[mlir][linalg] Move vectorization tests for pad + insert_slice Ops (nfc) #140877
+1,046
−592
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
This patch moves vectorization tests for `tensor.pad` and
`tensor.insert_slice` into dedicated files under a new subdirectory for
the vectorizer. The goal is to better organize the growing set of tests,
which are currently difficult to navigate.
This change is also a preparatory step for upcoming work: I’ll soon be
updating the vectorization logic for `tensor.pad` +
`tensor.insert_slice`. With the new structure in place, two things will
be clear in follow-up changes:
* Only tests related to `tensor.pad` and `tensor.insert_slice` are
being updated.
* Only the relevant tests will be touched (e.g., when changing mask
generation, only tests involving masking will be affected).
|
@llvm/pr-subscribers-mlir-linalg @llvm/pr-subscribers-mlir Author: Andrzej Warzyński (banach-space) ChangesThis patch moves vectorization tests for This change is also a preparatory step for upcoming work: I’ll soon be
Patch is 98.83 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/140877.diff 6 Files Affected:
diff --git a/mlir/test/Dialect/Linalg/vectorization-with-patterns.mlir b/mlir/test/Dialect/Linalg/vectorization-with-patterns.mlir
index 9f2ee47b45b3e..b282c57e3e4cb 100644
--- a/mlir/test/Dialect/Linalg/vectorization-with-patterns.mlir
+++ b/mlir/test/Dialect/Linalg/vectorization-with-patterns.mlir
@@ -889,207 +889,6 @@ module attributes {transform.with_named_sequence} {
// -----
-// CHECK-LABEL: func @pad_static(
-// CHECK-SAME: %[[ARG0:.*]]: tensor<2x?x2xf32>, %[[PAD:.*]]: f32
-// CHECK-NOT: tensor.pad
-// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
-// CHECK-DAG: %[[C2:.*]] = arith.constant 2 : index
-// CHECK-DAG: %[[INIT:.*]] = tensor.empty() : tensor<2x3x4xf32>
-// CHECK-DAG: %[[VEC:.*]] = vector.broadcast %[[PAD]] : f32 to vector<2x3x4xf32>
-// CHECK: %[[FILL:.*]] = vector.transfer_write %[[VEC]], %[[INIT]]{{.*}} : vector<2x3x4xf32>, tensor<2x3x4xf32>
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG0]][%[[C0]], %[[C0]], %[[C0]]], %[[PAD]] {in_bounds = [true, false, true]} : tensor<2x?x2xf32>, vector<2x3x2xf32>
-// CHECK: %[[RESULT:.*]] = vector.transfer_write %[[READ]], %[[FILL]][%[[C0]], %[[C0]], %[[C2]]] {in_bounds = [true, true, true]} : vector<2x3x2xf32>, tensor<2x3x4xf32>
-// CHECK: return %[[RESULT]]
-func.func @pad_static(%arg0: tensor<2x?x2xf32>, %pad_value: f32) -> tensor<2x3x4xf32> {
- %0 = tensor.pad %arg0 low[0, 0, 2] high[0, 1, 0] {
- ^bb0(%arg1: index, %arg2: index, %arg3: index):
- tensor.yield %pad_value : f32
- } : tensor<2x?x2xf32> to tensor<2x3x4xf32>
- return %0 : tensor<2x3x4xf32>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-// CHECK-LABEL: func @pad_static_source(
-// CHECK-SAME: %[[ARG0:.*]]: tensor<2x5x2xf32>, %[[PAD:.*]]: f32
-// CHECK-NOT: tensor.pad
-// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
-// CHECK-DAG: %[[C2:.*]] = arith.constant 2 : index
-// CHECK: %[[INIT:.*]] = tensor.empty() : tensor<2x6x4xf32>
-// CHECK: %[[VEC:.*]] = vector.broadcast %[[PAD]] : f32 to vector<2x6x4xf32>
-// CHECK: %[[FILL:.*]] = vector.transfer_write %[[VEC]], %[[INIT]][%[[C0]], %[[C0]], %[[C0]]] {in_bounds = [true, true, true]} : vector<2x6x4xf32>, tensor<2x6x4xf32>
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG0]][%[[C0]], %[[C0]], %[[C0]]], %{{.*}} {in_bounds = [true, true, true]} : tensor<2x5x2xf32>, vector<2x5x2xf32>
-// CHECK: %[[WRITE:.*]] = vector.transfer_write %[[READ]], %[[FILL]][%[[C0]], %[[C0]], %[[C2]]] {in_bounds = [true, true, true]} : vector<2x5x2xf32>, tensor<2x6x4xf32>
-// CHECK: return %[[WRITE]]
-func.func @pad_static_source(%arg0: tensor<2x5x2xf32>, %pad_value: f32) -> tensor<2x6x4xf32> {
- %0 = tensor.pad %arg0 low[0, 0, 2] high[0, 1, 0] {
- ^bb0(%arg1: index, %arg2: index, %arg3: index):
- tensor.yield %pad_value : f32
- } : tensor<2x5x2xf32> to tensor<2x6x4xf32>
- return %0 : tensor<2x6x4xf32>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-
-// -----
-
-// CHECK-LABEL: func @pad_static_dynamic(
-// CHECK-SAME: %[[SRC:.*]]: tensor<1x2x2x?xf32>, %[[LOW:.*]]: index, %[[HIGH:.*]]: index
-// CHECK-NOT: tensor.pad
-// CHECK-DAG: %[[C2:.*]] = arith.constant 2 : index
-// CHECK-DAG: %[[C3:.*]] = arith.constant 3 : index
-// CHECK-DAG: %[[C5:.*]] = arith.constant 5 : index
-// CHECK: %[[V0:.*]] = arith.addi %[[LOW]], %[[C2]] : index
-// CHECK: %[[V1:.*]] = arith.addi %[[V0]], %[[C3]] : index
-// CHECK: %[[V2:.*]] = arith.addi %[[HIGH]], %[[C5]] : index
-// CHECK: %[[DIM3:.*]] = tensor.dim %[[SRC]], %[[C3]] : tensor<1x2x2x?xf32>
-// CHECK: %[[V4:.*]] = arith.addi %[[DIM3]], %[[C3]] : index
-// CHECK: %[[V5:.*]] = arith.addi %[[V4]], %[[C2]] : index
-// CHECK: %[[INIT:.*]] = tensor.empty(%[[V1]], %[[V2]], %[[V5]]) : tensor<6x?x?x?xf32>
-// CHECK: %[[FILL:.*]] = linalg.fill ins(%{{.*}} : f32) outs(%[[INIT]] : tensor<6x?x?x?xf32>) -> tensor<6x?x?x?xf32>
-// CHECK: %[[SRCDIM:.*]] = tensor.dim %[[SRC]], %[[C3]] : tensor<1x2x2x?xf32>
-// CHECK: %[[RESULT:.*]] = tensor.insert_slice %[[SRC]] into %[[FILL]][2, %[[LOW]], 3, 3] [1, 2, 2, %[[SRCDIM]]] [1, 1, 1, 1] : tensor<1x2x2x?xf32> into tensor<6x?x?x?xf32>
-// CHECK: return %[[RESULT]]
-func.func @pad_static_dynamic(%arg0: tensor<1x2x2x?xf32>, %low: index, %high: index,
- %pad_value: f32) -> tensor<6x?x?x?xf32> {
- %0 = tensor.pad %arg0 low[2, %low, 3, 3] high[3, 3, %high, 2] {
- ^bb0(%arg1: index, %arg2: index, %arg3: index, %arg4: index):
- tensor.yield %pad_value : f32
- } : tensor<1x2x2x?xf32> to tensor<6x?x?x?xf32>
- return %0 : tensor<6x?x?x?xf32>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-// CHECK-LABEL: func @pad_static_complex(
-// CHECK-NOT: vector<
-func.func @pad_static_complex(%arg0: tensor<2x5x2xcomplex<f32>>, %pad_value: complex<f32>) -> tensor<2x6x4xcomplex<f32>> {
- %0 = tensor.pad %arg0 low[0, 0, 2] high[0, 1, 0] {
- ^bb0(%arg1: index, %arg2: index, %arg3: index):
- tensor.yield %pad_value : complex<f32>
- } : tensor<2x5x2xcomplex<f32>> to tensor<2x6x4xcomplex<f32>>
- return %0 : tensor<2x6x4xcomplex<f32>>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-func.func private @make_vector() -> tensor<12x13xf32>
-
-// CHECK-LABEL: func.func @pad_and_insert_slice_dest(
-// CHECK-SAME: %[[ARG_0:.*]]: tensor<1x5x6xf32>) -> tensor<1x12x13xf32> {
-// CHECK: %[[C0:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[CST:.*]] = arith.constant dense<5.000000e+00> : vector<1x12x13xf32>
-// CHECK: %[[C0_IDX:.*]] = arith.constant 0 : index
-// CHECK: %[[PAD_VAL:.*]] = arith.constant 5.000000e+00 : f32
-// CHECK: %[[EMPTY:.*]] = tensor.empty() : tensor<1x12x13xf32>
-// CHECK: %[[WRITE_1:.*]] = vector.transfer_write %[[CST]], %[[EMPTY]]{{\[}}%[[C0_IDX]], %[[C0_IDX]], %[[C0_IDX]]] {in_bounds = [true, true, true]} : vector<1x12x13xf32>, tensor<1x12x13xf32>
-// CHECK: %[[READ_1:.*]] = vector.transfer_read %[[ARG_0]]{{\[}}%[[C0_IDX]], %[[C0_IDX]], %[[C0_IDX]]], %[[PAD_VAL]] {in_bounds = [true, true, true]} : tensor<1x5x6xf32>, vector<1x5x6xf32>
-// CHECK: %[[WRITE_2:.*]] = vector.transfer_write %[[READ_1]], %[[WRITE_1]]{{\[}}%[[C0_IDX]], %[[C0_IDX]], %[[C0_IDX]]] {in_bounds = [true, true, true]} : vector<1x5x6xf32>, tensor<1x12x13xf32>
-// CHECK: %[[MAKE_VEC:.*]] = call @make_vector() : () -> tensor<12x13xf32>
-// CHECK: %[[READ_2:.*]] = vector.transfer_read %[[MAKE_VEC]]{{\[}}%[[C0_IDX]], %[[C0_IDX]]], %[[C0]] {in_bounds = [true, true]} : tensor<12x13xf32>, vector<12x13xf32>
-// CHECK: %[[RES:.*]] = vector.transfer_write %[[READ_2]], %[[WRITE_2]]{{\[}}%[[C0_IDX]], %[[C0_IDX]], %[[C0_IDX]]] {in_bounds = [true, true]} : vector<12x13xf32>, tensor<1x12x13xf32>
-// CHECK: return %[[RES]] : tensor<1x12x13xf32>
-func.func @pad_and_insert_slice_dest(
- %arg0: tensor<1x5x6xf32>) -> tensor<1x12x13xf32> {
- %c5 = arith.constant 5.0 : f32
- %0 = tensor.pad %arg0 low[0, 0, 0] high[0, 7, 7] {
- ^bb0(%arg2: index, %arg3: index, %arg4: index):
- tensor.yield %c5 : f32
- } : tensor<1x5x6xf32> to tensor<1x12x13xf32>
- %1 = call @make_vector() : () -> tensor<12x13xf32>
- %r = tensor.insert_slice %1 into %0[0, 0, 0][1, 12, 13][1, 1, 1] : tensor<12x13xf32> into tensor<1x12x13xf32>
- return %r : tensor<1x12x13xf32>
-}
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %3 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %4 = transform.get_parent_op %3 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %5 = transform.structured.vectorize_children_and_apply_patterns %4 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-// CHECK-LABEL: func @pad_tensor_non_const_pad_value
-// CHECK-SAME: %[[ARG0:.*]]: tensor<5x6xf32>
-// CHECK-NOT: tensor.pad
-// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
-// CHECK-DAG: %[[C3:.*]] = arith.constant 3 : index
-// CHECK-DAG: %[[C4:.*]] = arith.constant 4 : index
-// CHECK: %[[FILL:.*]] = tensor.generate
-// CHECK: %[[RES:.*]] = arith.mulf
-// CHECK: tensor.yield %[[RES]] : f32
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG0]][%[[C0]], %[[C0]]], %{{.*}} {in_bounds = [true, true]} : tensor<5x6xf32>, vector<5x6xf32>
-// CHECK: %[[WRITE:.*]] = vector.transfer_write %[[READ]], %[[FILL]][%[[C3]], %[[C4]]] {in_bounds = [true, true]} : vector<5x6xf32>, tensor<12x13xf32>
-// CHECK: return %[[WRITE]]
-func.func @pad_tensor_non_const_pad_value(%arg0: tensor<5x6xf32>) -> tensor<12x13xf32> {
- %c0 = arith.constant 0 : index
- %c5 = arith.constant 5.0 : f32
- %0 = tensor.pad %arg0 low[3, 4] high[4, 3] {
- ^bb0(%arg1: index, %arg2: index):
- %i1 = arith.index_cast %arg1 : index to i32
- %i2 = arith.index_cast %arg2 : index to i32
- %f1 = arith.sitofp %i1 : i32 to f32
- %f2 = arith.sitofp %i2 : i32 to f32
- %m = arith.mulf %f1, %f2 : f32
- tensor.yield %m : f32
- } : tensor<5x6xf32> to tensor<12x13xf32>
- return %0 : tensor<12x13xf32>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %3 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %4 = transform.get_parent_op %3 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %5 = transform.structured.vectorize_children_and_apply_patterns %4 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
// CHECK-LABEL: func @sum_exp
func.func @sum_exp(%input: tensor<4x16x8xf32>, %output: tensor<4x16xf32>)
-> tensor<4x16xf32>
@@ -1805,29 +1604,6 @@ module attributes {transform.with_named_sequence} {
// -----
-// CHECK-LABEL: func @test_masked_pad_static_dynamic
-func.func @test_masked_pad_static_dynamic(%arg0: tensor<1x2x2x?xf32>, %low: index, %high: index,
- %pad_value: f32) -> tensor<6x?x?x?xf32> {
- // CHECK: tensor.pad
- %0 = tensor.pad %arg0 low[2, %low, 3, 3] high[3, 3, %high, 2] {
- ^bb0(%arg1: index, %arg2: index, %arg3: index, %arg4: index):
- tensor.yield %pad_value : f32
- } : tensor<1x2x2x?xf32> to tensor<6x?x?x?xf32>
- return %0 : tensor<6x?x?x?xf32>
-}
-
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.pad"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
func.func @zero_dim_tensor(%input: tensor<f32>, %output: tensor<f32>) -> tensor<f32>
{
%0 = linalg.generic { indexing_maps = [ affine_map<() -> ()>, affine_map<() -> ()> ],
@@ -2001,94 +1777,3 @@ module attributes {transform.with_named_sequence} {
transform.yield
}
}
-
-// -----
-
-///----------------------------------------------------------------------------------------
-/// tensor.insert_slice
-///----------------------------------------------------------------------------------------
-
-// The pad value for xfer-read is neither needed nor available - use the default (0.0).
-
-// CHECK-LABEL: func @insert_static_slice_default_pad
-// CHECK-SAME: %[[ARG_0:.*]]: tensor<1x2x3xf32>,
-// CHECK-SAME: %[[ARG_1:.*]]: tensor<9x8x7x1x2x3xf32>) -> tensor<9x8x7x1x2x3xf32> {
-// CHECK: %[[PAD:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK: %[[C0:.*]] = arith.constant 0 : index
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG_0]]{{\[}}%[[C0]], %[[C0]], %[[C0]]], %[[PAD]] {in_bounds = [true, true, true]} : tensor<1x2x3xf32>, vector<1x2x3xf32>
-// CHECK: %[[WRITE:.*]] = vector.transfer_write %[[READ]], %[[ARG_1]]{{\[}}%[[C0]], %[[C0]], %[[C0]], %[[C0]], %[[C0]], %[[C0]]] {in_bounds = [true, true, true]} : vector<1x2x3xf32>, tensor<9x8x7x1x2x3xf32>
-// CHECK: return %[[WRITE]] : tensor<9x8x7x1x2x3xf32>
-func.func @insert_static_slice_default_pad(%arg0: tensor<1x2x3xf32>, %arg1: tensor<9x8x7x1x2x3xf32>) -> tensor<9x8x7x1x2x3xf32> {
- %res = tensor.insert_slice %arg0 into %arg1[0, 0, 0, 0, 0, 0] [1, 1, 1, 1, 2, 3][1, 1, 1, 1, 1, 1] : tensor<1x2x3xf32> into tensor<9x8x7x1x2x3xf32>
- return %res : tensor<9x8x7x1x2x3xf32>
-}
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.insert_slice"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 { vectorize_padding } : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-// Same as above, but there's a pad value available that should be used instead of the default value.
-
-// CHECK-LABEL: func.func @insert_static_slice_non_zero_pad
-// CHECK-SAME: %[[ARG_0:.*]]: tensor<1x2x3xf32>,
-// CHECK-SAME: %[[PAD:.*]]: f32) -> tensor<9x8x7x1x2x3xf32> {
-// CHECK: %[[EMPTY:.*]] = tensor.empty() : tensor<9x8x7x1x2x3xf32>
-// CHECK: %[[BC:.*]] = vector.broadcast %[[PAD]] : f32 to vector<9x8x7x1x2x3xf32>
-// CHECK: %[[WRITE:.*]] = vector.transfer_write %[[BC]], %[[EMPTY]]{{.*}} {in_bounds = [true, true, true, true, true, true]} : vector<9x8x7x1x2x3xf32>, tensor<9x8x7x1x2x3xf32>
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG_0]]{{.*}}, %[[PAD]] {in_bounds = [true, true, true]} : tensor<1x2x3xf32>, vector<1x2x3xf32>
-// CHECK: %[[RES:.*]] = vector.transfer_write %[[READ]], %[[WRITE]]{{.*}} {in_bounds = [true, true, true]} : vector<1x2x3xf32>, tensor<9x8x7x1x2x3xf32>
-// CHECK: return %[[RES]] : tensor<9x8x7x1x2x3xf32>
-func.func @insert_static_slice_non_zero_pad(%arg0: tensor<1x2x3xf32>, %pad : f32) -> tensor<9x8x7x1x2x3xf32> {
- %init = tensor.empty() : tensor<9x8x7x1x2x3xf32>
- %fill = linalg.fill ins(%pad : f32) outs(%init : tensor<9x8x7x1x2x3xf32>) -> tensor<9x8x7x1x2x3xf32>
- %res = tensor.insert_slice %arg0 into %fill[0, 0, 0, 0, 0, 0] [1, 1, 1, 1, 2, 3][1, 1, 1, 1, 1, 1] : tensor<1x2x3xf32> into tensor<9x8x7x1x2x3xf32>
- return %res : tensor<9x8x7x1x2x3xf32>
-}
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.insert_slice"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
-
-// -----
-
-// Same as above, but the source type has is dynamically shaped. This means
-// that the pad value is now required and the vector dim corresponding to the
-// dynamic shape has to be inferred from the shape of the destination tensor.
-
-// CHECK-LABEL: func.func @insert_dynamic_slice_non_zero_pad(
-// CHECK-SAME: %[[ARG_0:.*]]: tensor<1x?x3xf32>,
-// CHECK-SAME: %[[PAD:.*]]: f32,
-// CHECK-SAME: %[[SIZE:.*]]: index) -> tensor<9x8x7x1x2x3xf32> {
-// CHECK: %[[EMPTY:.*]] = tensor.empty() : tensor<9x8x7x1x2x3xf32>
-// CHECK: %[[BC:.*]] = vector.broadcast %[[PAD]] : f32 to vector<9x8x7x1x2x3xf32>
-// CHECK: %[[WRITE:.*]] = vector.transfer_write %[[BC]], %[[EMPTY]]{{.*}} {in_bounds = [true, true, true, true, true, true]} : vector<9x8x7x1x2x3xf32>, tensor<9x8x7x1x2x3xf32>
-// CHECK: %[[READ:.*]] = vector.transfer_read %[[ARG_0]]{{.*}}, %[[PAD]] {in_bounds = [true, false, true]} : tensor<1x?x3xf32>, vector<1x2x3xf32>
-// CHECK: %[[RES:.*]] = vector.transfer_write %[[READ]], %[[WRITE]]{{.*}} {in_bounds = [true, true, true]} : vector<1x2x3xf32>, tensor<9x8x7x1x2x3xf32>
-// CHECK: return %[[RES]] : tensor<9x8x7x1x2x3xf32>
-func.func @insert_dynamic_slice_non_zero_pad(%arg0: tensor<1x?x3xf32>, %pad : f32, %size: index) -> tensor<9x8x7x1x2x3xf32> {
- %init = tensor.empty() : tensor<9x8x7x1x2x3xf32>
- %fill = linalg.fill ins(%pad : f32) outs(%init : tensor<9x8x7x1x2x3xf32>) -> tensor<9x8x7x1x2x3xf32>
- %res = tensor.insert_slice %arg0 into %fill[0, 0, 0, 0, 0, 0] [1, 1, 1, 1, %size, 3][1, 1, 1, 1, 1, 1] : tensor<1x?x3xf32> into tensor<9x8x7x1x2x3xf32>
- return %res : tensor<9x8x7x1x2x3xf32>
-}
-
-module attributes {transform.with_named_sequence} {
- transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
- %0 = transform.structured.match ops{["tensor.insert_slice"]} in %arg1 : (!transform.any_op) -> !transform.any_op
- %1 = transform.get_parent_op %0 {isolated_from_above} : (!transform.any_op) -> !transform.any_op
- %2 = transform.structured.vectorize_children_and_apply_patterns %1 : (!transform.any_op) -> !transform.any_op
- transform.yield
- }
-}
diff --git a/mlir/test/Dialect/Linalg/vectorization.mlir b/mlir/test/Dialect/Linalg/vectorization.mlir
index 6b760a15afd56..8c6760fa50325 100644
--- a/mlir/test/Dialect/Linalg/vectorization.mlir
+++ b/mlir/test/Dialect/Linalg/vectorization.mlir
@@ -580,133 +580,6 @@ module attributes {transform.with_named_sequence} {
}
}
-// -----
-
-// CHECK-LABEL: func @test_masked_vectorize_pad
-func.func @test_masked_vectorize_pad(
- %0 : tensor<?x?xf32>, %h0 : index, %h1 : index)
- -> tensor<2x4xf32>
-{
- // CHECK-DAG: %[[c42:.*]] = arith.constant 4.243000e+01 : f32
- // CHECK-DAG: %[[c0:.*]] = arith.constant 0 : index
- // CHECK-DAG: %[[c0_0:.*]] = arith.constant 0 : index
- // CHECK: %[[d0:.*]] = tensor.dim {{.*}} : ...
[truncated]
|
javedabsar1
approved these changes
May 21, 2025
hanhanW
reviewed
May 21, 2025
There was a problem hiding this comment.
What is the rule of this? Why other vectorization tests are not moved? Will you move them in the future?
I dont have issues of finding tests. I typically just clicked into a few files and search the ops. I'd be confused about where to find tests if you land the change. Because some of them are in Linalg/ and some of them are in Linalg/vectorization. The other point is that I don't know which file I should use if I add a new support for vectorization.
|
(I think your rule may be moving all vectorization tests of non-linalg ops to |
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.
This patch moves vectorization tests for
tensor.padandtensor.insert_sliceinto dedicated files under a new subdirectory forthe vectorizer. The goal is to better organize the growing set of tests,
which are currently difficult to navigate.
This change is also a preparatory step for upcoming work: I’ll soon be
updating the vectorization logic for
tensor.pad+tensor.insert_slice. With the new structure in place, two things willbe clear in follow-up changes:
tensor.padandtensor.insert_slicearebeing updated.
generation, only tests involving masking will be affected).