colin
/
mlir-hlo
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Avoid Broadcast op if all shapes are (known to) be equal. 

The rank specialization case for shapes which are either of the same shape or a
scalar doesn't need to compute the final result shape.

PiperOrigin-RevId: 380129316
This commit is contained in:
Adrian Kuegel 2021-06-18 00:21:34 -07:00 committed by TensorFlow MLIR Team
parent d4a7901284
commit 4c282fb542
2 changed files with 24 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
lib/Dialect/mhlo/transforms/rank_specialization.cc
View File

@ -15,6 +15,7 @@ limitations under the License.
==============================================================================*/
#include "llvm/ADT/EquivalenceClasses.h"
#include "llvm/ADT/Optional.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
@ -341,17 +342,20 @@ SmallVector<Value, 8> MaterializeRankedOperations(
SmallVector<Value, 8> MaterializeFinalReshape(
OpBuilder &b, Location loc, chlo::RankSpecializationClusterOp op,
ValueRange unshaped_results) {
ValueRange unshaped_results, llvm::Optional<Value> shape = {}) {
if (!shape) {
// Compute result shape.
auto non_scalar_operands = llvm::make_filter_range(
op.operands(), [](Value v) { return !IsScalarTensorType(v.getType()); });
op.operands(),
[](Value v) { return !IsScalarTensorType(v.getType()); });
SmallVector<Value, 8> results;
auto operand_shapes =
llvm::to_vector<8>(llvm::map_range(non_scalar_operands, [&](Value v) {
return b.create<shape::ShapeOfOp>(loc, v).result();
}));
auto shape = b.create<shape::BroadcastOp>(
shape = b.create<shape::BroadcastOp>(
loc, shape::getExtentTensorType(b.getContext()), operand_shapes);
}
// Reshape results.
return llvm::to_vector<8>(
@ -359,7 +363,7 @@ SmallVector<Value, 8> MaterializeFinalReshape(
return b
.create<mhlo::DynamicReshapeOp>(
loc, DeriveUnrankedTensorTypes(unshaped.getType()), unshaped,
shape)
shape.getValue())
.result();
}));
}
@ -664,7 +668,8 @@ MaterializeRankSpecializationForSingleNonScalarShapeEquivalenceClass(
MaterializeRankedOperations(b, loc, bvm, op);
// Restore the results' expected shape.
return MaterializeFinalReshape(b, loc, op, unshaped_results);
return MaterializeFinalReshape(b, loc, op, unshaped_results,
non_scalar_shapes.front());
}
Value MaterializeRankSpecializationForTwoNonScalarShapeEquivalenceClasses(

22
tests/rank-specialization.mlir
View File

@ -159,8 +159,7 @@ func @sqrt(%arg : tensor<*xf32>) -> tensor<*xf32> {
// CHECK-SCF: %[[TMP0:.*]] = "mhlo.sqrt"(%[[FLAT_ARG]]) : (tensor<?xf32>)
// CHECK-SCF: %[[TMP1:.*]] = "mhlo.sqrt"(%[[TMP0]]) : (tensor<?xf32>)
// CHECK-SCF: %[[UNSHAPED_RES:.*]] = "mhlo.sqrt"(%[[TMP1]]) : (tensor<?xf32>)
// CHECK-SCF: %[[RES_SHAPE:.*]] = shape.shape_of %[[ARG]]
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[RES_SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: return %[[RES]]
// -----
@ -229,8 +228,7 @@ func @tan(%arg : tensor<*xf32>) -> tensor<*xf32> {
// CHECK-SCF: %[[TMP0:.*]] = chlo.tan %[[FLAT_ARG]] : tensor<?xf32>
// CHECK-SCF: %[[TMP1:.*]] = chlo.tan %[[TMP0]] : tensor<?xf32>
// CHECK-SCF: %[[UNSHAPED_RES:.*]] = chlo.tan %[[TMP1]] : tensor<?xf32>
// CHECK-SCF: %[[RES_SHAPE:.*]] = shape.shape_of %[[ARG]]
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[RES_SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: return %[[RES]]
// -----
@ -295,8 +293,7 @@ func @relu(%arg : tensor<*xf32>) -> tensor<*xf32> {
// CHECK-SCF: %[[FLAT_SHAPE:.*]] = tensor.from_elements %[[N]]
// CHECK-SCF: %[[FLAT_ARG:.*]] = "mhlo.dynamic_reshape"(%[[ARG]], %[[FLAT_SHAPE]]) : (tensor<*xf32>, tensor<1xindex>) -> tensor<?xf32>
// CHECK-SCF: %[[UNSHAPED_RES:.*]] = chlo.broadcast_maximum %[[FLAT_ARG]], %[[C0]] : (tensor<?xf32>, tensor<f32>)
// CHECK-SCF: %[[RES_SHAPE:.*]] = shape.shape_of %[[ARG]]
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[RES_SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: return %[[RES]]
// -----
@ -327,8 +324,7 @@ func @angle(%arg : tensor<*xcomplex<f32>>) -> tensor<*xf32> {
// CHECK-SCF: %[[IMAG:.*]] = "mhlo.imag"(%[[FLAT_ARG]]) : (tensor<?xcomplex<f32>>)
// CHECK-SCF: %[[REAL:.*]] = "mhlo.real"(%[[FLAT_ARG]]) : (tensor<?xcomplex<f32>>)
// CHECK-SCF: %[[UNSHAPED_RES:.*]] = mhlo.atan2 %[[IMAG]], %[[REAL]] : tensor<?xf32>
// CHECK-SCF: %[[RES_SHAPE:.*]] = shape.shape_of %[[ARG]]
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[RES_SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[SHAPE]]) : (tensor<?xf32>, tensor<?xindex>) -> tensor<*xf32>
// CHECK-SCF: return %[[RES]]
// -----
@ -610,10 +606,7 @@ func @all_equal_shapes_inferrable(%arg0: tensor<*xf64>, %arg1 : tensor<*xf64>)
// CHECK-SCF-DAG: %[[FLAT0:.*]] = "mhlo.dynamic_reshape"(%[[ARG0]], %[[FLAT_S]])
// CHECK-SCF-DAG: %[[FLAT1:.*]] = "mhlo.dynamic_reshape"(%[[ARG1]], %[[FLAT_S]])
// CHECK-SCF: %[[FLAT_RES:.*]] = mhlo.add %[[FLAT0]], %[[FLAT1]]
// CHECK-SCF-DAG: %[[S0:.*]] = shape.shape_of %[[ARG0]]
// CHECK-SCF-DAG: %[[S1:.*]] = shape.shape_of %[[ARG1]]
// CHECK-SCF-DAG: %[[RES_S:.*]] = shape.broadcast %8, %9
// CHECK-SCF-DAG: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[FLAT_RES]], %[[RES_S]])
// CHECK-SCF-DAG: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[FLAT_RES]], %[[S0]])
// CHECK-SCF: return %[[RES]]
// -----
@ -647,8 +640,5 @@ func @relu_grad(%arg0: tensor<*xf32>, %arg1: tensor<*xf32>) -> tensor<*xf32> {
// CHECK-SCF-DAG: %[[ZERO:.*]] = "chlo.constant_like"(%[[FLAT0]]) {value = 0.0{{.*}}+00 : f32}
// CHECK-SCF-DAG: %[[PRED:.*]] = "mhlo.compare"(%[[FLAT0]], %[[ZERO]]) {comparison_direction = "GT"}
// CHECK-SCF: %[[UNSHAPED_RES:.*]] = "mhlo.select"(%[[PRED]], %[[FLAT1]], %[[ZERO]])
// CHECK-SCF-DAG: %[[S0:.*]] = shape.shape_of %[[ARG0]]
// CHECK-SCF-DAG: %[[S1:.*]] = shape.shape_of %[[ARG1]]
// CHECK-SCF-DAG: %[[RES_SHAPE:.*]] = shape.broadcast %[[S1]], %[[S0]]
// CHECK-SCF-DAG: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[RES_SHAPE]])
// CHECK-SCF-DAG: %[[RES:.*]] = "mhlo.dynamic_reshape"(%[[UNSHAPED_RES]], %[[S1]])
// CHECK-SCF: return %[[RES]]