[mlir][vector] Update tests for xfer permutation lowering (3/N) #127320
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
* Remove `vector.create_mask` from tests. Instead, pass masks as arguments. This simplifies the tests without sacrificing test coverage. * Update `@xfer_read_minor_identity_tranposed_with_mask_scalable` to use similar shapes as other tests and to avoid using test Ops (e.g. `@test.some_use`). This improves consistency between tests. * Fix some comment typos.
|
@llvm/pr-subscribers-mlir-vector Author: Andrzej Warzyński (banach-space) Changes
Full diff: https://github.com/llvm/llvm-project/pull/127320.diff 1 Files Affected:
diff --git a/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir b/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
index 6705905633e0f..7029c9b15de8c 100644
--- a/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
+++ b/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
@@ -106,8 +106,8 @@ func.func @xfer_write_minor_identity_transposed_map_masked(
/// (neither a minor identity nor transposed minor identity map)
/// OUT 1: vector.broadcast + vector.transfer_write
/// (transposed minor identity)
-/// OUT 2: vector.transfer_write -> vector.broadcast + vector.transpose + vector.transfer_write
-/// (minor identity)
+/// OUT 2: vector.transfer_write -> vector.broadcast + vector.transpose
+/// + vector.transfer_write (minor identity)
///----------------------------------------------------------------------------------------
// CHECK-LABEL: func.func @xfer_write_non_minor_identity(
@@ -233,16 +233,16 @@ func.func @xfer_write_non_minor_identity_masked_scalable(
// CHECK-LABEL: func @xfer_write_non_minor_identity_masked_2
// CHECK-SAME: %[[DEST:.*]]: tensor<?x?x?x?xf32>
// CHECK-SAME: %[[VEC:.*]]: vector<14x8x16xf32>
-// CHECK-SAME: %[[DIM:.*]]: index, %[[IDX:.*]]: index) -> tensor<?x?x?x?xf32>
+// CHECK-SAME: %[[MASK:.*]]: vector<14x8x16xi1>
+// CHECK-SAME: %[[DIM:.*]]: index
// CHECK-NOT: vector.broadcast
-// CHECK: vector.mask %0 { vector.transfer_write %[[VEC]], %[[DEST]]{{.*}} : vector<14x8x16xf32>, tensor<?x?x?x?xf32> } : vector<14x8x16xi1> -> tensor<?x?x?x?xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_write %[[VEC]], %[[DEST]]{{.*}} : vector<14x8x16xf32>, tensor<?x?x?x?xf32> } : vector<14x8x16xi1> -> tensor<?x?x?x?xf32>
func.func @xfer_write_non_minor_identity_masked_2(
%dest : tensor<?x?x?x?xf32>,
%vec : vector<14x8x16xf32>,
- %dim : index,
+ %mask: vector<14x8x16xi1>,
%idx: index) -> tensor<?x?x?x?xf32> {
- %mask = vector.create_mask %dim, %dim, %dim : vector<14x8x16xi1>
%res = vector.mask %mask {
vector.transfer_write %vec, %dest[%idx, %idx, %idx, %idx] {
in_bounds = [false, false, true],
@@ -259,29 +259,27 @@ func.func @xfer_write_non_minor_identity_masked_2(
///
/// IN: vector.transfer_read
/// (_transposed_ minor identity permutation map, with 0 or more broadcast dims)
-/// OUT: vector.transpose + vector.transfer_write
+/// OUT: vector.transfer_read + vector.broadcast + vector.transpose
/// (minor identity permutation map with 0 or more leading broadcast dims)
///----------------------------------------------------------------------------------------
/// TODO: Inner broadcast dim - see also the block at the bottom of this file
-// CHECK-LABEL: func.func @xfer_read_minor_identity_tranposed_with_mask
+// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_with_mask
// CHECK-SAME: %[[MEM:.*]]: memref<?x?xf32>,
-// CHECK-SAME: %[[DIM_1:.*]]: index, %[[DIM_2:.*]]: index, %[[IDX:.*]]: index) -> vector<8x4x2xf32> {
+// CHECK-SAME: %[[MASK:.*]]: vector<2x4xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK: %[[PASS_THROUGH:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[MASK:.*]] = vector.create_mask %[[DIM_2]], %[[DIM_1]] : vector<2x4xi1>
// CHECK: %[[T_READ:.*]] = vector.transfer_read %[[MEM]]{{\[}}%[[IDX]], %[[IDX]]], %[[PASS_THROUGH]], %[[MASK]] {in_bounds = [true, true]} : memref<?x?xf32>, vector<2x4xf32>
// CHECK: %[[BCAST:.*]] = vector.broadcast %[[T_READ]] : vector<2x4xf32> to vector<8x2x4xf32>
// CHECK: %[[TRANSPOSE:.*]] = vector.transpose %[[BCAST]], [0, 2, 1] : vector<8x2x4xf32> to vector<8x4x2xf32>
// CHECK: return %[[TRANSPOSE]] : vector<8x4x2xf32>
-func.func @xfer_read_minor_identity_tranposed_with_mask(
+func.func @xfer_read_minor_identity_transposed_with_mask(
%mem: memref<?x?xf32>,
- %dim_1: index,
- %dim_2: index,
+ %mask: vector<2x4xi1>,
%idx: index) -> (vector<8x4x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim_2, %dim_1 : vector<2x4xi1>
%res = vector.transfer_read %mem[%idx, %idx], %pad, %mask {
in_bounds = [true, true, true],
permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
@@ -290,24 +288,22 @@ func.func @xfer_read_minor_identity_tranposed_with_mask(
return %res : vector<8x4x2xf32>
}
-// CHECK-LABEL: func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
+// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_with_mask_scalable(
// CHECK-SAME: %[[MEM:.*]]: memref<?x?xf32>,
-// CHECK-SAME: %[[DIM_1:.*]]: index, %[[DIM_2:.*]]: index, %[[IDX:.*]]: index) -> vector<8x[4]x2xf32> {
+// CHECK-SAME: %[[MASK:.*]]: vector<2x[4]xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK: %[[PAD:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[MASK:.*]] = vector.create_mask %[[DIM_2]], %[[DIM_1]] : vector<2x[4]xi1>
// CHECK: %[[T_READ:.*]] = vector.transfer_read %[[MEM]]{{\[}}%[[IDX]], %[[IDX]]], %[[PAD]], %[[MASK]] {in_bounds = [true, true]} : memref<?x?xf32>, vector<2x[4]xf32>
// CHECK: %[[BCAST:.*]] = vector.broadcast %[[T_READ]] : vector<2x[4]xf32> to vector<8x2x[4]xf32>
// CHECK: %[[TRANSPOSE:.*]] = vector.transpose %[[BCAST]], [0, 2, 1] : vector<8x2x[4]xf32> to vector<8x[4]x2xf32>
// CHECK: return %[[TRANSPOSE]] : vector<8x[4]x2xf32>
-func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
+func.func @xfer_read_minor_identity_transposed_with_mask_scalable(
%mem: memref<?x?xf32>,
- %dim_1: index,
- %dim_2: index,
+ %mask: vector<2x[4]xi1>,
%idx: index) -> (vector<8x[4]x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim_2, %dim_1 : vector<2x[4]xi1>
%res = vector.transfer_read %mem[%idx, %idx], %pad, %mask {
in_bounds = [true, true, true],
permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
@@ -319,24 +315,26 @@ func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
// Masked version is not supported
// CHECK-LABEL: func @xfer_read_minor_identity_transposed_masked(
-// CHECK-SAME: %[[DEST:.*]]: tensor<?x1xf32>,
-// CHECK-SAME: %[[MASK:.*]]: vector<4x1xi1>
+// CHECK-SAME: %[[DEST:.*]]: tensor<?x?xf32>,
+// CHECK-SAME: %[[MASK:.*]]: vector<2x4xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK-NOT: vector.transpose
-// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}}: tensor<?x1xf32>, vector<1x4x4xf32> } : vector<4x1xi1> -> vector<1x4x4xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}}: tensor<?x?xf32>, vector<8x4x2xf32> } : vector<2x4xi1> -> vector<8x4x2xf32>
func.func @xfer_read_minor_identity_transposed_masked(
- %dest: tensor<?x1xf32>,
- %mask : vector<4x1xi1>,
- %idx: index) {
+ %dest: tensor<?x?xf32>,
+ %mask: vector<2x4xi1>,
+ %idx: index) -> (vector<8x4x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %3 = vector.mask %mask {
- vector.transfer_read %dest[%idx, %idx], %pad {
- permutation_map = affine_map<(d0, d1) -> (d1, 0, d0)>
- } : tensor<?x1xf32>, vector<1x4x4xf32>
- } : vector<4x1xi1> -> vector<1x4x4xf32>
- "test.some_use"(%3) : (vector<1x4x4xf32>) -> ()
- return
+ %res = vector.mask %mask {
+ vector.transfer_read %dest[%idx, %idx], %pad, %mask {
+ in_bounds = [true, true, true],
+ permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
+ } : tensor<?x?xf32>, vector<8x4x2xf32>
+ } : vector<2x4xi1> -> vector<8x4x2xf32>
+
+ return %res : vector<8x4x2xf32>
}
// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_masked_scalable(
@@ -346,7 +344,7 @@ func.func @xfer_read_minor_identity_transposed_masked(
// CHECK: %[[T_READ:.*]] = vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}} : tensor<?x?xf32>, vector<8x[4]x2xf32> } : vector<2x[4]xi1> -> vector<8x[4]x2xf32>
func.func @xfer_read_minor_identity_transposed_masked_scalable(
%dest: tensor<?x?xf32>,
- %mask : vector<2x[4]xi1>,
+ %mask: vector<2x[4]xi1>,
%idx: index) -> vector<8x[4]x2xf32> {
%pad = arith.constant 0.000000e+00 : f32
@@ -388,17 +386,16 @@ func.func @xfer_read_minor_identitiy_bcast_dims_scalable(
// CHECK-LABEL: func.func @xfer_read_minor_identitiy_bcast_dims_masked
// CHECK-SAME: %[[MEM:.*]]: memref<?x?x?x?xf32>,
-// CHECK-SAME: %[[DIM:.*]]: index,
+// CHECK-SAME: %[[MASK:.*]]: vector<[4]x3xi1>
// CHECK-SAME: %[[IDX:.*]]: index) -> vector<8x[4]x2x3xf32> {
// CHECK-NOT: vector.broadcast
-// CHECK: %[[MASK:.*]] = vector.mask %0 { vector.transfer_read %[[MEM]]{{.*}} : memref<?x?x?x?xf32>, vector<8x[4]x2x3xf32> } : vector<[4]x3xi1> -> vector<8x[4]x2x3xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[MEM]]{{.*}} : memref<?x?x?x?xf32>, vector<8x[4]x2x3xf32> } : vector<[4]x3xi1> -> vector<8x[4]x2x3xf32>
func.func @xfer_read_minor_identitiy_bcast_dims_masked(
%mem: memref<?x?x?x?xf32>,
- %dim: index,
+ %mask: vector<[4]x3xi1>,
%idx: index) -> vector<8x[4]x2x3xf32> {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim, %dim: vector<[4]x3xi1>
%res = vector.mask %mask {
vector.transfer_read %mem[%idx, %idx, %idx, %idx], %pad {
|
|
@llvm/pr-subscribers-mlir Author: Andrzej Warzyński (banach-space) Changes
Full diff: https://github.com/llvm/llvm-project/pull/127320.diff 1 Files Affected:
diff --git a/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir b/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
index 6705905633e0f..7029c9b15de8c 100644
--- a/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
+++ b/mlir/test/Dialect/Vector/vector-transfer-permutation-lowering.mlir
@@ -106,8 +106,8 @@ func.func @xfer_write_minor_identity_transposed_map_masked(
/// (neither a minor identity nor transposed minor identity map)
/// OUT 1: vector.broadcast + vector.transfer_write
/// (transposed minor identity)
-/// OUT 2: vector.transfer_write -> vector.broadcast + vector.transpose + vector.transfer_write
-/// (minor identity)
+/// OUT 2: vector.transfer_write -> vector.broadcast + vector.transpose
+/// + vector.transfer_write (minor identity)
///----------------------------------------------------------------------------------------
// CHECK-LABEL: func.func @xfer_write_non_minor_identity(
@@ -233,16 +233,16 @@ func.func @xfer_write_non_minor_identity_masked_scalable(
// CHECK-LABEL: func @xfer_write_non_minor_identity_masked_2
// CHECK-SAME: %[[DEST:.*]]: tensor<?x?x?x?xf32>
// CHECK-SAME: %[[VEC:.*]]: vector<14x8x16xf32>
-// CHECK-SAME: %[[DIM:.*]]: index, %[[IDX:.*]]: index) -> tensor<?x?x?x?xf32>
+// CHECK-SAME: %[[MASK:.*]]: vector<14x8x16xi1>
+// CHECK-SAME: %[[DIM:.*]]: index
// CHECK-NOT: vector.broadcast
-// CHECK: vector.mask %0 { vector.transfer_write %[[VEC]], %[[DEST]]{{.*}} : vector<14x8x16xf32>, tensor<?x?x?x?xf32> } : vector<14x8x16xi1> -> tensor<?x?x?x?xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_write %[[VEC]], %[[DEST]]{{.*}} : vector<14x8x16xf32>, tensor<?x?x?x?xf32> } : vector<14x8x16xi1> -> tensor<?x?x?x?xf32>
func.func @xfer_write_non_minor_identity_masked_2(
%dest : tensor<?x?x?x?xf32>,
%vec : vector<14x8x16xf32>,
- %dim : index,
+ %mask: vector<14x8x16xi1>,
%idx: index) -> tensor<?x?x?x?xf32> {
- %mask = vector.create_mask %dim, %dim, %dim : vector<14x8x16xi1>
%res = vector.mask %mask {
vector.transfer_write %vec, %dest[%idx, %idx, %idx, %idx] {
in_bounds = [false, false, true],
@@ -259,29 +259,27 @@ func.func @xfer_write_non_minor_identity_masked_2(
///
/// IN: vector.transfer_read
/// (_transposed_ minor identity permutation map, with 0 or more broadcast dims)
-/// OUT: vector.transpose + vector.transfer_write
+/// OUT: vector.transfer_read + vector.broadcast + vector.transpose
/// (minor identity permutation map with 0 or more leading broadcast dims)
///----------------------------------------------------------------------------------------
/// TODO: Inner broadcast dim - see also the block at the bottom of this file
-// CHECK-LABEL: func.func @xfer_read_minor_identity_tranposed_with_mask
+// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_with_mask
// CHECK-SAME: %[[MEM:.*]]: memref<?x?xf32>,
-// CHECK-SAME: %[[DIM_1:.*]]: index, %[[DIM_2:.*]]: index, %[[IDX:.*]]: index) -> vector<8x4x2xf32> {
+// CHECK-SAME: %[[MASK:.*]]: vector<2x4xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK: %[[PASS_THROUGH:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[MASK:.*]] = vector.create_mask %[[DIM_2]], %[[DIM_1]] : vector<2x4xi1>
// CHECK: %[[T_READ:.*]] = vector.transfer_read %[[MEM]]{{\[}}%[[IDX]], %[[IDX]]], %[[PASS_THROUGH]], %[[MASK]] {in_bounds = [true, true]} : memref<?x?xf32>, vector<2x4xf32>
// CHECK: %[[BCAST:.*]] = vector.broadcast %[[T_READ]] : vector<2x4xf32> to vector<8x2x4xf32>
// CHECK: %[[TRANSPOSE:.*]] = vector.transpose %[[BCAST]], [0, 2, 1] : vector<8x2x4xf32> to vector<8x4x2xf32>
// CHECK: return %[[TRANSPOSE]] : vector<8x4x2xf32>
-func.func @xfer_read_minor_identity_tranposed_with_mask(
+func.func @xfer_read_minor_identity_transposed_with_mask(
%mem: memref<?x?xf32>,
- %dim_1: index,
- %dim_2: index,
+ %mask: vector<2x4xi1>,
%idx: index) -> (vector<8x4x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim_2, %dim_1 : vector<2x4xi1>
%res = vector.transfer_read %mem[%idx, %idx], %pad, %mask {
in_bounds = [true, true, true],
permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
@@ -290,24 +288,22 @@ func.func @xfer_read_minor_identity_tranposed_with_mask(
return %res : vector<8x4x2xf32>
}
-// CHECK-LABEL: func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
+// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_with_mask_scalable(
// CHECK-SAME: %[[MEM:.*]]: memref<?x?xf32>,
-// CHECK-SAME: %[[DIM_1:.*]]: index, %[[DIM_2:.*]]: index, %[[IDX:.*]]: index) -> vector<8x[4]x2xf32> {
+// CHECK-SAME: %[[MASK:.*]]: vector<2x[4]xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK: %[[PAD:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[MASK:.*]] = vector.create_mask %[[DIM_2]], %[[DIM_1]] : vector<2x[4]xi1>
// CHECK: %[[T_READ:.*]] = vector.transfer_read %[[MEM]]{{\[}}%[[IDX]], %[[IDX]]], %[[PAD]], %[[MASK]] {in_bounds = [true, true]} : memref<?x?xf32>, vector<2x[4]xf32>
// CHECK: %[[BCAST:.*]] = vector.broadcast %[[T_READ]] : vector<2x[4]xf32> to vector<8x2x[4]xf32>
// CHECK: %[[TRANSPOSE:.*]] = vector.transpose %[[BCAST]], [0, 2, 1] : vector<8x2x[4]xf32> to vector<8x[4]x2xf32>
// CHECK: return %[[TRANSPOSE]] : vector<8x[4]x2xf32>
-func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
+func.func @xfer_read_minor_identity_transposed_with_mask_scalable(
%mem: memref<?x?xf32>,
- %dim_1: index,
- %dim_2: index,
+ %mask: vector<2x[4]xi1>,
%idx: index) -> (vector<8x[4]x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim_2, %dim_1 : vector<2x[4]xi1>
%res = vector.transfer_read %mem[%idx, %idx], %pad, %mask {
in_bounds = [true, true, true],
permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
@@ -319,24 +315,26 @@ func.func @xfer_read_minor_identity_tranposed_with_mask_scalable(
// Masked version is not supported
// CHECK-LABEL: func @xfer_read_minor_identity_transposed_masked(
-// CHECK-SAME: %[[DEST:.*]]: tensor<?x1xf32>,
-// CHECK-SAME: %[[MASK:.*]]: vector<4x1xi1>
+// CHECK-SAME: %[[DEST:.*]]: tensor<?x?xf32>,
+// CHECK-SAME: %[[MASK:.*]]: vector<2x4xi1>
+// CHECK-SAME: %[[IDX:.*]]: index
// CHECK-NOT: vector.transpose
-// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}}: tensor<?x1xf32>, vector<1x4x4xf32> } : vector<4x1xi1> -> vector<1x4x4xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}}: tensor<?x?xf32>, vector<8x4x2xf32> } : vector<2x4xi1> -> vector<8x4x2xf32>
func.func @xfer_read_minor_identity_transposed_masked(
- %dest: tensor<?x1xf32>,
- %mask : vector<4x1xi1>,
- %idx: index) {
+ %dest: tensor<?x?xf32>,
+ %mask: vector<2x4xi1>,
+ %idx: index) -> (vector<8x4x2xf32>) {
%pad = arith.constant 0.000000e+00 : f32
- %3 = vector.mask %mask {
- vector.transfer_read %dest[%idx, %idx], %pad {
- permutation_map = affine_map<(d0, d1) -> (d1, 0, d0)>
- } : tensor<?x1xf32>, vector<1x4x4xf32>
- } : vector<4x1xi1> -> vector<1x4x4xf32>
- "test.some_use"(%3) : (vector<1x4x4xf32>) -> ()
- return
+ %res = vector.mask %mask {
+ vector.transfer_read %dest[%idx, %idx], %pad, %mask {
+ in_bounds = [true, true, true],
+ permutation_map = affine_map<(d0, d1) -> (0, d1, d0)>
+ } : tensor<?x?xf32>, vector<8x4x2xf32>
+ } : vector<2x4xi1> -> vector<8x4x2xf32>
+
+ return %res : vector<8x4x2xf32>
}
// CHECK-LABEL: func.func @xfer_read_minor_identity_transposed_masked_scalable(
@@ -346,7 +344,7 @@ func.func @xfer_read_minor_identity_transposed_masked(
// CHECK: %[[T_READ:.*]] = vector.mask %[[MASK]] { vector.transfer_read %[[DEST]]{{.*}} : tensor<?x?xf32>, vector<8x[4]x2xf32> } : vector<2x[4]xi1> -> vector<8x[4]x2xf32>
func.func @xfer_read_minor_identity_transposed_masked_scalable(
%dest: tensor<?x?xf32>,
- %mask : vector<2x[4]xi1>,
+ %mask: vector<2x[4]xi1>,
%idx: index) -> vector<8x[4]x2xf32> {
%pad = arith.constant 0.000000e+00 : f32
@@ -388,17 +386,16 @@ func.func @xfer_read_minor_identitiy_bcast_dims_scalable(
// CHECK-LABEL: func.func @xfer_read_minor_identitiy_bcast_dims_masked
// CHECK-SAME: %[[MEM:.*]]: memref<?x?x?x?xf32>,
-// CHECK-SAME: %[[DIM:.*]]: index,
+// CHECK-SAME: %[[MASK:.*]]: vector<[4]x3xi1>
// CHECK-SAME: %[[IDX:.*]]: index) -> vector<8x[4]x2x3xf32> {
// CHECK-NOT: vector.broadcast
-// CHECK: %[[MASK:.*]] = vector.mask %0 { vector.transfer_read %[[MEM]]{{.*}} : memref<?x?x?x?xf32>, vector<8x[4]x2x3xf32> } : vector<[4]x3xi1> -> vector<8x[4]x2x3xf32>
+// CHECK: vector.mask %[[MASK]] { vector.transfer_read %[[MEM]]{{.*}} : memref<?x?x?x?xf32>, vector<8x[4]x2x3xf32> } : vector<[4]x3xi1> -> vector<8x[4]x2x3xf32>
func.func @xfer_read_minor_identitiy_bcast_dims_masked(
%mem: memref<?x?x?x?xf32>,
- %dim: index,
+ %mask: vector<[4]x3xi1>,
%idx: index) -> vector<8x[4]x2x3xf32> {
%pad = arith.constant 0.000000e+00 : f32
- %mask = vector.create_mask %dim, %dim: vector<[4]x3xi1>
%res = vector.mask %mask {
vector.transfer_read %mem[%idx, %idx, %idx, %idx], %pad {
|
nujaa
approved these changes
Feb 17, 2025
| "test.some_use"(%3) : (vector<1x4x4xf32>) -> () | ||
| return | ||
| %res = vector.mask %mask { | ||
| vector.transfer_read %dest[%idx, %idx], %pad, %mask { |
There was a problem hiding this comment.
Hi, just checking if on purpose but it seems %mask has been added to the transfer_read here on top of using it in the vector.mask.
Also, in_bounds attributes have been added. does it change something for the test ?
There was a problem hiding this comment.
it seems %mask has been added to the transfer_read here on top of using it in the vector.mask.
Copy and paste error. I wanted to make this test a variation of @xfer_read_minor_identity_transposed_with_mask, but forgot to delete that %mask 😅
in_bounds was also added for consistency. Note that it doesn't change the output - the pattern won't apply as the "masked" version is not supported.
Welcome back :) And thank you so much for taking a look 🙏🏻 |
Remove mask that was added by mistake
…ng (3/N) Add a TODO re in_bounds
wldfngrs
pushed a commit
to wldfngrs/llvm-project
that referenced
this pull request
Feb 19, 2025
…#127320) * Remove `vector.create_mask` from tests. Instead, pass masks as arguments. This simplifies the tests without sacrificing test coverage. * Update `@xfer_read_minor_identity_tranposed_with_mask_scalable` to use similar shapes as other tests and to avoid using test Ops (e.g. `@test.some_use`). This improves consistency between tests. * Fix some comment typos.
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.
vector.create_maskfrom tests. Instead, pass masks asarguments. This simplifies the tests without sacrificing test
coverage.
@xfer_read_minor_identity_tranposed_with_mask_scalableto usesimilar shapes as other tests and to avoid using test Ops (e.g.
@test.some_use). This improves consistency between tests.