[mlir][tosa] Added folders for tosa.sub

[rsuderman]

Summary

Added folders for tosa.sub that handles bypassing sub-zero,
fold subtraction of two splat tensors.

mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td

def Tosa_SubOp : Tosa_Op<"sub", [
   let arguments = (ins
     Tosa_Tensor:$input1,
     Tosa_Tensor:$input2
   );
 
   let results = (outs
     Tosa_Tensor:$output
   );
+
+  let hasFolder = 1;
 }
 
 //===----------------------------------------------------------------------===//
 // Operator: table
 //===----------------------------------------------------------------------===//
 def Tosa_TableOp : Tosa_Op<"table", [
     DeclareOpInterfaceMethods<InferShapedTypeOpInterface,
                               ["inferReturnTypeComponents"]>,

mlir/lib/Dialect/Tosa/IR/TosaCanonicalizations.cpp

OpFoldResult AddOp::fold(ArrayRef<Attribute> operands) {
 
   if (!lhsAttr || !rhsAttr)
     return {};
 
   return BinaryFolder<std::plus<APInt>, std::plus<APFloat>>(lhsAttr, rhsAttr,
                                                             lhsTy);
 }
 
+OpFoldResult SubOp::fold(ArrayRef<Attribute> operands) {
+  auto lhsTy = getInput1().getType().dyn_cast<RankedTensorType>();
+  auto rhsTy = getInput2().getType().dyn_cast<RankedTensorType>();
+  auto resultTy = getType().dyn_cast<RankedTensorType>();
+  if (!lhsTy || !rhsTy || !resultTy)
+    return {};
+  if (lhsTy != rhsTy)
+    return {};
+
+  auto resultETy = resultTy.getElementType();
+  auto lhsAttr = operands[0].dyn_cast_or_null<DenseElementsAttr>();
+  auto rhsAttr = operands[1].dyn_cast_or_null<DenseElementsAttr>();
+
+  if (rhsAttr && rhsAttr.isSplat() && resultETy.isa<FloatType>()) {
+    if (rhsAttr.getSplatValue<APFloat>().isZero())
+      return getInput1();
+  }
+
+  if (rhsAttr && rhsAttr.isSplat() && resultETy.isa<IntegerType>()) {
+    if (rhsAttr.getSplatValue<APInt>().isZero())
+      return getInput1();
+  }
+
+  if (!lhsAttr || !rhsAttr)
+    return {};
+
+  return BinaryFolder<std::minus<APInt>, std::minus<APFloat>>(lhsAttr, rhsAttr,
+                                                              lhsTy);
+}
+
 namespace {
 template <typename Cmp>
 struct ComparisonFold {
   ComparisonFold() {}
   APInt operator()(const APInt &l, const APInt &r) {
     return APInt(1, Cmp()(l, r));
   }
 

mlir/test/Dialect/Tosa/constant-op-fold.mlir

func.func @fold_add_splat_f32() -> tensor<10xf32> {
   %add = "tosa.add"(%one, %two) : (tensor<10xf32>, tensor<10xf32>) -> tensor<10xf32>
   // CHECK: %[[THREE:.+]] = "tosa.const"() {value = dense<3.000000e+00> : tensor<10xf32>}
   // CHECK: return %[[THREE]]
   return %add : tensor<10xf32>
 }
 
 // -----
 
+// CHECK-LABEL: @fold_sub_zero_rhs_f32
+func.func @fold_sub_zero_rhs_f32(%arg0: tensor<f32>) -> tensor<f32> {
+  %zero = "tosa.const"() {value = dense<0.0> : tensor<f32>} : () -> tensor<f32>
+  %sub = "tosa.sub"(%arg0, %zero) : (tensor<f32>, tensor<f32>) -> tensor<f32>
+  // CHECK: return %arg0
+  return %sub : tensor<f32>
+}
+
+// -----
+
+// CHECK-LABEL: @fold_sub_zero_rhs_i32
+func.func @fold_sub_zero_rhs_i32(%arg0: tensor<i32>) -> tensor<i32> {
+  %zero = "tosa.const"() {value = dense<0> : tensor<i32>} : () -> tensor<i32>
+  %sub = "tosa.sub"(%arg0, %zero) : (tensor<i32>, tensor<i32>) -> tensor<i32>
+  // CHECK: return %arg0
+  return %sub : tensor<i32>
+}
+
+// -----
+
+// CHECK-LABEL: @fold_sub_splat_i32
+func.func @fold_sub_splat_i32() -> tensor<10xi32> {
+  %one = "tosa.const"() {value = dense<1> : tensor<10xi32>} : () -> tensor<10xi32>
+  %two = "tosa.const"() {value = dense<2> : tensor<10xi32>} : () -> tensor<10xi32>
+  %sub = "tosa.sub"(%one, %two) : (tensor<10xi32>, tensor<10xi32>) -> tensor<10xi32>
+  // CHECK: %[[THREE:.+]] = "tosa.const"() {value = dense<-1> : tensor<10xi32>}
+  // CHECK: return %[[THREE]]
+  return %sub : tensor<10xi32>
+}
+
+// -----
+
+// CHECK-LABEL: @fold_sub_splat_f32
+func.func @fold_sub_splat_f32() -> tensor<10xf32> {
+  %one = "tosa.const"() {value = dense<1.0> : tensor<10xf32>} : () -> tensor<10xf32>
+  %two = "tosa.const"() {value = dense<2.0> : tensor<10xf32>} : () -> tensor<10xf32>
+  %sub = "tosa.sub"(%one, %two) : (tensor<10xf32>, tensor<10xf32>) -> tensor<10xf32>
+  // CHECK: %[[THREE:.+]] = "tosa.const"() {value = dense<-1.000000e+00> : tensor<10xf32>}
+  // CHECK: return %[[THREE]]
+  return %sub : tensor<10xf32>
+}
+
+// -----
+
 // CHECK-LABEL: @fold_greater_splat_f32_true
 func.func @fold_greater_splat_f32_true() -> tensor<10xi1> {
   %one = "tosa.const"() {value = dense<4.0> : tensor<10xf32>} : () -> tensor<10xf32>
   %two = "tosa.const"() {value = dense<2.0> : tensor<10xf32>} : () -> tensor<10xf32>
   %add = "tosa.greater"(%one, %two) : (tensor<10xf32>, tensor<10xf32>) -> tensor<10xi1>
   // CHECK: %[[BOOL:.+]] = "tosa.const"() {value = dense<true> : tensor<10xi1>}
   // CHECK: return %[[BOOL]]
   return %add : tensor<10xi1>