b1fa4bbbc37..2d8180abdfec9 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/scalable-interleave.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/scalable-interleave.mlir
@@ -7,8 +7,8 @@
 func.func @entry() {
   %f1 = arith.constant 1.0 : f32
   %f2 = arith.constant 2.0 : f32
-  %v1 = vector.splat %f1 : vector<[4]xf32>
-  %v2 = vector.splat %f2 : vector<[4]xf32>
+  %v1 = vector.broadcast %f1 : f32 to vector<[4]xf32>
+  %v2 = vector.broadcast %f2 : f32 to vector<[4]xf32>
   vector.print %v1 : vector<[4]xf32>
   vector.print %v2 : vector<[4]xf32>
   //
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/interleave.mlir b/mlir/test/Integration/Dialect/Vector/CPU/interleave.mlir
index f812c25c0352f..740c742863919 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/interleave.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/interleave.mlir
@@ -6,8 +6,8 @@
 func.func @entry() {
   %f1 = arith.constant 1.0 : f32
   %f2 = arith.constant 2.0 : f32
-  %v1 = vector.splat %f1 : vector<2x4xf32>
-  %v2 = vector.splat %f2 : vector<2x4xf32>
+  %v1 = vector.broadcast %f1 : f32 to vector<2x4xf32>
+  %v2 = vector.broadcast %f2 : f32 to vector<2x4xf32>
   vector.print %v1 : vector<2x4xf32>
   vector.print %v2 : vector<2x4xf32>
   //
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-f32.mlir b/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-f32.mlir
index f7e2229321c00..e25795ab5f14d 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-f32.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-f32.mlir
@@ -14,9 +14,9 @@
 !vector_type_R = vector<7xf32>
 
 func.func @vector_outerproduct_splat_8x8(%fa: f32, %fb: f32, %fc: f32) -> !vector_type_C {
-  %a = vector.splat %fa: !vector_type_A
-  %b = vector.splat %fb: !vector_type_B
-  %c = vector.splat %fc: !vector_type_C
+  %a = vector.broadcast %fa: f32 to !vector_type_A
+  %b = vector.broadcast %fb: f32 to !vector_type_B
+  %c = vector.broadcast %fc: f32 to !vector_type_C
   %d = vector.outerproduct %a, %b, %c : !vector_type_A, !vector_type_B
   return %d: !vector_type_C
 }
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-i64.mlir b/mlir/test/Integration/Dialect/Vector/CPU/outerproduct-i64.mlir
index a19dfa1ce818e..0675102a...
[truncated]

@newling
Copy link
Contributor Author

newling commented Aug 6, 2025

@banach-space this PR touches ARM integration tests, so I'll wait for your stamp of approval

banach-space
banach-space approved these changes Aug 7, 2025
View reviewed changes
Copy link
Contributor

@banach-space banach-space left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks! LGTM

@newling newling merged commit 0cb98c7 into llvm:main Aug 7, 2025
17 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@banach-space banach-space banach-space approved these changes

@kuhar kuhar kuhar approved these changes

@dcaballe dcaballe Awaiting requested review from dcaballe dcaballe is a code owner

@nicolasvasilache nicolasvasilache Awaiting requested review from nicolasvasilache nicolasvasilache is a code owner

Assignees
No one assigned
Labels
mlir:arith mlir:sme mlir:sve mlir:vector mlir:vectorops mlir
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@newling @llvmbot @banach-space @kuhar