diff options
| author | Nicolas Vasilache <ntv@google.com> | 2019-11-14 08:10:36 -0800 |
|---|---|---|
| committer | A. Unique TensorFlower <gardener@tensorflow.org> | 2019-11-14 08:15:23 -0800 |
| commit | f2b6ae99913d0049c7929160aed5f213b1081abb (patch) | |
| tree | 41ccb1f9a181eaada07f21cbc7da75ef2d229575 /mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir | |
| parent | 7c28de4aef6da3ab2f53118ecf717e56c68352e7 (diff) | |
| download | bcm5719-llvm-f2b6ae99913d0049c7929160aed5f213b1081abb.tar.gz bcm5719-llvm-f2b6ae99913d0049c7929160aed5f213b1081abb.zip | |
Move VectorOps to Tablegen - (almost) NFC
This CL moves VectorOps to Tablegen and cleans up the implementation.
This is almost NFC but 2 changes occur:
1. an interface change occurs in the padding value specification in vector_transfer_read:
the value becomes non-optional. As a shortcut we currently use %f0 for all paddings.
This should become an OpInterface for vectorization in the future.
2. the return type of vector.type_cast is trivial and simplified to `memref<vector<...>>`
Relevant roundtrip and invalid tests that used to sit in core are moved to the vector dialect.
The op documentation is moved to the .td file.
PiperOrigin-RevId: 280430869
Diffstat (limited to 'mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir')
| -rw-r--r-- | mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir b/mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir index 4f61a26b4a8..765cd07ce7d 100644 --- a/mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir +++ b/mlir/test/Transforms/Vectorize/vectorize_transpose_2d.mlir @@ -22,7 +22,7 @@ func @vec2d(%A : memref<?x?x?xf32>) { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 32 // CHECK: affine.for %{{.*}} = 0 to %{{.*}} { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 256 - // CHECK: {{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> + // CHECK: {{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}], %{{.*}} {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> affine.for %i3 = 0 to %M { affine.for %i4 = 0 to %N { affine.for %i5 = 0 to %P { @@ -40,12 +40,12 @@ func @vec2d_imperfectly_nested(%A : memref<?x?x?xf32>) { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 32 { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 256 { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} { - // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> + // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}], %{{.*}} {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> // CHECK: affine.for %{{.*}} = 0 to %{{.*}} { // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 256 { - // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> + // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}], %{{.*}} {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> // CHECK: affine.for %{{.*}} = 0 to %{{.*}} step 256 { - // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> + // CHECK: %{{.*}} = vector.transfer_read %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}], %{{.*}} {permutation_map = #[[map_proj_d0d1d2_d2d1]]} : memref<?x?x?xf32>, vector<32x256xf32> affine.for %i0 = 0 to %0 { affine.for %i1 = 0 to %1 { affine.for %i2 = 0 to %2 { |
