Skip to content
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

Update tests to match changed op order #416

Closed
wants to merge 1 commit into from
Closed

Update tests to match changed op order #416

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 9 additions & 9 deletions test/Conversion/TorchToTosa/basic.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1103,15 +1103,15 @@ func.func @torch.aten.avg_pool2d$basic(%arg0: !torch.vtensor<[1,512,7,7],f32> )

// CHECK-LABEL: @torch.aten.max.dim$basic(
// CHECK-SAME: %[[ARG0:.*]]: tensor<3x2x3xf32>)
// CHECK: %[[VAL_0:.*]] = torch_c.from_builtin_tensor %[[ARG0]] : tensor<3x2x3xf32> -> !torch.vtensor<[3,2,3],f32>
// CHECK: %[[VAL_1:.*]] = torch_c.to_builtin_tensor %[[VAL_0]] : !torch.vtensor<[3,2,3],f32> -> tensor<3x2x3xf32>
// CHECK: %[[VAL_TRUE:.*]] = torch.constant.bool true
// CHECK: %[[VAL_I2:.*]] = torch.constant.int 2
// CHECK: %[[VAL_2:.*]] = tosa.reduce_max %[[VAL_1]] {axis = 2 : i32} : (tensor<3x2x3xf32>) -> tensor<3x2x1xf32>
// CHECK: %[[VAL_3:.*]] = tosa.argmax %[[VAL_1]] {axis = 2 : i32} : (tensor<3x2x3xf32>) -> tensor<3x2xi64>
// CHECK: %[[VAL_4:.*]] = tosa.reshape %[[VAL_3]] {new_shape = array<i64: 3, 2, 1>} : (tensor<3x2xi64>) -> tensor<3x2x1xi64>
// CHECK: %[[VAL_5:.*]] = torch_c.from_builtin_tensor %[[VAL_2]] : tensor<3x2x1xf32> -> !torch.vtensor<[3,2,1],f32>
// CHECK: %[[VAL_6:.*]] = torch_c.to_builtin_tensor %[[VAL_5]] : !torch.vtensor<[3,2,1],f32> -> tensor<3x2x1xf32>
// CHECK-DAG: %[[VAL_0:.*]] = torch_c.from_builtin_tensor %[[ARG0]] : tensor<3x2x3xf32> -> !torch.vtensor<[3,2,3],f32>
// CHECK-DAG: %[[VAL_1:.*]] = torch_c.to_builtin_tensor %[[VAL_0]] : !torch.vtensor<[3,2,3],f32> -> tensor<3x2x3xf32>
// CHECK-DAG: %[[VAL_TRUE:.*]] = torch.constant.bool true
// CHECK-DAG: %[[VAL_I2:.*]] = torch.constant.int 2
// CHECK-DAG: %[[VAL_2:.*]] = tosa.reduce_max %[[VAL_1]] {axis = 2 : i32} : (tensor<3x2x3xf32>) -> tensor<3x2x1xf32>
// CHECK-DAG: %[[VAL_3:.*]] = tosa.argmax %[[VAL_1]] {axis = 2 : i32} : (tensor<3x2x3xf32>) -> tensor<3x2xi64>
// CHECK-DAG: %[[VAL_4:.*]] = tosa.reshape %[[VAL_3]] {new_shape = array<i64: 3, 2, 1>} : (tensor<3x2xi64>) -> tensor<3x2x1xi64>
// CHECK-DAG: %[[VAL_5:.*]] = torch_c.from_builtin_tensor %[[VAL_2]] : tensor<3x2x1xf32> -> !torch.vtensor<[3,2,1],f32>
// CHECK-DAG: %[[VAL_6:.*]] = torch_c.to_builtin_tensor %[[VAL_5]] : !torch.vtensor<[3,2,1],f32> -> tensor<3x2x1xf32>
// CHECK: return %[[VAL_6]] : tensor<3x2x1xf32>
func.func @torch.aten.max.dim$basic(%arg0: tensor<3x2x3xf32>) -> tensor<3x2x1xf32> {
%0 = torch_c.from_builtin_tensor %arg0 : tensor<3x2x3xf32> -> !torch.vtensor<[3,2,3],f32>
Expand Down
30 changes: 15 additions & 15 deletions test/Dialect/TMTensor/bufferize.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,17 +4,17 @@
// CHECK-LABEL: func.func @scan_1d_inclusive(
// CHECK-SAME: %[[IN_TENSOR:.*]]: tensor<128xi32>, %[[OUT_TENSOR:.*]]: tensor<128xi32>,
// CHECK-SAME: %[[ACC_TENSOR:.*]]: tensor<i32>) -> (tensor<128xi32>, tensor<i32>) {
// CHECK: %[[IN_MEMREF:.*]] = bufferization.to_memref %[[IN_TENSOR]] : memref<128xi32>
// CHECK: %[[OUT_MEMREF_NEW:.*]] = memref.alloc() : memref<128xi32>
// CHECK: %[[ACC_MEMREF_NEW:.*]] = memref.alloc() : memref<i32>
// CHECK-DAG: %[[IN_MEMREF:.*]] = bufferization.to_memref %[[IN_TENSOR]] : memref<128xi32>
// CHECK-DAG: %[[OUT_MEMREF_NEW:.*]] = memref.alloc() : memref<128xi32>
// CHECK-DAG: %[[ACC_MEMREF_NEW:.*]] = memref.alloc() : memref<i32>
// CHECK-DAG: %[[OUT_TENSOR_NEW:.*]] = bufferization.to_tensor %[[OUT_MEMREF_NEW]] : memref<128xi32>
// CHECK-DAG: %[[ACC_TENSOR_NEW:.*]] = bufferization.to_tensor %[[ACC_MEMREF_NEW]] : memref<i32>
// CHECK: tm_tensor.scan dimension(0) inclusive(true) ins(%[[IN_MEMREF]] : memref<128xi32>)
// CHECK-SAME: outs(%[[OUT_MEMREF_NEW]], %[[ACC_MEMREF_NEW]] : memref<128xi32>, memref<i32>) {
// CHECK: ^bb0(%[[OUT_PREV_ELEMENT:.*]]: i32, %[[IN_ELEMENT:.*]]: i32):
// CHECK: %[[OUT_CURRENT_ELEMENT:.*]] = arith.addi %[[OUT_PREV_ELEMENT]], %[[IN_ELEMENT]] : i32
// CHECK: tm_tensor.yield %[[OUT_CURRENT_ELEMENT]] : i32
// CHECK: }
// CHECK-DAG: %[[OUT_TENSOR_NEW:.*]] = bufferization.to_tensor %[[OUT_MEMREF_NEW]] : memref<128xi32>
Copy link
Author

Choose a reason for hiding this comment

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

Can someone confirm that moving this above the last use of the memref is valid?

Copy link
Collaborator

Choose a reason for hiding this comment

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

Looks odd. Is this how the reference look upstream? They have a bump to that LLVM version, right?
Maybe we want to wait until I catched up with torch-mlir bumps?

Copy link
Author

Choose a reason for hiding this comment

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

Yes, upstream they have the same reference for some of the tests: https://github.com/llvm/torch-mlir/blob/8e0eafd022cd7555c8b58927d3238a7a89e9dbd4/test/Dialect/TMTensor/bufferize.mlir#L10 . I agree that we should try to get to llvm#3765 in torch-mlir .

// CHECK-DAG: %[[ACC_TENSOR_NEW:.*]] = bufferization.to_tensor %[[ACC_MEMREF_NEW]] : memref<i32>
// CHECK: return %[[OUT_TENSOR_NEW]], %[[ACC_TENSOR_NEW]] : tensor<128xi32>, tensor<i32>
func.func @scan_1d_inclusive(%in: tensor<128xi32>, %out: tensor<128xi32>, %acc: tensor<i32>) -> (tensor<128xi32>, tensor<i32>) {
%ret_out, %ret_acc = tm_tensor.scan dimension(0) inclusive(true)
Expand All @@ -32,17 +32,17 @@ func.func @scan_1d_inclusive(%in: tensor<128xi32>, %out: tensor<128xi32>, %acc:
// CHECK-SAME: %[[ACC_TENSOR:.*]]: tensor<i32>) -> (tensor<128xi32>, tensor<i32>) {
// CHECK-DAG: %[[IN_MEMREF:.*]] = bufferization.to_memref %[[IN_TENSOR]] : memref<128xi32>
// CHECK-DAG: %[[ACC_MEMREF:.*]] = bufferization.to_memref %[[ACC_TENSOR]] : memref<i32>
// CHECK: %[[OUT_MEMREF_NEW:.*]] = memref.alloc() : memref<128xi32>
// CHECK: %[[ACC_MEMREF_NEW:.*]] = memref.alloc() : memref<i32>
// CHECK: memref.copy %[[ACC_MEMREF]], %[[ACC_MEMREF_NEW]] : memref<i32> to memref<i32>
// CHECK-DAG: %[[OUT_MEMREF_NEW:.*]] = memref.alloc() : memref<128xi32>
// CHECK-DAG: %[[ACC_MEMREF_NEW:.*]] = memref.alloc() : memref<i32>
// CHECK-DAG: %[[OUT_TENSOR_NEW:.*]] = bufferization.to_tensor %[[OUT_MEMREF_NEW]] : memref<128xi32>
// CHECK-DAG: %[[ACC_TENSOR_NEW:.*]] = bufferization.to_tensor %[[ACC_MEMREF_NEW]] : memref<i32>
// CHECK-DAG: memref.copy %[[ACC_MEMREF]], %[[ACC_MEMREF_NEW]] : memref<i32> to memref<i32>
// CHECK: tm_tensor.scan dimension(0) inclusive(false) ins(%[[IN_MEMREF]] : memref<128xi32>)
// CHECK-SAME: outs(%[[OUT_MEMREF_NEW]], %[[ACC_MEMREF_NEW]] : memref<128xi32>, memref<i32>) {
// CHECK: ^bb0(%[[OUT_PREV_ELEMENT:.*]]: i32, %[[IN_ELEMENT:.*]]: i32):
// CHECK: %[[OUT_CURRENT_ELEMENT:.*]] = arith.addi %[[OUT_PREV_ELEMENT]], %[[IN_ELEMENT]] : i32
// CHECK: tm_tensor.yield %[[OUT_CURRENT_ELEMENT]] : i32
// CHECK: }
// CHECK-DAG: %[[OUT_TENSOR_NEW:.*]] = bufferization.to_tensor %[[OUT_MEMREF_NEW]] : memref<128xi32>
// CHECK-DAG: %[[ACC_TENSOR_NEW:.*]] = bufferization.to_tensor %[[ACC_MEMREF_NEW]] : memref<i32>
// CHECK: return %[[OUT_TENSOR_NEW]], %[[ACC_TENSOR_NEW]] : tensor<128xi32>, tensor<i32>
func.func @scan_1d_exclusive(%in: tensor<128xi32>, %out: tensor<128xi32>, %acc: tensor<i32>) -> (tensor<128xi32>, tensor<i32>) {
%ret_out, %ret_acc = tm_tensor.scan dimension(0) inclusive(false)
Expand All @@ -62,14 +62,14 @@ func.func @scan_1d_exclusive(%in: tensor<128xi32>, %out: tensor<128xi32>, %acc:
// CHECK-DAG: %[[UPDATES_MEMREF:.*]] = bufferization.to_memref %[[UPDATES_TENSOR]] : memref<3xi32>
// CHECK-DAG: %[[INDICES_MEMREF:.*]] = bufferization.to_memref %[[INDICES_TENSOR]] : memref<3x1xi32>
// CHECK-DAG: %[[ORIG_MEMREF:.*]] = bufferization.to_memref %[[ORIG_TENSOR]] : memref<8xi32>
// CHECK: %[[ORIG_MEMREF_NEW:.*]] = memref.alloc() : memref<8xi32>
// CHECK-DAG: %[[ORIG_MEMREF_NEW:.*]] = memref.alloc() : memref<8xi32>
// CHECK-DAG: %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[ORIG_MEMREF_NEW]] : memref<8xi32>
// CHECK: memref.copy %[[ORIG_MEMREF]], %[[ORIG_MEMREF_NEW]] : memref<8xi32> to memref<8xi32>
// CHECK: tm_tensor.scatter {dimension_map = array<i64: 0>} unique_indices(true) ins(%[[UPDATES_MEMREF]], %[[INDICES_MEMREF]]
// CHECK-SAME: : memref<3xi32>, memref<3x1xi32>) outs(%[[ORIG_MEMREF_NEW]] : memref<8xi32>) {
// CHECK: ^bb0(%[[UPDATE_SCALAR:.*]]: i32, %[[ORIG_SCALAR:.*]]: i32):
// CHECK: tm_tensor.yield %[[UPDATE_SCALAR]] : i32
// CHECK: }
// CHECK: %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[ORIG_MEMREF_NEW]] : memref<8xi32>
// CHECK: return %[[OUT_TENSOR]] : tensor<8xi32>
func.func @scatter_update_scalar_1D(
%original: tensor<8xi32>, %indices: tensor<3x1xi32>,
Expand All @@ -90,16 +90,16 @@ func.func @scatter_update_scalar_1D(
// CHECK-DAG: %[[UPDATES_MEMREF:.*]] = bufferization.to_memref %[[UPDATES_TENSOR]] : memref<3xi32>
// CHECK-DAG: %[[INDICES_MEMREF:.*]] = bufferization.to_memref %[[INDICES_TENSOR]] : memref<3x1xi32>
// CHECK-DAG: %[[ORIG_MEMREF:.*]] = bufferization.to_memref %[[ORIG_TENSOR]] : memref<8xi32>
// CHECK: %[[ORIG_MEMREF_NEW:.*]] = memref.alloc() : memref<8xi32>
// CHECK: memref.copy %[[ORIG_MEMREF]], %[[ORIG_MEMREF_NEW]] : memref<8xi32> to memref<8xi32>
// CHECK-DAG: %[[ORIG_MEMREF_NEW:.*]] = memref.alloc() : memref<8xi32>
// CHECK-DAG: memref.copy %[[ORIG_MEMREF]], %[[ORIG_MEMREF_NEW]] : memref<8xi32> to memref<8xi32>
// CHECK-DAG: %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[ORIG_MEMREF_NEW]] : memref<8xi32>
// CHECK: tm_tensor.scatter {dimension_map = array<i64: 0>} unique_indices(true) ins(%[[UPDATES_MEMREF]], %[[INDICES_MEMREF]]
// CHECK-SAME: : memref<3xi32>, memref<3x1xi32>) outs(%[[ORIG_MEMREF_NEW]] : memref<8xi32>) {
// CHECK: ^bb0(%[[UPDATE_SCALAR:.*]]: i32, %[[ORIG_SCALAR:.*]]: i32):
// CHECK: %[[CST1:.*]] = arith.constant 1 : i32
// CHECK: %[[ADD:.*]] = arith.addi %[[ORIG_SCALAR]], %[[CST1]] : i32
// CHECK: tm_tensor.yield %[[ADD]] : i32
// CHECK: }
// CHECK: %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[ORIG_MEMREF_NEW]] : memref<8xi32>
// CHECK: return %[[OUT_TENSOR]] : tensor<8xi32>
func.func @scatter_add_scalar_1D(
%original: tensor<8xi32>, %indices: tensor<3x1xi32>,
Expand Down
Loading