[mlir][SCF] NFC - Drop SCF EDSC usage

[nicolasvasilache]

Summary

Drop the SCF dialect EDSC subdirectory and update all uses.

[ftynse]

mlir/lib/Dialect/Linalg/Transforms/Loops.cpp

27–60

This could use a doc.

[mehdi_amini]

mlir/lib/Dialect/Linalg/Transforms/Loops.cpp

27–60

Anonymous namespace?

[nicolasvasilache]

Address.

[nicolasvasilache]

Thanks!

mlir/include/mlir/Dialect/Affine/EDSC/Builders.h

 Value sge(Value lhs, Value rhs);
 Value ult(Value lhs, Value rhs);
 Value ule(Value lhs, Value rhs);
 Value ugt(Value lhs, Value rhs);
 Value uge(Value lhs, Value rhs);
 
 } // namespace op
 
-/// Arithmetic operator overloadings.
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator+(Value e) {
-  using op::operator+;
-  return static_cast<Value>(*this) + e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator-(Value e) {
-  using op::operator-;
-  return static_cast<Value>(*this) - e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator*(Value e) {
-  using op::operator*;
-  return static_cast<Value>(*this) * e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator/(Value e) {
-  using op::operator/;
-  return static_cast<Value>(*this) / e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator%(Value e) {
-  using op::operator%;
-  return static_cast<Value>(*this) % e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator^(Value e) {
-  using op::operator^;
-  return static_cast<Value>(*this) ^ e;
-}
-
-/// Assignment-arithmetic operator overloadings.
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator+=(Value e) {
-  using op::operator+;
-  return Store(*this + e, getBase(), indices);
-}
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator-=(Value e) {
-  using op::operator-;
-  return Store(*this - e, getBase(), indices);
-}
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator*=(Value e) {
-  using op::operator*;
-  return Store(*this * e, getBase(), indices);
-}
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator/=(Value e) {
-  using op::operator/;
-  return Store(*this / e, getBase(), indices);
-}
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator%=(Value e) {
-  using op::operator%;
-  return Store(*this % e, getBase(), indices);
-}
-template <typename Load, typename Store>
-Store TemplatedIndexedValue<Load, Store>::operator^=(Value e) {
-  using op::operator^;
-  return Store(*this ^ e, getBase(), indices);
-}
-
-/// Logical operator overloadings.
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator&&(Value e) {
-  using op::operator&&;
-  return static_cast<Value>(*this) && e;
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::operator||(Value e) {
-  using op::operator||;
-  return static_cast<Value>(*this) || e;
-}
-
-/// Comparison operator overloadings.
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::eq(Value e) {
-  return eq(value, e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::ne(Value e) {
-  return ne(value, e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::slt(Value e) {
-  using op::slt;
-  return slt(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::sle(Value e) {
-  using op::sle;
-  return sle(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::sgt(Value e) {
-  using op::sgt;
-  return sgt(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::sge(Value e) {
-  using op::sge;
-  return sge(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::ult(Value e) {
-  using op::ult;
-  return ult(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::ule(Value e) {
-  using op::ule;
-  return ule(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::ugt(Value e) {
-  using op::ugt;
-  return ugt(static_cast<Value>(*this), e);
-}
-template <typename Load, typename Store>
-Value TemplatedIndexedValue<Load, Store>::uge(Value e) {
-  using op::uge;
-  return uge(static_cast<Value>(*this), e);
-}
-
 } // namespace edsc
 } // namespace mlir
 
 #endif // MLIR_DIALECT_AFFINE_EDSC_BUILDERS_H_

mlir/include/mlir/Dialect/Affine/EDSC/Intrinsics.h

 
 using affine_apply = ValueBuilder<AffineApplyOp>;
 using affine_if = OperationBuilder<AffineIfOp>;
 using affine_load = ValueBuilder<AffineLoadOp>;
 using affine_min = ValueBuilder<AffineMinOp>;
 using affine_max = ValueBuilder<AffineMaxOp>;
 using affine_store = OperationBuilder<AffineStoreOp>;
 
-/// Provide an index notation around affine_load and affine_store.
-using AffineIndexedValue = TemplatedIndexedValue<affine_load, affine_store>;
-
 } // namespace intrinsics
 } // namespace edsc
 } // namespace mlir
 
 #endif // MLIR_DIALECT_STANDARDOPS_EDSC_INTRINSICS_H_

mlir/include/mlir/Dialect/Linalg/Utils/Utils.h

 
 //===----------------------------------------------------------------------===//
 // Loop nest utilities
 //===----------------------------------------------------------------------===//
 
 /// Utility class used to generate nested loops with ranges described by
 /// `loopRanges` and loop type described by the `iteratorTypes`. `bodyBuilderFn`
 /// is used to generate the body of the innermost loop. It is passed a range
-/// of loop induction variables.
+/// of loop induction variables and a range of iterArgs.
 template <typename LoopTy>
 struct GenerateLoopNest {
-  using IndexedValueTy =
-      typename std::conditional<std::is_same<LoopTy, AffineForOp>::value,
-                                edsc::intrinsics::AffineIndexedValue,
-                                edsc::intrinsics::MemRefIndexedValue>::type;
-
-  static void
-  doit(ArrayRef<Range> loopRanges, LinalgOp linalgOp,
-       ArrayRef<Attribute> iteratorTypes,
-       function_ref<scf::ValueVector(ValueRange, ValueRange)> bodyBuilderFn,
+  static void doit(OpBuilder &b, Location loc, ArrayRef<Range> loopRanges,
+                   LinalgOp linalgOp, ArrayRef<Attribute> iteratorTypes,
+                   function_ref<scf::ValueVector(OpBuilder &, Location,
+                                                 ValueRange, ValueRange)>
+                       bodyBuilderFn,
                    Optional<LinalgLoopDistributionOptions> = None);
 };
 
 } // namespace linalg
 } // namespace mlir
 
 #endif // MLIR_DIALECT_LINALG_UTILS_H_

mlir/include/mlir/Dialect/MemRef/EDSC/Intrinsics.h

 using memref_dim = ValueBuilder<memref::DimOp>;
 using memref_load = ValueBuilder<memref::LoadOp>;
 using memref_store = OperationBuilder<memref::StoreOp>;
 using memref_sub_view = ValueBuilder<memref::SubViewOp>;
 using memref_tensor_load = ValueBuilder<memref::TensorLoadOp>;
 using memref_tensor_store = OperationBuilder<memref::TensorStoreOp>;
 using memref_view = ValueBuilder<memref::ViewOp>;
 
-/// Provide an index notation around memref_load and memref_store.
-using MemRefIndexedValue =
-    TemplatedIndexedValue<intrinsics::memref_load, intrinsics::memref_store>;
 } // namespace intrinsics
 } // namespace edsc
 } // namespace mlir
 
 static inline ::llvm::SmallVector<mlir::Value, 8>
 getMemRefSizes(mlir::Value memRef) {
   using namespace mlir;
   using namespace mlir::edsc;

mlir/include/mlir/Dialect/SCF/EDSC/Builders.h

-//===- Builders.h - MLIR Declarative Builder Classes ------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// Provides intuitive composable interfaces for building structured MLIR
-// snippets in a declarative fashion.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef MLIR_DIALECT_SCF_EDSC_BUILDERS_H_
-#define MLIR_DIALECT_SCF_EDSC_BUILDERS_H_
-
-#include "mlir/Dialect/SCF/SCF.h"
-#include "mlir/EDSC/Builders.h"
-#include "mlir/IR/Builders.h"
-#include "mlir/IR/Types.h"
-
-namespace mlir {
-namespace edsc {
-
-/// Adapters for building loop nests using the builder and the location stored
-/// in ScopedContext. Actual builders are in scf::buildLoopNest.
-scf::LoopNest loopNestBuilder(ValueRange lbs, ValueRange ubs,
-                                 ValueRange steps,
-                                 function_ref<void(ValueRange)> fun = nullptr);
-scf::LoopNest loopNestBuilder(Value lb, Value ub, Value step,
-                                 function_ref<void(Value)> fun = nullptr);
-scf::LoopNest loopNestBuilder(
-    Value lb, Value ub, Value step, ValueRange iterArgInitValues,
-    function_ref<scf::ValueVector(Value, ValueRange)> fun = nullptr);
-scf::LoopNest loopNestBuilder(
-    ValueRange lbs, ValueRange ubs, ValueRange steps,
-    ValueRange iterArgInitValues,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> fun = nullptr);
-
-/// Adapters for building if conditions using the builder and the location
-/// stored in ScopedContext. 'thenBody' is mandatory, 'elseBody' can be omitted
-/// if the condition should not have an 'else' part.
-/// When `ifOp` is specified, the scf::IfOp is captured. This is particularly
-/// convenient for 0-result conditions.
-ValueRange conditionBuilder(TypeRange results, Value condition,
-                            function_ref<scf::ValueVector()> thenBody,
-                            function_ref<scf::ValueVector()> elseBody = nullptr,
-                            scf::IfOp *ifOp = nullptr);
-ValueRange conditionBuilder(Value condition, function_ref<void()> thenBody,
-                            function_ref<void()> elseBody = nullptr,
-                            scf::IfOp *ifOp = nullptr);
-
-} // namespace edsc
-} // namespace mlir
-
-#endif // MLIR_DIALECT_SCF_EDSC_BUILDERS_H_

mlir/include/mlir/Dialect/SCF/EDSC/Intrinsics.h

-//===- Intrinsics.h - MLIR EDSC Intrinsics for SCF --------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM
-// Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-#ifndef MLIR_DIALECT_SCF_EDSC_INTRINSICS_H_
-#define MLIR_DIALECT_SCF_EDSC_INTRINSICS_H_
-
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
-
-namespace mlir {
-namespace edsc {
-namespace intrinsics {
-
-using loop_yield = OperationBuilder<scf::YieldOp>;
-
-} // namespace intrinsics
-} // namespace edsc
-} // namespace mlir
-
-#endif // MLIR_DIALECT_SCF_EDSC_INTRINSICS_H_

mlir/include/mlir/EDSC/Builders.h

 
 private:
   // Only one of Value or type may be set.
   Type type;
   Value value;
   SmallVector<AffineExpr, 4> exprs;
 };
 
-/// A TemplatedIndexedValue brings an index notation over the template Load and
-/// Store parameters. Assigning to an IndexedValue emits an actual `Store`
-/// operation, while converting an IndexedValue to a Value emits an actual
-/// `Load` operation.
-template <typename Load, typename Store>
-class TemplatedIndexedValue {
-public:
-  explicit TemplatedIndexedValue(Value v) : value(v) {}
-
-  TemplatedIndexedValue(const TemplatedIndexedValue &rhs) = default;
-
-  TemplatedIndexedValue operator()() { return *this; }
-  /// Returns a new `TemplatedIndexedValue`.
-  TemplatedIndexedValue operator()(Value index) {
-    TemplatedIndexedValue res(value);
-    res.indices.push_back(index);
-    return res;
-  }
-  template <typename... Args>
-  TemplatedIndexedValue operator()(Value index, Args... indices) {
-    return TemplatedIndexedValue(value, index).append(indices...);
-  }
-  TemplatedIndexedValue operator()(ValueRange indices) {
-    return TemplatedIndexedValue(value, indices);
-  }
-
-  /// Emits a `store`.
-  Store operator=(const TemplatedIndexedValue &rhs) {
-    return Store(rhs, value, indices);
-  }
-  Store operator=(Value rhs) { return Store(rhs, value, indices); }
-
-  /// Emits a `load` when converting to a Value.
-  operator Value() const { return Load(value, indices); }
-
-  /// Returns the base memref.
-  Value getBase() const { return value; }
-
-  /// Returns the underlying memref.
-  MemRefType getMemRefType() const {
-    return value.getType().template cast<MemRefType>();
-  }
-
-  /// Returns the underlying MemRef elemental type cast as `T`.
-  template <typename T>
-  T getElementalTypeAs() const {
-    return value.getType()
-        .template cast<MemRefType>()
-        .getElementType()
-        .template cast<T>();
-  }
-
-  /// Arithmetic operator overloadings.
-  Value operator+(Value e);
-  Value operator-(Value e);
-  Value operator*(Value e);
-  Value operator/(Value e);
-  Value operator%(Value e);
-  Value operator^(Value e);
-  Value operator+(TemplatedIndexedValue e) {
-    return *this + static_cast<Value>(e);
-  }
-  Value operator-(TemplatedIndexedValue e) {
-    return *this - static_cast<Value>(e);
-  }
-  Value operator*(TemplatedIndexedValue e) {
-    return *this * static_cast<Value>(e);
-  }
-  Value operator/(TemplatedIndexedValue e) {
-    return *this / static_cast<Value>(e);
-  }
-  Value operator%(TemplatedIndexedValue e) {
-    return *this % static_cast<Value>(e);
-  }
-  Value operator^(TemplatedIndexedValue e) {
-    return *this ^ static_cast<Value>(e);
-  }
-
-  /// Assignment-arithmetic operator overloadings.
-  Store operator+=(Value e);
-  Store operator-=(Value e);
-  Store operator*=(Value e);
-  Store operator/=(Value e);
-  Store operator%=(Value e);
-  Store operator^=(Value e);
-  Store operator+=(TemplatedIndexedValue e) {
-    return this->operator+=(static_cast<Value>(e));
-  }
-  Store operator-=(TemplatedIndexedValue e) {
-    return this->operator-=(static_cast<Value>(e));
-  }
-  Store operator*=(TemplatedIndexedValue e) {
-    return this->operator*=(static_cast<Value>(e));
-  }
-  Store operator/=(TemplatedIndexedValue e) {
-    return this->operator/=(static_cast<Value>(e));
-  }
-  Store operator%=(TemplatedIndexedValue e) {
-    return this->operator%=(static_cast<Value>(e));
-  }
-  Store operator^=(TemplatedIndexedValue e) {
-    return this->operator^=(static_cast<Value>(e));
-  }
-
-  /// Logical operator overloadings.
-  Value operator&&(Value e);
-  Value operator||(Value e);
-  Value operator&&(TemplatedIndexedValue e) {
-    return *this && static_cast<Value>(e);
-  }
-  Value operator||(TemplatedIndexedValue e) {
-    return *this || static_cast<Value>(e);
-  }
-
-  /// Comparison operator overloadings.
-  Value eq(Value e);
-  Value ne(Value e);
-  Value slt(Value e);
-  Value sle(Value e);
-  Value sgt(Value e);
-  Value sge(Value e);
-  Value ult(Value e);
-  Value ule(Value e);
-  Value ugt(Value e);
-  Value uge(Value e);
-  Value slt(TemplatedIndexedValue e) {
-    return slt(*this, static_cast<Value>(e));
-  }
-  Value sle(TemplatedIndexedValue e) {
-    return sle(*this, static_cast<Value>(e));
-  }
-  Value sgt(TemplatedIndexedValue e) {
-    return sgt(*this, static_cast<Value>(e));
-  }
-  Value sge(TemplatedIndexedValue e) {
-    return sge(*this, static_cast<Value>(e));
-  }
-  Value ult(TemplatedIndexedValue e) {
-    return ult(*this, static_cast<Value>(e));
-  }
-  Value ule(TemplatedIndexedValue e) {
-    return ule(*this, static_cast<Value>(e));
-  }
-  Value ugt(TemplatedIndexedValue e) {
-    return ugt(*this, static_cast<Value>(e));
-  }
-  Value uge(TemplatedIndexedValue e) {
-    return uge(*this, static_cast<Value>(e));
-  }
-
-private:
-  TemplatedIndexedValue(Value value, ValueRange indices)
-      : value(value), indices(indices.begin(), indices.end()) {}
-
-  TemplatedIndexedValue &append() { return *this; }
-
-  template <typename T, typename... Args>
-  TemplatedIndexedValue &append(T index, Args... indices) {
-    this->indices.push_back(static_cast<Value>(index));
-    append(indices...);
-    return *this;
-  }
-  Value value;
-  SmallVector<Value, 8> indices;
-};
-
 } // namespace edsc
 } // namespace mlir
 
 #endif // MLIR_EDSC_BUILDERS_H_

mlir/lib/Dialect/GPU/Transforms/MemoryPromotion.cpp

 //===- MemoryPromotion.cpp - Utilities for moving data across GPU memories ===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 //
 // This file implements utilities that allow one to create IR moving the data
 // across different levels of the GPU memory hierarchy.
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/GPU/MemoryPromotion.h"
 #include "mlir/Dialect/GPU/GPUDialect.h"
 #include "mlir/Dialect/MemRef/EDSC/Intrinsics.h"
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
+#include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/StandardOps/EDSC/Intrinsics.h"
 #include "mlir/Pass/Pass.h"
 #include "mlir/Transforms/LoopUtils.h"
 
 using namespace mlir;
 using namespace mlir::edsc;
 using namespace mlir::edsc::intrinsics;
 using namespace mlir::gpu;
static StringRef getDimName(unsigned dim) {
   llvm_unreachable("dimension ID overflow");
 }
 
 /// Emits the (imperfect) loop nest performing the copy between "from" and "to"
 /// values using the bounds derived from the "from" value. Emits at least
 /// GPUDialect::getNumWorkgroupDimensions() loops, completing the nest with
 /// single-iteration loops. Maps the innermost loops to thread dimensions, in
 /// reverse order to enable access coalescing in the innermost loop.
-static void insertCopyLoops(OpBuilder &builder, Location loc,
+static void insertCopyLoops(OpBuilder &b, Location loc,
                             MemRefBoundsCapture &bounds, Value from, Value to) {
   // Create EDSC handles for bounds.
   unsigned rank = bounds.rank();
   SmallVector<Value, 4> lbs, ubs, steps;
 
   // Make sure we have enough loops to use all thread dimensions, these trivial
   // loops should be outermost and therefore inserted first.
   if (rank < GPUDialect::getNumWorkgroupDimensions()) {
static void insertCopyLoops(OpBuilder &b, Location loc,
   ubs.append(bounds.getUbs().begin(), bounds.getUbs().end());
 
   // Emit constant operations for steps.
   steps.reserve(lbs.size());
   llvm::transform(bounds.getSteps(), std::back_inserter(steps),
                   [](int64_t step) { return std_constant_index(step); });
 
   // Obtain thread identifiers and block sizes, necessary to map to them.
-  auto indexType = builder.getIndexType();
+  auto indexType = b.getIndexType();
   SmallVector<Value, 3> threadIds, blockDims;
   for (unsigned i = 0; i < 3; ++i) {
-    auto dimName = builder.getStringAttr(getDimName(i));
-    threadIds.push_back(
-        builder.create<gpu::ThreadIdOp>(loc, indexType, dimName));
-    blockDims.push_back(
-        builder.create<gpu::BlockDimOp>(loc, indexType, dimName));
+    auto dimName = b.getStringAttr(getDimName(i));
+    threadIds.push_back(b.create<gpu::ThreadIdOp>(loc, indexType, dimName));
+    blockDims.push_back(b.create<gpu::BlockDimOp>(loc, indexType, dimName));
   }
 
   // Produce the loop nest with copies.
   SmallVector<Value, 8> ivs(lbs.size());
-  loopNestBuilder(lbs, ubs, steps, [&](ValueRange loopIvs) {
+  mlir::scf::buildLoopNest(
+      b, loc, lbs, ubs, steps,
+      [&](OpBuilder &b, Location loc, ValueRange loopIvs) {
         ivs.assign(loopIvs.begin(), loopIvs.end());
         auto activeIvs = llvm::makeArrayRef(ivs).take_back(rank);
-    MemRefIndexedValue fromHandle(from), toHandle(to);
-    toHandle(activeIvs) = fromHandle(activeIvs);
+        Value loaded = b.create<memref::LoadOp>(loc, from, activeIvs);
+        b.create<memref::StoreOp>(loc, loaded, to, activeIvs);
       });
 
   // Map the innermost loops to threads in reverse order.
   for (auto en :
        llvm::enumerate(llvm::reverse(llvm::makeArrayRef(ivs).take_back(
            GPUDialect::getNumWorkgroupDimensions())))) {
     Value v = en.value();
     auto loop = cast<scf::ForOp>(v.getParentRegion()->getParentOp());
     mapLoopToProcessorIds(loop, {threadIds[en.index()]},
static void insertCopies(Region &region, Location loc, Value from, Value to) {
   auto toType = to.getType().cast<MemRefType>();
   (void)fromType;
   (void)toType;
   assert(fromType.getShape() == toType.getShape());
   assert(fromType.getRank() != 0);
   assert(llvm::hasSingleElement(region) &&
          "unstructured control flow not supported");
 
-  OpBuilder builder(region.getContext());
-  builder.setInsertionPointToStart(&region.front());
+  OpBuilder b(region.getContext());
+  b.setInsertionPointToStart(&region.front());
 
-  ScopedContext edscContext(builder, loc);
+  ScopedContext edscContext(b, loc);
   MemRefBoundsCapture fromBoundsCapture(from);
-  insertCopyLoops(builder, loc, fromBoundsCapture, from, to);
-  builder.create<gpu::BarrierOp>(loc);
+  insertCopyLoops(b, loc, fromBoundsCapture, from, to);
+  b.create<gpu::BarrierOp>(loc);
 
-  builder.setInsertionPoint(&region.front().back());
-  builder.create<gpu::BarrierOp>(loc);
-  insertCopyLoops(builder, loc, fromBoundsCapture, to, from);
+  b.setInsertionPoint(&region.front().back());
+  b.create<gpu::BarrierOp>(loc);
+  insertCopyLoops(b, loc, fromBoundsCapture, to, from);
 }
 
 /// Promotes a function argument to workgroup memory in the given function. The
 /// copies will be inserted in the beginning and in the end of the function.
 void mlir::promoteToWorkgroupMemory(GPUFuncOp op, unsigned arg) {
   Value value = op.getArgument(arg);
   auto type = value.getType().dyn_cast<MemRefType>();
   assert(type && type.hasStaticShape() && "can only promote memrefs");

mlir/lib/Dialect/Linalg/EDSC/Builders.cpp

 //===- Builders.cpp - MLIR Declarative Linalg Builders --------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/IR/Builders.h"
 #include "mlir/Dialect/Affine/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Linalg/EDSC/Builders.h"
 #include "mlir/Dialect/Linalg/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Math/EDSC/Intrinsics.h"
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
 #include "mlir/Dialect/StandardOps/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Utils/StructuredOpsUtils.h"
 #include "mlir/IR/AffineExpr.h"
 
 using namespace mlir;
 using namespace mlir::edsc;
 using namespace mlir::edsc::intrinsics;
 using namespace mlir::linalg;

mlir/lib/Dialect/Linalg/Transforms/Loops.cpp

 //===- Loops.cpp - conversion from Linalg named and generic ops to loops --===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 #include "PassDetail.h"
-#include "mlir/Dialect/Affine/EDSC/Intrinsics.h"
-#include "mlir/Dialect/Linalg/EDSC/FoldedIntrinsics.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
 #include "mlir/Dialect/Linalg/IR/LinalgTypes.h"
 #include "mlir/Dialect/Linalg/Passes.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
-#include "mlir/Dialect/MemRef/EDSC/Intrinsics.h"
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
-#include "mlir/Dialect/StandardOps/EDSC/Intrinsics.h"
 #include "mlir/IR/AffineExpr.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/Support/LLVM.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "mlir/Transforms/FoldUtils.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 #include "llvm/ADT/TypeSwitch.h"
 
 using namespace mlir;
-using namespace mlir::edsc;
-using namespace mlir::edsc::intrinsics;
 using namespace mlir::linalg;
 
-using edsc::op::operator+;
+namespace {
+/// Helper struct to build simple arithmetic quantities with minimal type
+/// inference support.
+struct ArithBuilder {
+  ArithBuilder(OpBuilder &b, Location loc) : b(b), loc(loc) {}
+
+  Value select(Value cmp, Value lhs, Value rhs) {
+    return b.create<SelectOp>(loc, cmp, lhs, rhs);
+  }
+  Value slt(Value lhs, Value rhs) {
+    if (lhs.getType().isa<IntegerType>())
+      return b.create<CmpIOp>(loc, CmpIPredicate::slt, lhs, rhs);
+    return b.create<CmpFOp>(loc, CmpFPredicate::OLT, lhs, rhs);
+  }
+  Value sgt(Value lhs, Value rhs) {
+    if (lhs.getType().isa<IntegerType>())
+      return b.create<CmpIOp>(loc, CmpIPredicate::sgt, lhs, rhs);
+    return b.create<CmpFOp>(loc, CmpFPredicate::OGT, lhs, rhs);
+  }
+  Value add(Value lhs, Value rhs) {
+    if (lhs.getType().isa<IntegerType>())
+      return b.create<AddIOp>(loc, lhs, rhs);
+    return b.create<AddFOp>(loc, lhs, rhs);
+  }
+  Value mul(Value lhs, Value rhs) {
+    if (lhs.getType().isa<IntegerType>())
+      return b.create<MulIOp>(loc, lhs, rhs);
+    return b.create<MulFOp>(loc, lhs, rhs);
+  }
+
+  OpBuilder &b;
+  Location loc;
+};
+} // namespace
 
-static SmallVector<Value, 8> makeCanonicalAffineApplies(OpBuilder &b,
-                                                        Location loc,
-                                                        AffineMap map,
-                                                        ArrayRef<Value> vals) {
+static SmallVector<Value> makeCanonicalAffineApplies(OpBuilder &b, Location loc,
   if (map.isEmpty())
     return {};
 
   assert(map.getNumInputs() == vals.size());
-  SmallVector<Value, 8> res;
+  SmallVector<Value> res;
   res.reserve(map.getNumResults());
   auto dims = map.getNumDims();
   for (auto e : map.getResults()) {
     auto exprMap = AffineMap::get(dims, map.getNumSymbols(), e);
-    SmallVector<Value, 4> operands(vals.begin(), vals.end());
+    SmallVector<Value> operands(vals.begin(), vals.end());
     canonicalizeMapAndOperands(&exprMap, &operands);
-    res.push_back(affine_apply(exprMap, operands));
+    res.push_back(b.create<AffineApplyOp>(loc, exprMap, operands));
   }
   return res;
 }
 
-template <typename IndexedValueType, typename OpType>
-static void inlineRegionAndEmitStore(OpType op, ArrayRef<Value> indexedValues,
-                                     ArrayRef<SmallVector<Value, 8>> indexing,
+template <typename LoadOpTy, typename StoreOpTy, typename OpType>
+static void inlineRegionAndEmitStore(OpBuilder &b, Location loc, OpType op,
+                                     ArrayRef<Value> indexedValues,
+                                     ArrayRef<SmallVector<Value>> indexing,
                                      ArrayRef<Value> outputBuffers) {
-  assert(op->getNumRegions() == 1 && "Expected single region op");
-  auto &b = ScopedContext::getBuilderRef();
   auto &block = op->getRegion(0).front();
   BlockAndValueMapping map;
   map.map(block.getArguments(), indexedValues);
   for (auto &op : block.without_terminator()) {
     auto *newOp = b.clone(op, map);
     map.map(op.getResults(), newOp->getResults());
   }
 
-  Operation &terminator = block.back();
-  assert(isa<linalg::YieldOp>(terminator) &&
-         "expected a yield op in the end of the region");
-  for (unsigned i = 0, e = terminator.getNumOperands(); i < e; ++i) {
-    IndexedValueType O(outputBuffers[i]);
-    O(indexing[i]) = map.lookupOrDefault(terminator.getOperand(i));
+  Operation *terminator = block.getTerminator();
+  for (OpOperand &operand : terminator->getOpOperands()) {
+    Value toStore = map.lookupOrDefault(operand.get());
+    b.create<StoreOpTy>(loc, toStore, outputBuffers[operand.getOperandNumber()],
+                        indexing[operand.getOperandNumber()]);
   }
 }
 
 // Returns a pair that contains input indices and output indices of a
 // SingleInputPoolingOp `op`.
 struct InputAndOutputIndices {
-  SmallVector<Value, 8> inputs;
-  SmallVector<Value, 8> outputs;
+  SmallVector<Value> inputs;
+  SmallVector<Value> outputs;
 };
 template <typename SingleInputPoolingOp>
-static InputAndOutputIndices getInputAndOutputIndices(ArrayRef<Value> allIvs,
-                                                      SingleInputPoolingOp op) {
-  auto &b = ScopedContext::getBuilderRef();
-  auto loc = ScopedContext::getLocation();
+static InputAndOutputIndices
+getInputAndOutputIndices(OpBuilder &b, Location loc, ArrayRef<Value> allIvs,
   auto mapsRange = op.indexing_maps().template getAsRange<AffineMapAttr>();
   auto maps = llvm::to_vector<8>(
       llvm::map_range(mapsRange, [](AffineMapAttr a) { return a.getValue(); }));
   return InputAndOutputIndices{
       makeCanonicalAffineApplies(b, loc, maps[0], allIvs),
       makeCanonicalAffineApplies(b, loc, maps[2], allIvs)};
 }
 

 ///          store %14#0, %arg1[%i, %j, %k] :
 ///            memref<?x?x?Xf32, stride_specification>
 ///          store %14#1, %arg2[%i, %k, %j] :
 ///            memref<?x?x?Xf32, stride_specification>
 ///       }
 ///      }
 ///    }
 /// ```
-template <typename IndexedValueType>
-static void emitScalarImplementation(ArrayRef<Value> allIvs,
+template <typename LoadOpTy, typename StoreOpTy>
+static void emitScalarImplementation(OpBuilder &b, Location loc,
+                                     ArrayRef<Value> allIvs,
                                      LinalgOp linalgOp) {
   assert(linalgOp.hasBufferSemantics() &&
          "expected linalg op with buffer semantics");
-  auto &b = ScopedContext::getBuilderRef();
-  auto loc = ScopedContext::getLocation();
   unsigned nInputs = linalgOp.getNumInputs();
   unsigned nOutputs = linalgOp.getNumOutputs();
-  SmallVector<Value, 4> indexedValues;
+  SmallVector<Value> indexedValues;
   indexedValues.reserve(nInputs + nOutputs);
 
-  auto allIvsPlusDims = SmallVector<Value, 4>(allIvs.begin(), allIvs.end());
+  auto allIvsPlusDims = SmallVector<Value>(allIvs.begin(), allIvs.end());
 
   // TODO: Avoid the loads if the corresponding argument of the
   // region has no uses.
   // 1.a. Emit load from input views.
   for (unsigned i = 0; i < nInputs; ++i) {
     auto indexing = makeCanonicalAffineApplies(
         b, loc, linalgOp.getInputIndexingMap(i), allIvsPlusDims);
-    // Passing through IndexedValueType emits the proper load operation.
-    indexedValues.push_back(IndexedValueType(linalgOp.getInput(i))(indexing));
+    indexedValues.push_back(
+        b.create<LoadOpTy>(loc, linalgOp.getInput(i), indexing));
   }
   // 1.b. Emit load from output views.
   for (unsigned i = 0; i < nOutputs; ++i) {
     auto indexing = makeCanonicalAffineApplies(
         b, loc, linalgOp.getOutputIndexingMap(i), allIvsPlusDims);
-    // Passing through IndexedValueType emits the proper load operation.
     indexedValues.push_back(
-        IndexedValueType(linalgOp.getOutputBuffer(i))(indexing));
+        b.create<LoadOpTy>(loc, linalgOp.getOutputBuffer(i), indexing));
   }
 
   // TODO: When a region inliner exists, use it.
   // 2. Inline region, currently only works for a single basic block.
   // 3. Emit store.
-  SmallVector<SmallVector<Value, 8>, 8> indexing;
-  SmallVector<Value, 8> outputBuffers;
+  SmallVector<SmallVector<Value>, 8> indexing;
+  SmallVector<Value> outputBuffers;
   for (unsigned i = 0; i < nOutputs; ++i) {
     indexing.push_back(makeCanonicalAffineApplies(
         b, loc, linalgOp.getOutputIndexingMap(i), allIvsPlusDims));
     outputBuffers.push_back(linalgOp.getOutputBuffer(i));
   }
-  inlineRegionAndEmitStore<IndexedValueType>(linalgOp, indexedValues, indexing,
-                                             outputBuffers);
+  inlineRegionAndEmitStore<LoadOpTy, StoreOpTy>(b, loc, linalgOp, indexedValues,
+                                                indexing, outputBuffers);
 }
 
 // Create a padded view into the given `input` tensor using the 'indices'
 // to access the tensor. `skipPadding` lists the dimensions for which no padding
 // is needed e.g. the non-spatial dimensions for convolutions.
-template <typename IndexedValueType>
-Value getPaddedInput(Value input, ArrayRef<Value> indices,
-                     ArrayRef<int> skipPadding, Value padValue) {
-  // TODO: add a level of indirection to linalg.generic.
-
-  IndexedValueType indexedInput(input);
-
-  auto *context = ScopedContext::getContext();
-  Value zeroIndex = std_constant_index(0);
-  SmallVector<Value, 8> conds;
-  SmallVector<Value, 8> clampedImIdx;
+Value getPaddedInput(OpBuilder &b, Location loc, Value input,
+                     ArrayRef<Value> indices, ArrayRef<int> skipPadding,
+                     Value padValue) {
+  Value zeroIndex = b.create<ConstantIndexOp>(loc, 0);
+  SmallVector<Value> conds;
+  SmallVector<Value> clampedImIdx;
   for (auto iter : llvm::enumerate(indices)) {
     int idx = iter.index();
     auto dim = iter.value();
     if (is_contained(skipPadding, idx)) {
       clampedImIdx.push_back(dim);
       continue;
     }
 
-    using edsc::op::sge;
-    using edsc::op::slt;
-    using edsc::op::operator||;
-    Value leftOutOfBound = slt(dim, zeroIndex);
+    Value leftOutOfBound =
+        b.create<CmpIOp>(loc, CmpIPredicate::slt, dim, zeroIndex);
     if (conds.empty())
       conds.push_back(leftOutOfBound);
     else
-      conds.push_back(conds.back() || leftOutOfBound);
-    Value rightBound = memref_dim(input, idx);
-    conds.push_back(conds.back() || (sge(dim, rightBound)));
+      conds.push_back(b.create<OrOp>(loc, conds.back(), leftOutOfBound));
+    Value rightBound = b.create<memref::DimOp>(loc, input, idx);
+    Value rightOutOfBound =
+        b.create<CmpIOp>(loc, CmpIPredicate::sge, dim, rightBound);
+    conds.push_back(b.create<OrOp>(loc, conds.back(), rightOutOfBound));
 
     // When padding is involved, the indices will only be shifted to negative,
     // so having a max op is enough.
-    auto maxMap = AffineMap::get(/*dimCount=*/1, 0,
-                                 {getAffineDimExpr(/*position=*/0, context),
-                                  getAffineConstantExpr(0, context)},
-                                 context);
-    clampedImIdx.push_back(affine_max(dim.getType(), maxMap, ValueRange{dim}));
+    MLIRContext *ctx = input.getContext();
+    AffineExpr m = getAffineDimExpr(/*position=*/0, ctx),
+               zero = getAffineConstantExpr(0, ctx);
+    AffineMap maxMap =
+        AffineMap::inferFromExprList(ArrayRef<ArrayRef<AffineExpr>>{{m, zero}})
+            .front();
+    clampedImIdx.push_back(b.create<AffineMaxOp>(loc, maxMap, ValueRange{dim}));
   }
 
-  Value readInput = indexedInput(clampedImIdx);
-  return conds.empty() ? readInput
-                       : (Value)std_select(conds.back(), padValue, readInput);
+  Value readInput = b.create<memref::LoadOp>(loc, input, clampedImIdx);
+  if (conds.empty())
+    return readInput;
+
+  return b.create<SelectOp>(loc, conds.back(), padValue, readInput);
 }
 
 namespace {
 
 /// The padding value for a given Op depends on the semantics of the Op.
 /// The identity value for ConvOp and PoolingSumOp is 0, for PoolingMaxOp is
 /// -inf or minInt and for PoolingMinOp is inf or maxInt.
 template <typename OpType> Attribute getPadValueAttr(Type type) {
   llvm_unreachable("Unexpected op type for getPadValueAttr");
   return {};
 }
 
 template <> Attribute getPadValueAttr<PoolingMaxOp>(Type type) {
-  auto &b = ScopedContext::getBuilderRef();
   if (auto floatType = type.dyn_cast<FloatType>()) {
-    return b.getFloatAttr(
-        floatType,
-        APFloat::getInf(floatType.getFloatSemantics(), /*Negative*/ true));
+    return OpBuilder(type.getContext())
+        .getFloatAttr(floatType, APFloat::getInf(floatType.getFloatSemantics(),
+                                                 /*Negative*/ true));
   }
   if (auto intType = type.dyn_cast<IntegerType>()) {
     unsigned width = intType.getWidth();
     // The select instruction used to lower the PoolingMin uses a signed
     // comparison, use a signed constant irrespective of the signedness of the
     // integer type.
-    return b.getIntegerAttr(intType, APInt::getSignedMinValue(width));
+    return OpBuilder(type.getContext())
+        .getIntegerAttr(intType, APInt::getSignedMinValue(width));
   }
   llvm_unreachable("Unsupported data type for PoolingMaxOp");
   return {};
 }
 
 template <> Attribute getPadValueAttr<PoolingMinOp>(Type type) {
-  auto &b = ScopedContext::getBuilderRef();
   if (auto floatType = type.dyn_cast<FloatType>()) {
-    return b.getFloatAttr(floatType,
-                          APFloat::getInf(floatType.getFloatSemantics()));
+    return OpBuilder(type.getContext())
+        .getFloatAttr(floatType,
   }
   if (auto intType = type.dyn_cast<IntegerType>()) {
     unsigned width = intType.getWidth();
     // The select instruction used to lower the PoolingMin uses a signed
     // comparison, use a signed constant irrespective of the signedness of the
     // integer type.
-    return b.getIntegerAttr(intType, APInt::getSignedMaxValue(width));
+    return OpBuilder(type.getContext())
+        .getIntegerAttr(intType, APInt::getSignedMaxValue(width));
   }
   llvm_unreachable("Unsupported data type for PoolingMinOp");
   return {};
 }
 
 template <> Attribute getPadValueAttr<PoolingSumOp>(Type type) {
-  auto &b = ScopedContext::getBuilderRef();
-  return b.getZeroAttr(type);
+  return OpBuilder(type.getContext()).getZeroAttr(type);
 }
 
 template <> Attribute getPadValueAttr<ConvOp>(Type type) {
-  auto &b = ScopedContext::getBuilderRef();
-  return b.getZeroAttr(type);
+  return OpBuilder(type.getContext()).getZeroAttr(type);
 }
 
 } // namespace
 
 /// Returns true is `convOp` has a non-zero padding.
 static bool hasPadding(ConvOp convOp) {
   for (unsigned i = 0, e = convOp.getNumSpatialDimensions(); i < e; ++i) {
     if (convOp.getLowPad(i) > 0 || convOp.getHighPad(i) > 0)
       return true;
   }
   return false;
 }
 
-template <typename IndexedValueType>
-static void emitScalarImplementation(ArrayRef<Value> allIvs, ConvOp convOp) {
+template <typename LoadOpTy, typename StoreOpTy>
+static void emitScalarImplementation(OpBuilder &b, Location loc,
+                                     ArrayRef<Value> allIvs, ConvOp convOp) {
   assert(convOp.hasBufferSemantics() &&
          "expected linalg op with buffer semantics");
-  auto &b = ScopedContext::getBuilderRef();
-  auto loc = ScopedContext::getLocation();
   auto mapsRange = convOp.indexing_maps().getAsRange<AffineMapAttr>();
   auto maps = llvm::to_vector<8>(
       llvm::map_range(mapsRange, [](AffineMapAttr a) { return a.getValue(); }));
-  SmallVector<Value, 8> fIdx(
-      makeCanonicalAffineApplies(b, loc, maps[0], allIvs));
-  SmallVector<Value, 8> imIdx(
-      makeCanonicalAffineApplies(b, loc, maps[1], allIvs));
-  SmallVector<Value, 8> oIdx(
-      makeCanonicalAffineApplies(b, loc, maps[2], allIvs));
+  SmallVector<Value> fIdx(makeCanonicalAffineApplies(b, loc, maps[0], allIvs));
+  SmallVector<Value> imIdx(makeCanonicalAffineApplies(b, loc, maps[1], allIvs));
+  SmallVector<Value> oIdx(makeCanonicalAffineApplies(b, loc, maps[2], allIvs));
 
-  IndexedValueType F(convOp.filter()), O(convOp.output());
+  Value filter = convOp.filter(), output = convOp.output();
 
   // Emit scalar form. Padded conv involves an affine.max in the memory access
   // which is not allowed by affine.load. Override to use an MemRefIndexedValue
   // when there is non-zero padding.
   if (hasPadding(convOp)) {
     Type type = convOp.input().getType().cast<MemRefType>().getElementType();
-    Value padValue = std_constant(type, getPadValueAttr<ConvOp>(type));
-    Value paddedInput = getPaddedInput<MemRefIndexedValue>(
-        convOp.input(), imIdx,
+    Value padValue =
+        b.create<ConstantOp>(loc, type, getPadValueAttr<ConvOp>(type));
+    Value paddedInput =
+        getPaddedInput(b, loc, convOp.input(), imIdx,
                        /* Only need to pad the window dimensions */
                        {0, static_cast<int>(imIdx.size()) - 1}, padValue);
-    O(oIdx) += F(fIdx) * paddedInput;
+    Value filterVal = b.create<LoadOpTy>(loc, filter, fIdx);
+    Value mulVal = ArithBuilder(b, loc).mul(filterVal, paddedInput);
+    Value outputVal = b.create<LoadOpTy>(loc, output, oIdx);
+    Value addVal = ArithBuilder(b, loc).add(mulVal, outputVal);
+    b.create<StoreOpTy>(loc, addVal, output, oIdx);
   } else {
-    IndexedValueType I(convOp.input());
-    O(oIdx) += F(fIdx) * I(imIdx);
+    Value inputVal = b.create<LoadOpTy>(loc, convOp.input(), imIdx);
+    Value filterVal = b.create<LoadOpTy>(loc, filter, fIdx);
+    Value mulVal = ArithBuilder(b, loc).mul(filterVal, inputVal);
+    Value outputVal = b.create<LoadOpTy>(loc, output, oIdx);
+    Value addVal = ArithBuilder(b, loc).add(mulVal, outputVal);
+    b.create<StoreOpTy>(loc, addVal, output, oIdx);
   }
 }
 
 template <typename PoolingOp> static bool hasPadding(PoolingOp poolingOp) {
   for (unsigned i = 0, e = poolingOp.getNumWindowLoops(); i < e; ++i) {
     if (poolingOp.getLowPad(i) > 0 || poolingOp.getHighPad(i) > 0)
       return true;
   }
   return false;
 }
 
-template <typename IndexedValueType, typename PoolingOp>
-static Value getPoolingInput(PoolingOp op, ArrayRef<Value> inputIndices) {
+template <typename LoadOpTy, typename StoreOpTy, typename PoolingOp>
+static Value getPoolingInput(OpBuilder &b, Location loc, PoolingOp op,
+                             ArrayRef<Value> inputIndices) {
   if (hasPadding(op)) {
     Type type =
         op.input().getType().template cast<MemRefType>().getElementType();
-    Value padValue = std_constant(type, getPadValueAttr<PoolingOp>(type));
-    return getPaddedInput<MemRefIndexedValue>(op.input(), inputIndices,
-                                              /*Pad every dimension*/ {},
-                                              padValue);
-  }
-  IndexedValueType input(op.input());
-  return input(inputIndices);
-}
-
-template <typename IndexedValueType, typename OpType>
-void emitPoolingMinMaxScalarImplementation(ArrayRef<Value> allIvs, OpType op) {
-  InputAndOutputIndices indices = getInputAndOutputIndices(allIvs, op);
-  // Emit scalar form.
-  IndexedValueType output(op.output());
-  Value lhs = output(indices.outputs);
-  Value rhs = getPoolingInput<IndexedValueType>(op, indices.inputs);
-  using edsc::op::sgt;
-  using edsc::op::slt;
-  Value value = std::is_same<OpType, PoolingMinOp>()
-                    ? std_select(slt(lhs, rhs), lhs, rhs)
-                    : std_select(sgt(lhs, rhs), lhs, rhs);
-  output(indices.outputs) = value;
-}
-
-template <typename IndexedValueType>
-static void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingMaxOp op) {
-  emitPoolingMinMaxScalarImplementation<IndexedValueType, PoolingMaxOp>(allIvs,
-                                                                        op);
-}
-
-template <typename IndexedValueType>
-static void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingMinOp op) {
-  emitPoolingMinMaxScalarImplementation<IndexedValueType, PoolingMinOp>(allIvs,
-                                                                        op);
+    Value padValue =
+        b.create<ConstantOp>(loc, type, getPadValueAttr<PoolingOp>(type));
+    return getPaddedInput(b, loc, op.input(), inputIndices,
+                          /*Pad every dimension*/ {}, padValue);
+  }
+  return b.create<LoadOpTy>(loc, op.input(), inputIndices);
+}
+
+template <typename LoadOpTy, typename StoreOpTy, typename OpType>
+void emitPoolingMinMaxScalarImplementation(OpBuilder &b, Location loc,
+                                           ArrayRef<Value> allIvs, OpType op) {
+  InputAndOutputIndices indices = getInputAndOutputIndices(b, loc, allIvs, op);
+  Value lhs = b.create<LoadOpTy>(loc, op.output(), indices.outputs);
+  Value rhs = getPoolingInput<LoadOpTy, StoreOpTy>(b, loc, op, indices.inputs);
+  Value value = llvm::TypeSwitch<Operation *, Value>(op)
+                    .Case([&](PoolingMinOp poolingOp) {
+                      return ArithBuilder(b, loc).select(
+                          ArithBuilder(b, loc).slt(lhs, rhs), lhs, rhs);
+                    })
+                    .Case([&](PoolingMaxOp poolingOp) {
+                      return ArithBuilder(b, loc).select(
+                          ArithBuilder(b, loc).sgt(lhs, rhs), lhs, rhs);
+                    })
+                    .Default([&](auto) { return Value(); });
+  b.create<StoreOpTy>(loc, value, op.output(), indices.outputs);
 }
 
-template <typename IndexedValueType>
-static void emitScalarImplementation(ArrayRef<Value> allIvs, PoolingSumOp op) {
-  auto indices = getInputAndOutputIndices(allIvs, op);
-  IndexedValueType output(op.output());
-
-  // Emit scalar form.
-  output(indices.outputs) +=
-      getPoolingInput<IndexedValueType>(op, indices.inputs);
+template <typename LoadOpTy, typename StoreOpTy>
+static void emitScalarImplementation(OpBuilder &b, Location loc,
+                                     ArrayRef<Value> allIvs, PoolingMaxOp op) {
+  emitPoolingMinMaxScalarImplementation<LoadOpTy, StoreOpTy, PoolingMaxOp>(
+      b, loc, allIvs, op);
+}
+
+template <typename LoadOpTy, typename StoreOpTy>
+static void emitScalarImplementation(OpBuilder &b, Location loc,
+                                     ArrayRef<Value> allIvs, PoolingMinOp op) {
+  emitPoolingMinMaxScalarImplementation<LoadOpTy, StoreOpTy, PoolingMinOp>(
+      b, loc, allIvs, op);
+}
+
+template <typename LoadOpTy, typename StoreOpTy>
+static void emitScalarImplementation(OpBuilder &b, Location loc,
+                                     ArrayRef<Value> allIvs, PoolingSumOp op) {
+  auto indices = getInputAndOutputIndices(b, loc, allIvs, op);
+  Value inputVal =
+      getPoolingInput<LoadOpTy, StoreOpTy>(b, loc, op, indices.inputs);
+  Value outputVal = b.create<LoadOpTy>(loc, op.output(), indices.outputs);
+  Value added = ArithBuilder(b, loc).add(outputVal, inputVal);
+  b.create<StoreOpTy>(loc, added, op.output(), indices.outputs);
 }
 
 /// Replace the index operations in the body of the loop nest by the matching
 /// induction variables.
 static void replaceIndexOpsByInductionVariables(LinalgOp linalgOp,
                                                 PatternRewriter &rewriter,
                                                 ArrayRef<Operation *> loopOps) {
   // Extract the induction variables of the loop nest from outer to inner.
for (IndexOp indexOp :
          llvm::make_early_inc_range(loopOp.getLoopBody().getOps<IndexOp>()))
       rewriter.replaceOp(indexOp, allIvs[indexOp.dim()]);
   }
 }
 
 template <typename LoopTy>
 static Optional<LinalgLoops> linalgOpToLoopsImpl(PatternRewriter &rewriter,
                                                  LinalgOp linalgOp) {
-  using IndexedValueTy = typename GenerateLoopNest<LoopTy>::IndexedValueTy;
-  ScopedContext scope(rewriter, linalgOp.getLoc());
+  using LoadOpTy =
+      typename std::conditional<std::is_same<LoopTy, AffineForOp>::value,
+                                AffineLoadOp, memref::LoadOp>::type;
+  using StoreOpTy =
+      typename std::conditional<std::is_same<LoopTy, AffineForOp>::value,
+                                AffineStoreOp, memref::StoreOp>::type;
 
   // Canonicalize indexed_generic operations before lowering them to loops.
   if (isa<IndexedGenericOp>(linalgOp))
     return llvm::None;
 
   // The flattened loopToOperandRangesMaps is expected to be an invertible
   // permutation map (which is asserted in the inverse calculation).
   assert(linalgOp.hasBufferSemantics() &&
          "expected linalg op with buffer semantics");
 
   auto loopRanges = linalgOp.createLoopRanges(rewriter, linalgOp.getLoc());
   auto iteratorTypes = llvm::to_vector<4>(linalgOp.iterator_types().getValue());
 
-  SmallVector<Value, 4> allIvs;
+  SmallVector<Value> allIvs;
   GenerateLoopNest<LoopTy>::doit(
-      loopRanges, linalgOp, iteratorTypes,
-      [&](ValueRange ivs, ValueRange iterArgs) -> scf::ValueVector {
+      rewriter, linalgOp.getLoc(), loopRanges, linalgOp, iteratorTypes,
+      [&](OpBuilder &b, Location loc, ValueRange ivs,
+          ValueRange iterArgs) -> scf::ValueVector {
         assert(iterArgs.empty() && "unexpected iterArgs");
         allIvs.append(ivs.begin(), ivs.end());
         llvm::TypeSwitch<Operation *>(linalgOp)
             .Case<ConvOp, PoolingMaxOp, PoolingMinOp, PoolingSumOp, LinalgOp>(
                 [&](auto op) {
-                  emitScalarImplementation<IndexedValueTy>(allIvs, op);
+                  emitScalarImplementation<LoadOpTy, StoreOpTy>(b, loc, allIvs,
+                                                                op);
                 })
             .Default([&](Operation *op) { assert(false && "unexpected op"); });
         return scf::ValueVector{};
       });
   // Number of loop ops might be different from the number of ivs since some
   // loops like affine.parallel and scf.parallel have multiple ivs.
   SetVector<Operation *> loopSet;
   for (Value iv : allIvs) {
if (!llvm::all_of(tiledLoop.outputs(), [&](Value arg) {
       return failure();
 
     // TODO: Build loop nest with `scf.for` and `scf.parallel` depending on the
     // iterator type.
     scf::buildLoopNest(rewriter, loc, tiledLoop.lowerBound(),
                        tiledLoop.upperBound(), tiledLoop.step(),
                        [&](OpBuilder &builder, Location loc, ValueRange ivs) {
                          // Move body without its terminator.
-                         SmallVector<Value, 16> newBlockArgs;
+                         SmallVector<Value> newBlockArgs;
                          newBlockArgs.append(ivs.begin(), ivs.end());
                          newBlockArgs.append(tiledLoop.inputs().begin(),
                                              tiledLoop.inputs().end());
                          newBlockArgs.append(tiledLoop.outputs().begin(),
                                              tiledLoop.outputs().end());
                          Block *newBody = rewriter.getInsertionBlock();
                          rewriter.mergeBlocks(tiledLoop.getBody(), newBody,
                                               newBlockArgs);

mlir/lib/Dialect/Linalg/Transforms/Tiling.cpp

 #include "mlir/Dialect/Affine/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Linalg/EDSC/FoldedIntrinsics.h"
 #include "mlir/Dialect/Linalg/IR/LinalgTypes.h"
 #include "mlir/Dialect/Linalg/Passes.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 #include "mlir/Dialect/MemRef/EDSC/Intrinsics.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
 #include "mlir/Dialect/StandardOps/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/IR/AffineExpr.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/Transforms/FoldUtils.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 
 #include "llvm/Support/CommandLine.h"
if (!options.interchangeVector.empty()) {
     assert(invPermutationMap);
     applyPermutationToVector(loopRanges, interchangeVector);
     applyPermutationToVector(iteratorTypes, interchangeVector);
   }
 
   // 2. Create the tiled loops.
   LinalgOp res = op;
   SmallVector<Value, 4> ivs, tensorResults;
-  GenerateLoopNest<LoopTy>::doit(
-      loopRanges, op, iteratorTypes,
-      [&](ValueRange localIvs, ValueRange iterArgs) -> scf::ValueVector {
-        auto &b = ScopedContext::getBuilderRef();
-        auto loc = ScopedContext::getLocation();
-        ivs.assign(localIvs.begin(), localIvs.end());
+  auto tiledLoopBodyBuilder = [&](OpBuilder &b, Location loc,
+                                  ValueRange localIvs,
+                                  ValueRange iterArgs) -> scf::ValueVector {
 
-        // When an `interchangeVector` is present, it has been applied to the
-        // loop ranges and the iterator types. Apply its inverse to the
-        // resulting loop `ivs` to match the op definition.
-        SmallVector<Value, 4> interchangedIvs;
-        if (!options.interchangeVector.empty())
-          interchangedIvs = applyMapToValues(b, loc, invPermutationMap, ivs);
-        else
-          interchangedIvs.assign(ivs.begin(), ivs.end());
 
-        assert(op.getNumOutputTensors() == iterArgs.size() &&
-               "num output tensors must match number of loop iter arguments");
 
-        auto operands = llvm::to_vector<4>(op.getInputs());
-        SmallVector<Value, 4> outputBuffers = op.getOutputBuffers();
-        // TODO: thanks to simplifying assumption we do not need to worry about
-        // order of output buffers and tensors: there is only ever one kind.
-        assert(outputBuffers.empty() || iterArgs.empty());
-        operands.append(outputBuffers.begin(), outputBuffers.end());
-        operands.append(iterArgs.begin(), iterArgs.end());
-        auto sizeBounds =
-            applyMapToValues(b, loc, shapeSizesToLoopsMap, allShapeSizes);
-        SmallVector<Value, 4> tiledOperands = makeTiledShapes(
-            b, loc, op, operands, interchangedIvs, tileSizes, sizeBounds);
-        auto nonShapedOperands = op.getAssumedNonShapedOperands();
-        tiledOperands.append(nonShapedOperands.begin(),
-                             nonShapedOperands.end());
+    tiledOperands.append(nonShapedOperands.begin(), nonShapedOperands.end());
 
-        // TODO: use an interface/adaptor to avoid leaking position in
-        // `tiledOperands`.
-        SmallVector<Type, 4> resultTensorTypes;
-        for (OpOperand *opOperand : op.getOutputTensorsOpOperands())
-          resultTensorTypes.push_back(
-              tiledOperands[opOperand->getOperandNumber()].getType());
 
-        res = op.clone(b, loc, resultTensorTypes, tiledOperands);
 
-        // Insert a subtensor_insert for each output tensor.
-        unsigned resultIdx = 0;
-        for (OpOperand *opOperand : op.getOutputTensorsOpOperands()) {
-          // TODO: use an interface/adaptor to avoid leaking position in
-          // `tiledOperands`.
-          Value outputTensor = tiledOperands[opOperand->getOperandNumber()];
-          if (auto subtensor = outputTensor.getDefiningOp<SubTensorOp>()) {
-            tensorResults.push_back(b.create<SubTensorInsertOp>(
-                loc, subtensor.source().getType(), res->getResult(resultIdx),
-                subtensor.source(), subtensor.offsets(), subtensor.sizes(),
-                subtensor.strides(), subtensor.static_offsets(),
-                subtensor.static_sizes(), subtensor.static_strides()));
-          } else {
-            tensorResults.push_back(res->getResult(resultIdx));
-          }
-          ++resultIdx;
-        }
-        return scf::ValueVector(tensorResults.begin(), tensorResults.end());
-      },
-      options.distribution);
+  };
+  GenerateLoopNest<LoopTy>::doit(b, op.getLoc(), loopRanges, op, iteratorTypes,
+                                 tiledLoopBodyBuilder, options.distribution);
 
   // 3. Transform IndexOp results w.r.t. the tiling.
   transformIndexOps(b, res, ivs, loopIndexToRangeIndex);
 
   // 4. Gather the newly created loops and return them with the new op.
   SmallVector<Operation *, 8> loops;
   loops.reserve(ivs.size());
   for (auto iv : ivs) {

mlir/lib/Dialect/Linalg/Utils/Utils.cpp

 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 
 #include "mlir/Dialect/Affine/EDSC/Intrinsics.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
 #include "mlir/Dialect/Linalg/IR/LinalgTypes.h"
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/StandardOps/EDSC/Intrinsics.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/IR/AffineExpr.h"
 #include "mlir/IR/AffineExprVisitor.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/IR/Matchers.h"
 #include "mlir/IR/OpImplementation.h"
IntegerAttr getSmallestBoundingIndex(Value size) {
   if (boundingConst && *boundingConst >= 0)
     return Builder(size.getContext()).getIndexAttr(*boundingConst);
   return nullptr;
 }
 
 /// Specialization to build an scf "for" nest.
 template <>
 void GenerateLoopNest<scf::ForOp>::doit(
-    ArrayRef<Range> loopRanges, LinalgOp linalgOp,
+    OpBuilder &b, Location loc, ArrayRef<Range> loopRanges, LinalgOp linalgOp,
     ArrayRef<Attribute> iteratorTypes,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> bodyBuilderFn,
+    function_ref<scf::ValueVector(OpBuilder &, Location, ValueRange,
+                                  ValueRange)>
+        bodyBuilderFn,
     Optional<LinalgLoopDistributionOptions> distributionOptions) {
   auto iterArgInitValues = linalgOp.getOutputTensors();
   // Create procInfo so it dominates loops, if appropriate.
-  OpBuilder &builder = edsc::ScopedContext::getBuilderRef();
-  Location loc = edsc::ScopedContext::getLocation();
-
   SmallVector<ProcInfo, 4> procInfo;
   SmallVector<DistributionMethod, 0> distributionMethod;
   if (distributionOptions.hasValue()) {
     // Collect loop ranges for parallel dimensions.
     SmallVector<Range, 2> parallelLoopRanges;
     for (auto iteratorType : enumerate(iteratorTypes))
       if (isParallelIteratorType(iteratorType.value()))
         parallelLoopRanges.push_back(loopRanges[iteratorType.index()]);
 
     // Get their distribution schemes.
     distributionMethod = distributionOptions->distributionMethod;
     if (distributionMethod.size() < parallelLoopRanges.size())
       parallelLoopRanges.resize(distributionMethod.size());
-    procInfo = distributionOptions->procInfo(builder, loc, parallelLoopRanges);
+    procInfo = distributionOptions->procInfo(b, loc, parallelLoopRanges);
   }
 
   SmallVector<Value, 4> lbs, ubs, steps;
   unpackRanges(loopRanges, lbs, ubs, steps);
-  LoopNest loopNest =
-      edsc::loopNestBuilder(lbs, ubs, steps, iterArgInitValues, bodyBuilderFn);
+  LoopNest loopNest = mlir::scf::buildLoopNest(
+      b, loc, lbs, ubs, steps, iterArgInitValues, bodyBuilderFn);
 
   if (!distributionOptions || loopNest.loops.empty())
     return;
 
   // Filter out scf.for loops that were created out of parallel dimensions.
   SmallVector<scf::ForOp, 4> loops;
   for (auto iteratorType : enumerate(iteratorTypes))
     if (isParallelIteratorType(iteratorType.value()))
       loops.push_back(loopNest.loops[iteratorType.index()]);
 
   // Distribute - only supports cyclic distribution for now.
   for (auto it : llvm::zip(loops, procInfo, distributionMethod))
     if (std::get<2>(it) == DistributionMethod::Cyclic)
       mapLoopToProcessorIds(std::get<0>(it), std::get<1>(it).procId,
                             std::get<1>(it).nprocs);
 }
 
 /// Specialization to build affine "for" nest.
 template <>
 void GenerateLoopNest<AffineForOp>::doit(
-    ArrayRef<Range> loopRanges, LinalgOp linalgOp,
+    OpBuilder &b, Location loc, ArrayRef<Range> loopRanges, LinalgOp linalgOp,
     ArrayRef<Attribute> iteratorTypes,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> bodyBuilderFn,
+    function_ref<scf::ValueVector(OpBuilder &, Location, ValueRange,
+                                  ValueRange)>
+        bodyBuilderFn,
     Optional<LinalgLoopDistributionOptions>) {
   auto iterArgInitValues = linalgOp.getOutputTensors();
   assert(iterArgInitValues.empty() && "unexpected AffineForOp init values");
   SmallVector<Value, 4> lbs, ubs, steps;
   unpackRanges(loopRanges, lbs, ubs, steps);
 
   // Affine loops require constant steps.
   SmallVector<int64_t, 4> constantSteps;
   constantSteps.reserve(steps.size());
   for (Value v : steps) {
     auto op = v.getDefiningOp<ConstantIndexOp>();
     assert(op && "Affine loops require constant steps");
     constantSteps.push_back(op.getValue());
   }
 
-  auto bodyBuilderWithoutIterArgsFn = [&](ValueRange ivs) {
-    bodyBuilderFn(ivs, {});
-  };
-  edsc::affineLoopNestBuilder(lbs, ubs, constantSteps,
-                              bodyBuilderWithoutIterArgsFn);
+  mlir::buildAffineLoopNest(b, loc, lbs, ubs, constantSteps,
+                            [&](OpBuilder &b, Location loc, ValueRange ivs) {
+                              bodyBuilderFn(b, loc, ivs, {});
+                            });
 }
 
 /// Specialization to build an linalg.tiled_loop
 template <>
 void GenerateLoopNest<TiledLoopOp>::doit(
-    ArrayRef<Range> loopRanges, LinalgOp linalgOp,
+    OpBuilder &b, Location loc, ArrayRef<Range> loopRanges, LinalgOp linalgOp,
     ArrayRef<Attribute> iteratorTypes,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> bodyBuilderFn,
+    function_ref<scf::ValueVector(OpBuilder &, Location, ValueRange,
+                                  ValueRange)>
+        bodyBuilderFn,
     Optional<LinalgLoopDistributionOptions>) {
-  OpBuilder &builder = edsc::ScopedContext::getBuilderRef();
-  Location loc = edsc::ScopedContext::getLocation();
   SmallVector<ProcInfo, 2> procInfo;
-
   SmallVector<Value, 4> lbs, ubs, steps;
   unpackRanges(loopRanges, lbs, ubs, steps);
 
   auto wrappedBuilderFn = [&](OpBuilder &nestedBuilder, Location nestedLoc,
                               ValueRange ivs, ValueRange inputs,
                               ValueRange outputs) {
-    ScopedContext context(nestedBuilder, nestedLoc);
-    scf::ValueVector results = bodyBuilderFn(ivs, linalgOp.getOutputTensors());
+    scf::ValueVector results = bodyBuilderFn(nestedBuilder, nestedLoc, ivs,
+                                             linalgOp.getOutputTensors());
     nestedBuilder.create<linalg::YieldOp>(nestedLoc, results);
   };
 
-  auto tiledLoop = builder.create<TiledLoopOp>(
+  auto tiledLoop = b.create<TiledLoopOp>(
       loc, lbs, ubs, steps, linalgOp.getInputs(), linalgOp.getOutputs(),
-      builder.getArrayAttr(iteratorTypes), wrappedBuilderFn);
+      b.getArrayAttr(iteratorTypes), wrappedBuilderFn);
 
   // Replace inputs/outputs with the corresponding region args.
   auto isInsideTiledLoop = [&](OpOperand &operand) {
     return operand.getOwner()->getBlock() == tiledLoop.getBody();
   };
   for (auto it :
        llvm::zip(linalgOp.getInputs(), tiledLoop.getRegionInputArgs()))
     std::get<0>(it).replaceUsesWithIf(std::get<1>(it), isInsideTiledLoop);
   for (auto it :
        llvm::zip(linalgOp.getOutputs(), tiledLoop.getRegionOutputArgs()))
     std::get<0>(it).replaceUsesWithIf(std::get<1>(it), isInsideTiledLoop);
 }
 
 /// Update the `lb`, `ub` and `step` to get per processor `lb`, `ub` and `step`.
-void updateBoundsForCyclicDistribution(OpBuilder &builder, Location loc,
-                                       Value procId, Value nprocs, Value &lb,
-                                       Value &ub, Value &step) {
+void updateBoundsForCyclicDistribution(OpBuilder &b, Location loc, Value procId,
+                                       Value nprocs, Value &lb, Value &ub,
+                                       Value &step) {
   using edsc::op::operator+;
   using edsc::op::operator*;
   lb = lb + (procId * step);
   step = nprocs * step;
 }
 
 /// Generates a loop nest consisting of scf.parallel and scf.for, depending
 /// on the `iteratorTypes.` Consecutive parallel loops create a single
 /// scf.parallel operation; each sequential loop creates a new scf.for
 /// operation. The body of the innermost loop is populated by
 /// `bodyBuilderFn` that accepts a range of induction variables for all
 /// loops. `ivStorage` is used to store the partial list of induction
 /// variables.
 // TODO: this function can be made iterative instead. However, it
 // will have at most as many recursive calls as nested loops, which rarely
 // exceeds 10.
-static void
-generateParallelLoopNest(ValueRange lbs, ValueRange ubs, ValueRange steps,
-                         ArrayRef<Attribute> iteratorTypes,
-                         function_ref<void(ValueRange)> bodyBuilderFn,
-                         SmallVectorImpl<Value> &ivStorage,
-                         ArrayRef<DistributionMethod> distributionMethod = {}) {
+static void generateParallelLoopNest(
+    OpBuilder &b, Location loc, ValueRange lbs, ValueRange ubs,
+    ValueRange steps, ArrayRef<Attribute> iteratorTypes,
+    function_ref<void(OpBuilder &, Location, ValueRange)> bodyBuilderFn,
   assert(lbs.size() == ubs.size());
   assert(lbs.size() == steps.size());
   assert(lbs.size() == iteratorTypes.size());
 
   // If there are no (more) loops to be generated, generate the body and be
   // done with it.
-  if (iteratorTypes.empty())
-    return bodyBuilderFn(ivStorage);
+  if (iteratorTypes.empty()) {
+    bodyBuilderFn(b, loc, ivStorage);
+    return;
+  }
 
   // Find the outermost parallel loops and drop their types from the list.
   unsigned nLoops = iteratorTypes.size();
   unsigned nOuterPar =
       nLoops - iteratorTypes.drop_while(isParallelIteratorType).size();
 
   // If there are no outer parallel loops, generate one sequential loop and
   // recurse. Note that we wouldn't have dropped anything from `iteratorTypes`
   // in this case.
   if (nOuterPar == 0) {
-    edsc::loopNestBuilder(lbs[0], ubs[0], steps[0], [&](Value iv) {
-      ivStorage.push_back(iv);
-      generateParallelLoopNest(lbs.drop_front(), ubs.drop_front(),
-                               steps.drop_front(), iteratorTypes.drop_front(),
-                               bodyBuilderFn, ivStorage, distributionMethod);
+    LoopNest singleLoop = buildLoopNest(
+        b, loc, lbs.take_front(), ubs.take_front(), steps.take_front(),
+        [&](OpBuilder &b, Location loc, ValueRange ivs) {
+          ivStorage.append(ivs.begin(), ivs.end());
+          generateParallelLoopNest(b, loc, lbs.drop_front(), ubs.drop_front(),
+                                   steps.drop_front(),
+                                   iteratorTypes.drop_front(), bodyBuilderFn,
+                                   ivStorage, distributionMethod);
         });
     return;
   }
   if (distributionMethod.empty()) {
     // Generate a single parallel loop-nest operation for all outermost
     // parallel loops and recurse.
-    edsc::OperationBuilder<scf::ParallelOp>(
-        lbs.take_front(nOuterPar), ubs.take_front(nOuterPar),
+    b.create<scf::ParallelOp>(
+        loc, lbs.take_front(nOuterPar), ubs.take_front(nOuterPar),
         steps.take_front(nOuterPar),
         [&](OpBuilder &nestedBuilder, Location nestedLoc, ValueRange localIvs) {
-          edsc::ScopedContext context(nestedBuilder, nestedLoc);
           ivStorage.append(localIvs.begin(), localIvs.end());
           generateParallelLoopNest(
-              lbs.drop_front(nOuterPar), ubs.drop_front(nOuterPar),
-              steps.drop_front(nOuterPar), iteratorTypes.drop_front(nOuterPar),
-              bodyBuilderFn, ivStorage,
+              nestedBuilder, nestedLoc, lbs.drop_front(nOuterPar),
+              ubs.drop_front(nOuterPar), steps.drop_front(nOuterPar),
+              iteratorTypes.drop_front(nOuterPar), bodyBuilderFn, ivStorage,
               (distributionMethod.size() < nOuterPar)
                   ? ArrayRef<DistributionMethod>()
                   : distributionMethod.drop_front(nOuterPar));
         });
     return;
   }
 
   // Process all consecutive similarly distributed loops simultaneously.
   DistributionMethod methodToUse = distributionMethod[0];
   unsigned numProcessed = 1;
   for (unsigned i = 1; i < nOuterPar && i < distributionMethod.size(); ++i) {
     if (distributionMethod[i] != methodToUse)
       break;
     numProcessed++;
   }
 
   switch (methodToUse) {
   case DistributionMethod::Cyclic: {
     // Generate a single parallel loop-nest operation for all outermost
     // parallel loops and recurse.
-    edsc::OperationBuilder<scf::ParallelOp>(
-        lbs.take_front(numProcessed), ubs.take_front(numProcessed),
+    b.create<scf::ParallelOp>(
+        loc, lbs.take_front(numProcessed), ubs.take_front(numProcessed),
         steps.take_front(numProcessed),
         [&](OpBuilder &nestedBuilder, Location nestedLoc, ValueRange localIvs) {
-          edsc::ScopedContext context(nestedBuilder, nestedLoc);
           ivStorage.append(localIvs.begin(), localIvs.end());
           generateParallelLoopNest(
-              lbs.drop_front(numProcessed), ubs.drop_front(numProcessed),
-              steps.drop_front(numProcessed),
+              nestedBuilder, nestedLoc, lbs.drop_front(numProcessed),
+              ubs.drop_front(numProcessed), steps.drop_front(numProcessed),
               iteratorTypes.drop_front(numProcessed), bodyBuilderFn, ivStorage,
               (distributionMethod.size() < numProcessed)
                   ? ArrayRef<DistributionMethod>()
                   : distributionMethod.drop_front(numProcessed));
         });
     return;
   }
   case DistributionMethod::CyclicNumProcsGeNumIters: {
     // Check (for the processed loops) that the iteration is in-bounds.
     using edsc::op::slt;
     using edsc::op::operator&&;
     Value cond = slt(lbs[0], ubs[0]);
     for (unsigned i = 1; i < numProcessed; ++i)
       cond = cond && slt(lbs[i], ubs[i]);
     ivStorage.append(lbs.begin(), std::next(lbs.begin(), numProcessed));
-    edsc::conditionBuilder(cond, [&]() {
+    b.create<scf::IfOp>(loc, cond, [&](OpBuilder &b, Location loc) {
       generateParallelLoopNest(
-          lbs.drop_front(numProcessed), ubs.drop_front(numProcessed),
+          b, loc, lbs.drop_front(numProcessed), ubs.drop_front(numProcessed),
           steps.drop_front(numProcessed),
           iteratorTypes.drop_front(numProcessed), bodyBuilderFn, ivStorage,
           distributionMethod.drop_front(numProcessed));
+      b.create<scf::YieldOp>(loc, ValueRange{});
     });
     return;
   }
   case DistributionMethod::CyclicNumProcsEqNumIters:
     // No check/loops needed here. Set the `%iv` to be the `%lb` and proceed
     // with inner loop generation.
     ivStorage.append(lbs.begin(), std::next(lbs.begin(), numProcessed));
     generateParallelLoopNest(
-        lbs.drop_front(numProcessed), ubs.drop_front(numProcessed),
+        b, loc, lbs.drop_front(numProcessed), ubs.drop_front(numProcessed),
         steps.drop_front(numProcessed), iteratorTypes.drop_front(numProcessed),
         bodyBuilderFn, ivStorage, distributionMethod.drop_front(numProcessed));
     return;
   }
 }
 
 /// Specialization for generating a mix of parallel and sequential scf loops.
 template <>
 void GenerateLoopNest<scf::ParallelOp>::doit(
-    ArrayRef<Range> loopRanges, LinalgOp linalgOp,
+    OpBuilder &b, Location loc, ArrayRef<Range> loopRanges, LinalgOp linalgOp,
     ArrayRef<Attribute> iteratorTypes,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> bodyBuilderFn,
+    function_ref<scf::ValueVector(OpBuilder &, Location, ValueRange,
+                                  ValueRange)>
+        bodyBuilderFn,
     Optional<LinalgLoopDistributionOptions> distributionOptions) {
   auto iterArgInitValues = linalgOp.getOutputTensors();
   assert(iterArgInitValues.empty() && "unexpected ParallelOp init values");
   // This function may be passed more iterator types than ranges.
   assert(iteratorTypes.size() >= loopRanges.size() &&
          "expected iterator type for all ranges");
   iteratorTypes = iteratorTypes.take_front(loopRanges.size());
   SmallVector<Value, 8> lbsStorage, ubsStorage, stepsStorage, ivs;
   unsigned numLoops = iteratorTypes.size();
   ivs.reserve(numLoops);
   lbsStorage.reserve(numLoops);
   ubsStorage.reserve(numLoops);
   stepsStorage.reserve(numLoops);
 
   // Get the loop lb, ub, and step.
   unpackRanges(loopRanges, lbsStorage, ubsStorage, stepsStorage);
 
   // Modify the lb, ub, and step based on the distribution options.
   SmallVector<DistributionMethod, 0> distributionMethod;
   if (distributionOptions) {
     auto &options = distributionOptions.getValue();
-    OpBuilder &builder = edsc::ScopedContext::getBuilderRef();
+    OpBuilder &b = edsc::ScopedContext::getBuilderRef();
     Location loc = edsc::ScopedContext::getLocation();
     distributionMethod.assign(distributionOptions->distributionMethod.begin(),
                               distributionOptions->distributionMethod.end());
     SmallVector<Range, 2> parallelLoopRanges;
     for (auto iteratorType : enumerate(iteratorTypes)) {
       if (isParallelIteratorType(iteratorType.value()))
         parallelLoopRanges.push_back(loopRanges[iteratorType.index()]);
     }
     if (distributionMethod.size() < parallelLoopRanges.size())
       parallelLoopRanges.resize(distributionMethod.size());
     SmallVector<ProcInfo, 2> procInfo =
-        options.procInfo(builder, loc, parallelLoopRanges);
+        options.procInfo(b, loc, parallelLoopRanges);
     unsigned index = 0;
     for (auto iteratorType : enumerate(iteratorTypes)) {
       if (index >= procInfo.size())
         break;
       if (isParallelIteratorType(iteratorType.value())) {
         unsigned i = iteratorType.index();
-        updateBoundsForCyclicDistribution(builder, loc, procInfo[index].procId,
+        updateBoundsForCyclicDistribution(b, loc, procInfo[index].procId,
                                           procInfo[index].nprocs, lbsStorage[i],
                                           ubsStorage[i], stepsStorage[i]);
         index++;
       }
     }
   }
   ValueRange lbs(lbsStorage), ubs(ubsStorage), steps(stepsStorage);
-  auto bodyBuilderWithoutIterArgsFn = [&](ValueRange ivs) {
-    bodyBuilderFn(ivs, {});
-  };
-  generateParallelLoopNest(lbs, ubs, steps, iteratorTypes,
-                           bodyBuilderWithoutIterArgsFn, ivs,
-                           distributionMethod);
+  generateParallelLoopNest(
+      b, loc, lbs, ubs, steps, iteratorTypes,
+      [&](OpBuilder &b, Location loc, ValueRange ivs) {
+        bodyBuilderFn(b, loc, ivs, {});
+      },
+      ivs, distributionMethod);
 
   assert(ivs.size() == iteratorTypes.size() && "did not generate enough loops");
 }
 
-SmallVector<Value, 4> makeTiledShapes(OpBuilder &builder, Location loc,
+SmallVector<Value, 4> makeTiledShapes(OpBuilder &b, Location loc,
                                       LinalgOp linalgOp,
                                       ArrayRef<Value> tiledOperands,
                                       ValueRange ivs, ValueRange tileSizes,
                                       ArrayRef<Value> sizeBounds) {
   assert(ivs.size() == static_cast<size_t>(llvm::count_if(
                            llvm::make_range(tileSizes.begin(), tileSizes.end()),
                            [](Value v) { return !isZero(v); })) &&
          "expected as many ivs as non-zero sizes");
for (unsigned idx = 0, idxIvs = 0, e = tileSizes.size(); idx < e; ++idx) {
     lbs.push_back(isTiled ? ivs[idxIvs++] : (Value)std_constant_index(0));
     // Before composing, we need to make range a closed interval.
     Value size = isTiled ? tileSizes[idx] : sizeBounds[idx];
     subShapeSizes.push_back(size - std_constant_index(1));
     LLVM_DEBUG(llvm::dbgs() << "lb: " << lbs.back() << "\n");
     LLVM_DEBUG(llvm::dbgs() << "size: " << subShapeSizes.back() << "\n");
   }
 
-  MLIRContext *context = builder.getContext();
+  MLIRContext *context = b.getContext();
   SmallVector<Value, 4> tiledShapes;
   tiledShapes.reserve(tiledOperands.size());
   for (auto en : llvm::enumerate(tiledOperands)) {
     Value shapedOp = en.value();
     LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: for operand " << shapedOp);
     ShapedType shapedType = shapedOp.getType().cast<ShapedType>();
     unsigned rank = shapedType.getRank();
     AffineMap map = linalgOp.getIndexingMap(en.index());
for (auto en : llvm::enumerate(tiledOperands)) {
     // Construct a new subview / subtensor for the tile.
     SmallVector<OpFoldResult, 4> offsets, sizes, strides;
     offsets.reserve(rank);
     sizes.reserve(rank);
     strides.reserve(rank);
     for (unsigned r = 0; r < rank; ++r) {
       LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: for dim#" << r);
       if (!isTiled(map.getSubMap({r}), tileSizes)) {
-        offsets.push_back(builder.getIndexAttr(0));
+        offsets.push_back(b.getIndexAttr(0));
         Value dim = memref_dim(shapedOp, r).value;
         sizes.push_back(dim);
-        strides.push_back(builder.getIndexAttr(1));
+        strides.push_back(b.getIndexAttr(1));
         LLVM_DEBUG(llvm::dbgs() << ": not tiled: use size: " << dim << "\n");
         continue;
       }
       LLVM_DEBUG(llvm::dbgs() << ": tiled: figure out subsize...\n");
 
       // Tiling creates a new slice at the proper index, the slice step is 1
       // (i.e. the op does not subsample, stepping occurs in the loop).
       auto m = map.getSubMap({r});
       LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: submap: " << map << "\n");
-      auto offset = applyMapToValues(builder, loc, m, lbs).front();
+      auto offset = applyMapToValues(b, loc, m, lbs).front();
       offsets.push_back(offset);
-      auto closedIntSize =
-          applyMapToValues(builder, loc, m, subShapeSizes).front();
+      auto closedIntSize = applyMapToValues(b, loc, m, subShapeSizes).front();
       // Resulting size needs to be made half open interval again.
       auto size = closedIntSize + std_constant_index(1);
       LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: raw size: " << size << "\n");
 
       // The size of the subview / subtensor should be trimmed to avoid
       // out-of-bounds accesses, unless we statically know the subshape size
       // divides the shape size evenly.
       int64_t shapeSize = shapedType.getDimSize(r);
       auto sizeCst = size.getDefiningOp<ConstantIndexOp>();
       if (ShapedType::isDynamic(shapeSize) || !sizeCst ||
           (shapeSize % sizeCst.getValue()) != 0) {
         LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: shapeSize=" << shapeSize
                                 << ", size: " << size
                                 << ": make sure in bound with affine.min\n");
         AffineExpr dim0, dim1, dim2;
         bindDims(context, dim0, dim1, dim2);
         // Compute min(size, dim - offset) to avoid out-of-bounds accesses.
-        auto minMap = AffineMap::get(
-            /*dimCount=*/3, /*symbolCount=*/0, {dim0, dim1 - dim2}, context);
-        Value d = memref_dim(shapedOp, r);
+        AffineMap minMap =
+            AffineMap::inferFromExprList(
+                ArrayRef<ArrayRef<AffineExpr>>{{dim0, dim1 - dim2}})
+                .front();
+        Value d = b.create<memref::DimOp>(loc, shapedOp, r);
         SmallVector<Value, 4> operands{size, d, offset};
         fullyComposeAffineMapAndOperands(&minMap, &operands);
-        size = affine_min(builder.getIndexType(), minMap, operands);
+        size = b.create<AffineMinOp>(loc, b.getIndexType(), minMap, operands);
       }
 
       sizes.push_back(size);
       LLVM_DEBUG(llvm::dbgs()
                  << "makeTiledShapes: new offset: " << offset << "\n");
       LLVM_DEBUG(llvm::dbgs() << "makeTiledShapes: new size: " << size << "\n");
-      strides.push_back(builder.getIndexAttr(1));
+      strides.push_back(b.getIndexAttr(1));
     }
 
     if (shapedType.isa<MemRefType>())
-      tiledShapes.push_back(builder.create<memref::SubViewOp>(
-          loc, shapedOp, offsets, sizes, strides));
+      tiledShapes.push_back(
+          b.create<memref::SubViewOp>(loc, shapedOp, offsets, sizes, strides));
     else
       tiledShapes.push_back(
-          builder.create<SubTensorOp>(loc, shapedOp, offsets, sizes, strides));
+          b.create<SubTensorOp>(loc, shapedOp, offsets, sizes, strides));
   }
 
   return tiledShapes;
 }
 
 } // namespace linalg
 } // namespace mlir

mlir/lib/Dialect/SCF/CMakeLists.txt

 add_mlir_dialect_library(MLIRSCF
   SCF.cpp
-  EDSC/Builders.cpp
 
   ADDITIONAL_HEADER_DIRS
   ${MLIR_MAIN_INCLUDE_DIR}/mlir/LoopOps
 
   DEPENDS
   MLIRSCFOpsIncGen
 
   LINK_LIBS PUBLIC
-  MLIREDSC
   MLIRIR
   MLIRLoopLikeInterface
   MLIRMemRef
   MLIRSideEffectInterfaces
   MLIRStandard
   )
 
 add_subdirectory(Transforms)

mlir/lib/Dialect/SCF/EDSC/Builders.cpp

-//===- Builders.cpp - MLIR Declarative Builder Classes --------------------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#include "mlir/Dialect/SCF/EDSC/Builders.h"
-#include "mlir/Dialect/SCF/SCF.h"
-#include "mlir/IR/AffineExpr.h"
-#include "mlir/IR/AffineMap.h"
-
-using namespace mlir;
-using namespace mlir::edsc;
-
-mlir::scf::LoopNest
-mlir::edsc::loopNestBuilder(ValueRange lbs, ValueRange ubs, ValueRange steps,
-                            function_ref<void(ValueRange)> fun) {
-  // Delegates actual construction to scf::buildLoopNest by wrapping `fun` into
-  // the expected function interface.
-  assert(ScopedContext::getContext() && "EDSC ScopedContext not set up");
-  return mlir::scf::buildLoopNest(
-      ScopedContext::getBuilderRef(), ScopedContext::getLocation(), lbs, ubs,
-      steps, [&](OpBuilder &builder, Location loc, ValueRange ivs) {
-        ScopedContext context(builder, loc);
-        if (fun)
-          fun(ivs);
-      });
-}
-
-mlir::scf::LoopNest
-mlir::edsc::loopNestBuilder(Value lb, Value ub, Value step,
-                            function_ref<void(Value)> fun) {
-  // Delegates to the ValueRange-based version by wrapping the lambda.
-  auto wrapper = [&](ValueRange ivs) {
-    assert(ivs.size() == 1);
-    if (fun)
-      fun(ivs[0]);
-  };
-  return loopNestBuilder(ValueRange(lb), ValueRange(ub), ValueRange(step),
-                         wrapper);
-}
-
-mlir::scf::LoopNest mlir::edsc::loopNestBuilder(
-    Value lb, Value ub, Value step, ValueRange iterArgInitValues,
-    function_ref<scf::ValueVector(Value, ValueRange)> fun) {
-  // Delegates actual construction to scf::buildLoopNest by wrapping `fun` into
-  // the expected function interface.
-  assert(ScopedContext::getContext() && "EDSC ScopedContext not set up");
-  return mlir::scf::buildLoopNest(
-      ScopedContext::getBuilderRef(), ScopedContext::getLocation(), lb, ub,
-      step, iterArgInitValues,
-      [&](OpBuilder &builder, Location loc, ValueRange ivs, ValueRange args) {
-        assert(ivs.size() == 1 && "expected one induction variable");
-        ScopedContext context(builder, loc);
-        if (fun)
-          return fun(ivs[0], args);
-        return scf::ValueVector(iterArgInitValues.begin(),
-                                iterArgInitValues.end());
-      });
-}
-
-mlir::scf::LoopNest mlir::edsc::loopNestBuilder(
-    ValueRange lbs, ValueRange ubs, ValueRange steps,
-    ValueRange iterArgInitValues,
-    function_ref<scf::ValueVector(ValueRange, ValueRange)> fun) {
-  // Delegates actual construction to scf::buildLoopNest by wrapping `fun` into
-  // the expected function interface.
-  assert(ScopedContext::getContext() && "EDSC ScopedContext not set up");
-  return mlir::scf::buildLoopNest(
-      ScopedContext::getBuilderRef(), ScopedContext::getLocation(), lbs, ubs,
-      steps, iterArgInitValues,
-      [&](OpBuilder &builder, Location loc, ValueRange ivs, ValueRange args) {
-        ScopedContext context(builder, loc);
-        if (fun)
-          return fun(ivs, args);
-        return scf::ValueVector(iterArgInitValues.begin(),
-                                iterArgInitValues.end());
-      });
-}
-
-static std::function<void(OpBuilder &, Location)>
-wrapIfBody(function_ref<scf::ValueVector()> body, TypeRange expectedTypes) {
-  (void)expectedTypes;
-  return [=](OpBuilder &builder, Location loc) {
-    ScopedContext context(builder, loc);
-    scf::ValueVector returned = body();
-    assert(ValueRange(returned).getTypes() == expectedTypes &&
-           "'if' body builder returned values of unexpected type");
-    builder.create<scf::YieldOp>(loc, returned);
-  };
-}
-
-ValueRange
-mlir::edsc::conditionBuilder(TypeRange results, Value condition,
-                             function_ref<scf::ValueVector()> thenBody,
-                             function_ref<scf::ValueVector()> elseBody,
-                             scf::IfOp *ifOp) {
-  assert(ScopedContext::getContext() && "EDSC ScopedContext not set up");
-  assert(thenBody && "thenBody is mandatory");
-
-  auto newOp = ScopedContext::getBuilderRef().create<scf::IfOp>(
-      ScopedContext::getLocation(), results, condition,
-      wrapIfBody(thenBody, results), wrapIfBody(elseBody, results));
-  if (ifOp)
-    *ifOp = newOp;
-  return newOp.getResults();
-}
-
-static std::function<void(OpBuilder &, Location)>
-wrapZeroResultIfBody(function_ref<void()> body) {
-  return [=](OpBuilder &builder, Location loc) {
-    ScopedContext context(builder, loc);
-    body();
-    builder.create<scf::YieldOp>(loc);
-  };
-}
-
-ValueRange mlir::edsc::conditionBuilder(Value condition,
-                                        function_ref<void()> thenBody,
-                                        function_ref<void()> elseBody,
-                                        scf::IfOp *ifOp) {
-  assert(ScopedContext::getContext() && "EDSC ScopedContext not set up");
-  assert(thenBody && "thenBody is mandatory");
-
-  auto newOp = ScopedContext::getBuilderRef().create<scf::IfOp>(
-      ScopedContext::getLocation(), condition, wrapZeroResultIfBody(thenBody),
-      elseBody ? llvm::function_ref<void(OpBuilder &, Location)>(
-                     wrapZeroResultIfBody(elseBody))
-               : llvm::function_ref<void(OpBuilder &, Location)>(nullptr));
-  if (ifOp)
-    *ifOp = newOp;
-  return {};
-}

mlir/test/Dialect/Linalg/affine.mlir

linalg.matmul ins(%A, %B: memref<?x?xf32>, memref<?x?xf32>)
                outs(%C: memref<?x?xf32>)
   return
 }
 
 // CHECK-LABEL: func @matmul(%{{.*}}: memref<?xi8>,
 // CHECK-SAME: [[M:arg[0-9]+]]: index
 // CHECK-SAME: [[N:arg[0-9]+]]: index
 // CHECK-SAME: [[K:arg[0-9]+]]: index
-//       CHECK: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK: affine.for %{{.*}}  = 0 to %{{.*}} {
-//       CHECK:   affine.for %{{.*}} = 0 to %{{.*}} {
-//       CHECK:     affine.for %{{.*}} = 0 to %{{.*}} {
-//   CHECK-DAG:       %[[a:.*]] = affine.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//   CHECK-DAG:       %[[b:.*]] = affine.load %[[B]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECK: %[[A:.*]] = memref.view %{{.*}} : memref<?xi8> to memref<?x?xf32>
+//       CHECK: %[[B:.*]] = memref.view %{{.*}} : memref<?xi8> to memref<?x?xf32>
+//       CHECK: %[[C:.*]] = memref.view %{{.*}} : memref<?xi8> to memref<?x?xf32>
+//       CHECK: affine.for
+//       CHECK:   affine.for
+//       CHECK:     affine.for
+//   CHECK-DAG:       %[[a:.*]] = affine.load %[[A]]{{.*}} : memref<?x?xf32>
+//   CHECK-DAG:       %[[b:.*]] = affine.load %[[B]]{{.*}} : memref<?x?xf32>
 //   CHECK-DAG:       %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
-//   CHECK-DAG:       %[[c:.*]] = affine.load %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//   CHECK-DAG:       %[[c:.*]] = affine.load %[[C]]{{.*}} : memref<?x?xf32>
 //   CHECK-DAG:       %[[res:.*]] = addf %[[c]], %[[inc]] : f32
-//       CHECK:       affine.store %[[res]], %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECK:       affine.store %[[res]], %[[C]]{{.*}} : memref<?x?xf32>
 
 func @conv_view3(%arg0: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg1: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg2: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>) {
   linalg.conv(%arg0, %arg1, %arg2) {strides = [2]}: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>
   return
 }
 
 // CHECK-LABEL: func @conv_view3(
 //  CHECK: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
 //       CHECK:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECK:   %[[Q:.*]] = memref.dim %arg0, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECK:   %[[K:.*]] = memref.dim %arg0, %c2 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECK:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECK:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECK:   affine.for %{{.*}} = 0 to %[[B]] {
-//       CHECK:     affine.for %{{.*}} = 0 to %[[X0]] {
-//       CHECK:       affine.for %{{.*}} = 0 to %[[K]] {
-//       CHECK:         affine.for %{{.*}} = 0 to %[[Q]] {
-//       CHECK:           affine.for %{{.*}} = 0 to %[[Z0]] {
-//       CHECK:            %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
+//       CHECK:   affine.for {{.*}}0 to %[[B]] {
+//       CHECK:     affine.for {{.*}}0 to %[[X0]] {
+//       CHECK:       affine.for {{.*}}0 to %[[K]] {
+//       CHECK:         affine.for {{.*}}0 to %[[Q]] {
+//       CHECK:           affine.for {{.*}}0 to %[[Z0]] {
+//       CHECK:            %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]]{{.*}}
 //       No padding needed here; only affine loads.
 //       CHECK-NEXT:       affine.load
 //       CHECK-NEXT:       affine.load
 
 func @conv_padding(%arg0: memref<?x?x?x?xf32>,
                    %arg1: memref<?x?x?x?xf32>,
                    %arg2: memref<?x?x?x?xf32>) {
   linalg.conv(%arg0, %arg1, %arg2) {dilations = [1, 1],
                                     padding = dense<[[0, 1], [1, 1]]> : tensor<2x2xi64>,
                                     strides = [1, 1]} :
     memref<?x?x?x?xf32>, memref<?x?x?x?xf32>, memref<?x?x?x?xf32>
   return
 }
 // CHECK-LABEL: func @conv_padding
 //       CHECK: %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>) {
 //       CHECK:   %[[ZERO:.*]] = constant 0.000000e+00 : f32
 //       CHECK:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32>
 //       CHECK:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32>
 //       CHECK:   %[[Q:.*]] =  memref.dim %arg0, %c2 : memref<?x?x?x?xf32>
 //       CHECK:   %[[K:.*]] =  memref.dim %arg0, %c3 : memref<?x?x?x?xf32>
 //       CHECK:   %[[B:.*]] =  memref.dim %arg1, %c0 : memref<?x?x?x?xf32>
 //       CHECK:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32>
 //       CHECK:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32>
-//       CHECK:   affine.for %{{.*}} = 0 to %[[B]] {
-//       CHECK:     affine.for %{{.*}} = 0 to %[[X0]] {
-//       CHECK:       affine.for %{{.*}} = 0 to %[[X1]] {
-//       CHECK:         affine.for %{{.*}} = 0 to %[[K]] {
-//       CHECK:           affine.for %{{.*}} = 0 to %[[Q]] {
-//       CHECK:             affine.for %{{.*}} = 0 to %[[Z0]] {
-//       CHECK:               affine.for %{{.*}} = 0 to %[[Z1]] {
-//       CHECK:                 %[[SUM0:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
-//       CHECK:                 %[[SUM1:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
+//       CHECK:   affine.for {{.*}}0 to %[[B]] {
+//       CHECK:     affine.for {{.*}}0 to %[[X0]] {
+//       CHECK:       affine.for {{.*}}0 to %[[X1]] {
+//       CHECK:         affine.for {{.*}}0 to %[[K]] {
+//       CHECK:           affine.for {{.*}}0 to %[[Q]] {
+//       CHECK:             affine.for {{.*}}0 to %[[Z0]] {
+//       CHECK:               affine.for {{.*}}0 to %[[Z1]] {
+//       CHECK:                 %[[SUM0:.*]] = affine.apply #{{.*}}
+//       CHECK:                 %[[SUM1:.*]] = affine.apply #{{.*}}
 //       CHECK:                 %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[SUM0]])
 //       CHECK:                 %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[SUM1]])
 // Padded conv involves an affine.max in the memory access and this is not
 // allowed by affine.load. Use memref.load in such cases.
-//       CHECK:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECK:                 %{{.*}} = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECK:                 %{{.*}} = affine.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECK:                 %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECK:                 %{{.*}} = affine.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECK:                 %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECK:                 affine.store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
+//       CHECK:                 memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
+//       CHECK:                 select {{.*}} : f32
+//       CHECK:                 affine.load
+//       CHECK:                 mulf {{.*}} : f32
+//       CHECK:                 affine.load
+//       CHECK:                 addf {{.*}} : f32
+//       CHECK:                 affine.store
 
 //----------------------------------------------------------------------------//
 // Named ops to loops.
 //----------------------------------------------------------------------------//
 func @named_batch_matmul(%A: memref<?x?x?xf32>, %B: memref<?x?x?xf32>, %C: memref<?x?x?xf32>) {
   linalg.batch_matmul ins(%A, %B: memref<?x?x?xf32>, memref<?x?x?xf32>)
                      outs(%C : memref<?x?x?xf32>)
   return
 }
 // CHECK-LABEL: @named_batch_matmul
 //  CHECK-SAME: %[[mA:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECK-SAME: %[[mB:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECK-SAME: %[[mC:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //       CHECK: %[[B:.*]] = memref.dim %[[mA]], %c0 : memref<?x?x?xf32>
 //       CHECK: %[[M:.*]] = memref.dim %[[mA]], %c1 : memref<?x?x?xf32>
 //       CHECK: %[[K:.*]] = memref.dim %[[mA]], %c2 : memref<?x?x?xf32>
 //       CHECK: %[[N:.*]] = memref.dim %[[mB]], %c2 : memref<?x?x?xf32>
-//       CHECK: affine.for %[[b:.*]] = 0 to %[[B]] {
-//       CHECK:   affine.for %[[m:.*]] = 0 to %[[M]] {
-//       CHECK:     affine.for %[[n:.*]] = 0 to %[[N]] {
-//       CHECK:       affine.for %[[k:.*]] = 0 to %[[K]] {
+//       CHECK: affine.for %[[b:.*]] = {{.*}}0 to %[[B]] {
+//       CHECK:   affine.for %[[m:.*]] = {{.*}}0 to %[[M]] {
+//       CHECK:     affine.for %[[n:.*]] = {{.*}}0 to %[[N]] {
+//       CHECK:       affine.for %[[k:.*]] = {{.*}}0 to %[[K]] {
 //       CHECK:       %[[va:.*]] = affine.load %[[mA]][%[[b]], %[[m]], %[[k]]] : memref<?x?x?xf32>
 //       CHECK:       %[[vb:.*]] = affine.load %[[mB]][%[[b]], %[[k]], %[[n]]] : memref<?x?x?xf32>
 //       CHECK:       %[[vc:.*]] = affine.load %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
 //       CHECK:       %[[inc:.*]] = mulf %[[va]], %[[vb]] : f32
 //       CHECK:       %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
 //       CHECK:       affine.store %[[res]], %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
 
 // CHECK-LABEL: func @pooling_max_min

mlir/test/Dialect/Linalg/loops.mlir

-// RUN: mlir-opt %s -convert-linalg-to-loops | FileCheck --check-prefix=CHECKLOOP %s
+// RUN: mlir-opt %s -convert-linalg-to-loops | FileCheck %s
 // RUN: mlir-opt %s -convert-linalg-to-parallel-loops | FileCheck --check-prefix=CHECKPARALLEL %s
 
 // Test that we can lower all the way to LLVM without crashing, don't check results here.
 // RUN: mlir-opt %s -convert-linalg-to-loops -convert-linalg-to-llvm -o=/dev/null 2>&1
 
-// CHECKLOOP-DAG: #[[$strided1D:.*]] = affine_map<(d0)[s0] -> (d0 + s0)>
-// CHECKLOOP-DAG: #[[$strided2D:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
-// CHECKLOOP-DAG: #[[$strided3D:.*]] = affine_map<(d0, d1, d2)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2 + d2)>
-// CHECKLOOP-DAG: #[[$strided4D:.*]] = affine_map<(d0, d1, d2, d3)[s0, s1, s2, s3] -> (d0 * s1 + s0 + d1 * s2 + d2 * s3 + d3)>
-// CHECKLOOP-DAG: #[[$clampMinMap:.*]] = affine_map<(d0) -> (d0, 0)>
-
-// CHECKLOOP-DAG: #[[$stride1Dilation1:.*]] = affine_map<(d0, d1) -> (d0  + d1)>
-// CHECKLOOP-DAG: #[[$stride2Dilation1:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1)>
-// CHECKLOOP-DAG: #[[$stride2Dilation4:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1 * 4)>
-// CHECKLOOP-DAG: #[[$stride3Dilation5:.*]] = affine_map<(d0, d1) -> (d0 * 3 + d1 * 5)>
-// CHECKLOOP-DAG: #[[$stride1Dilation1Padding1:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 1)>
-// CHECKLOOP-DAG: #[[$stride1Dilation1Padding2:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 2)>
+// CHECK-DAG: #[[$strided1D:.*]] = affine_map<(d0)[s0] -> (d0 + s0)>
+// CHECK-DAG: #[[$strided2D:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
+// CHECK-DAG: #[[$strided3D:.*]] = affine_map<(d0, d1, d2)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2 + d2)>
+// CHECK-DAG: #[[$strided4D:.*]] = affine_map<(d0, d1, d2, d3)[s0, s1, s2, s3] -> (d0 * s1 + s0 + d1 * s2 + d2 * s3 + d3)>
+// CHECK-DAG: #[[$clampMinMap:.*]] = affine_map<(d0) -> (d0, 0)>
+
+// CHECK-DAG: #[[$stride1Dilation1:.*]] = affine_map<(d0, d1) -> (d0  + d1)>
+// CHECK-DAG: #[[$stride2Dilation1:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1)>
+// CHECK-DAG: #[[$stride2Dilation4:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1 * 4)>
+// CHECK-DAG: #[[$stride3Dilation5:.*]] = affine_map<(d0, d1) -> (d0 * 3 + d1 * 5)>
+// CHECK-DAG: #[[$stride1Dilation1Padding1:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 1)>
+// CHECK-DAG: #[[$stride1Dilation1Padding2:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 2)>
 
 // CHECKPARALLEL-DAG: #[[$strided1D:.*]] = affine_map<(d0)[s0] -> (d0 + s0)>
 // CHECKPARALLEL-DAG: #[[$strided2D:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
 // CHECKPARALLEL-DAG: #[[$strided3D:.*]] = affine_map<(d0, d1, d2)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2 + d2)>
 // CHECKPARALLEL-DAG: #[[$strided4D:.*]] = affine_map<(d0, d1, d2, d3)[s0, s1, s2, s3] -> (d0 * s1 + s0 + d1 * s2 + d2 * s3 + d3)>
 // CHECKPARALLEL-DAG: #[[$clampMinMap:.*]] = affine_map<(d0) -> (d0, 0)>
 
 // CHECKPARALLEL-DAG: #[[$stride1Dilation1:.*]] = affine_map<(d0, d1) -> (d0  + d1)>
 // CHECKPARALLEL-DAG: #[[$stride2Dilation1:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1)>
 // CHECKPARALLEL-DAG: #[[$stride2Dilation4:.*]] = affine_map<(d0, d1) -> (d0 * 2 + d1 * 4)>
 // CHECKPARALLEL-DAG: #[[$stride3Dilation5:.*]] = affine_map<(d0, d1) -> (d0 * 3 + d1 * 5)>
 // CHECKPARALLEL-DAG: #[[$stride1Dilation1Padding1:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 1)>
 // CHECKPARALLEL-DAG: #[[$stride1Dilation1Padding2:.*]] = affine_map<(d0, d1) -> (d0 + d1 - 2)>
 
-
 func @matmul(%arg0: memref<?xi8>, %M: index, %N: index, %K: index) {
   %c0 = constant 0 : index
   %c1 = constant 1 : index
   %A = memref.view %arg0[%c0][%M, %K] : memref<?xi8> to memref<?x?xf32>
   %B = memref.view %arg0[%c0][%K, %N] : memref<?xi8> to memref<?x?xf32>
   %C = memref.view %arg0[%c0][%M, %N] : memref<?xi8> to memref<?x?xf32>
   linalg.matmul ins(%A, %B: memref<?x?xf32>, memref<?x?xf32>)
                outs(%C: memref<?x?xf32>)
   return
 }
-// CHECKLOOP-LABEL: func @matmul(%{{.*}}: memref<?xi8>,
-//  CHECKLOOP-SAME: [[M:arg[0-9]+]]: index
-//  CHECKLOOP-SAME: [[N:arg[0-9]+]]: index
-//  CHECKLOOP-SAME: [[K:arg[0-9]+]]: index
-//       CHECKLOOP: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECKLOOP: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECKLOOP: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECKLOOP: scf.for %{{.*}} = %{{.*}} to %[[M]] step %{{.*}} {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[N]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//   CHECKLOOP-DAG:       %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//   CHECKLOOP-DAG:       %[[b:.*]] = memref.load %[[B]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//   CHECKLOOP-DAG:       %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
-//   CHECKLOOP-DAG:       %[[c:.*]] = memref.load %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//   CHECKLOOP-DAG:       %[[res:.*]] = addf %[[c]], %[[inc]] : f32
-//       CHECKLOOP:       store %[[res]], %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @matmul(%{{.*}}: memref<?xi8>,
+//  CHECK-SAME: [[M:arg[0-9]+]]: index
+//  CHECK-SAME: [[N:arg[0-9]+]]: index
+//  CHECK-SAME: [[K:arg[0-9]+]]: index
+//       CHECK: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
+//       CHECK: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
+//       CHECK: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
+//       CHECK: scf.for {{.*}} to %[[M]]
+//       CHECK:   scf.for {{.*}} to %[[N]]
+//       CHECK:     scf.for {{.*}} to %[[K]]
+//   CHECK-DAG:       %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//   CHECK-DAG:       %[[b:.*]] = memref.load %[[B]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//   CHECK-DAG:       %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
+//   CHECK-DAG:       %[[c:.*]] = memref.load %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//   CHECK-DAG:       %[[res:.*]] = addf %[[c]], %[[inc]] : f32
+//       CHECK:       store %[[res]], %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @matmul(%{{.*}}: memref<?xi8>,
 //  CHECKPARALLEL-SAME: [[M:arg[0-9]+]]: index
 //  CHECKPARALLEL-SAME: [[N:arg[0-9]+]]: index
 //  CHECKPARALLEL-SAME: [[K:arg[0-9]+]]: index
 //       CHECKPARALLEL: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
 //       CHECKPARALLEL: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
 //       CHECKPARALLEL: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECKPARALLEL: scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[M]], %[[N]]) step (%{{.*}}, %{{.*}} {
-//       CHECKPARALLEL:   scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
+//       CHECKPARALLEL: scf.parallel {{.*}} to (%[[M]], %[[N]]) step (%{{.*}}, %{{.*}} {
+//       CHECKPARALLEL:   scf.for {{.*}} to %[[K]]
 //   CHECKPARALLEL-DAG:     %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 //   CHECKPARALLEL-DAG:     %[[b:.*]] = memref.load %[[B]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 //   CHECKPARALLEL-DAG:     %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
 //   CHECKPARALLEL-DAG:     %[[c:.*]] = memref.load %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 //   CHECKPARALLEL-DAG:     %[[res:.*]] = addf %[[c]], %[[inc]] : f32
 //       CHECKPARALLEL:     store %[[res]], %[[C]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 
 
 
 func @matvec(%arg0: memref<?xi8>, %M: index, %N: index) {
   %c0 = constant 0 : index
   %c1 = constant 1 : index
   %2 = memref.view %arg0[%c0][%M, %N] : memref<?xi8> to memref<?x?xf32>
   %3 = memref.view %arg0[%c0][%M] : memref<?xi8> to memref<?xf32>
   %4 = memref.view %arg0[%c0][%N] : memref<?xi8> to memref<?xf32>
   linalg.matvec ins(%2, %3: memref<?x?xf32>, memref<?xf32>)
                outs(%4 : memref<?xf32>)
   return
 }
-// CHECKLOOP-LABEL: func @matvec(%{{.*}}: memref<?xi8>,
-//  CHECKLOOP-SAME: [[M:arg[0-9]+]]: index
-//  CHECKLOOP-SAME: [[K:arg[0-9]+]]: index
-//       CHECKLOOP: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECKLOOP: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
-//       CHECKLOOP: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
-//       CHECKLOOP: scf.for %{{.*}} = %{{.*}} to %[[M]] step %{{.*}} {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//   CHECKLOOP-DAG:     %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//   CHECKLOOP-DAG:     %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
-//   CHECKLOOP-DAG:     %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
-//   CHECKLOOP-DAG:     %[[c:.*]] = memref.load %[[C]][%{{.*}}] : memref<?xf32>
-//   CHECKLOOP-DAG:     %[[res:.*]] = addf %[[c]], %[[inc]] : f32
-//       CHECKLOOP:     store %[[res]], %[[C]][%{{.*}}] : memref<?xf32>
+// CHECK-LABEL: func @matvec(%{{.*}}: memref<?xi8>,
+//  CHECK-SAME: [[M:arg[0-9]+]]: index
+//  CHECK-SAME: [[K:arg[0-9]+]]: index
+//       CHECK: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
+//       CHECK: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
+//       CHECK: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
+//       CHECK: scf.for {{.*}} to %[[M]]
+//       CHECK:   scf.for {{.*}} to %[[K]]
+//   CHECK-DAG:     %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//   CHECK-DAG:     %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
+//   CHECK-DAG:     %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
+//   CHECK-DAG:     %[[c:.*]] = memref.load %[[C]][%{{.*}}] : memref<?xf32>
+//   CHECK-DAG:     %[[res:.*]] = addf %[[c]], %[[inc]] : f32
+//       CHECK:     store %[[res]], %[[C]][%{{.*}}] : memref<?xf32>
 
 // CHECKPARALLEL-LABEL: func @matvec(%{{.*}}: memref<?xi8>,
 //  CHECKPARALLEL-SAME: [[M:arg[0-9]+]]: index
 //  CHECKPARALLEL-SAME: [[K:arg[0-9]+]]: index
 //       CHECKPARALLEL: %[[A:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?x?xf32>
 //       CHECKPARALLEL: %[[B:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
 //       CHECKPARALLEL: %[[C:.*]] = memref.view %{{.*}}[{{.*}}] : memref<?xi8> to memref<?xf32>
 //       CHECKPARALLEL: scf.parallel (%{{.*}}) = (%{{.*}}) to (%[[M]]) step (%{{.*}}) {
-//       CHECKPARALLEL:   scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
+//       CHECKPARALLEL:   scf.for {{.*}} to %[[K]]
 //   CHECKPARALLEL-DAG:     %[[a:.*]] = memref.load %[[A]][%{{.*}}, %{{.*}}] : memref<?x?xf32>
 //   CHECKPARALLEL-DAG:     %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
 //   CHECKPARALLEL-DAG:     %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
 //   CHECKPARALLEL-DAG:     %[[c:.*]] = memref.load %[[C]][%{{.*}}] : memref<?xf32>
 //   CHECKPARALLEL-DAG:     %[[res:.*]] = addf %[[c]], %[[inc]] : f32
 //       CHECKPARALLEL:     store %[[res]], %[[C]][%{{.*}}] : memref<?xf32>
 
 
 func @dot(%arg0: memref<?xi8>, %M: index) {
   %c0 = constant 0 : index
   %c1 = constant 1 : index
   %1 = memref.view %arg0[%c0][%M] : memref<?xi8> to memref<?xf32>
   %2 = memref.view %arg0[%c0][%M] : memref<?xi8> to memref<?xf32>
   %3 = memref.view %arg0[%c0][] : memref<?xi8> to memref<f32>
   linalg.dot ins(%1, %2 : memref<?xf32>, memref<?xf32>)
             outs(%3 : memref<f32>)
   return
 }
-// CHECKLOOP-LABEL: func @dot(%{{.*}}: memref<?xi8>,
-//  CHECKLOOP-SAME: [[K:arg[0-9]+]]: index
-//       CHECKLOOP: %[[A:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
-//       CHECKLOOP: %[[B:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
-//       CHECKLOOP: %[[C:.*]] = memref.view %{{.*}}[{{.*}}][] : memref<?xi8> to memref<f32>
-//       CHECKLOOP: scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//   CHECKLOOP-DAG:   %[[a:.*]] = memref.load %[[A]][%{{.*}}] : memref<?xf32>
-//   CHECKLOOP-DAG:   %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
-//   CHECKLOOP-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
-//   CHECKLOOP-DAG:   %[[c:.*]] = memref.load %[[C]][] : memref<f32>
-//   CHECKLOOP-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
-//       CHECKLOOP:   store %[[res]], %[[C]][] : memref<f32>
+// CHECK-LABEL: func @dot(%{{.*}}: memref<?xi8>,
+//  CHECK-SAME: [[K:arg[0-9]+]]: index
+//       CHECK: %[[A:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
+//       CHECK: %[[B:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
+//       CHECK: %[[C:.*]] = memref.view %{{.*}}[{{.*}}][] : memref<?xi8> to memref<f32>
+//       CHECK: scf.for {{.*}} to %[[K]]
+//   CHECK-DAG:   %[[a:.*]] = memref.load %[[A]][%{{.*}}] : memref<?xf32>
+//   CHECK-DAG:   %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
+//   CHECK-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
+//   CHECK-DAG:   %[[c:.*]] = memref.load %[[C]][] : memref<f32>
+//   CHECK-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
+//       CHECK:   store %[[res]], %[[C]][] : memref<f32>
 
 // CHECKPARALLEL-LABEL: func @dot(%{{.*}}: memref<?xi8>,
 //  CHECKPARALLEL-SAME: [[K:arg[0-9]+]]: index
 //       CHECKPARALLEL: %[[A:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
 //       CHECKPARALLEL: %[[B:.*]] = memref.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?xf32>
 //       CHECKPARALLEL: %[[C:.*]] = memref.view %{{.*}}[{{.*}}][] : memref<?xi8> to memref<f32>
-//       CHECKPARALLEL: scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
+//       CHECKPARALLEL: scf.for {{.*}} to %[[K]]
 //   CHECKPARALLEL-DAG:   %[[a:.*]] = memref.load %[[A]][%{{.*}}] : memref<?xf32>
 //   CHECKPARALLEL-DAG:   %[[b:.*]] = memref.load %[[B]][%{{.*}}] : memref<?xf32>
 //   CHECKPARALLEL-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
 //   CHECKPARALLEL-DAG:   %[[c:.*]] = memref.load %[[C]][] : memref<f32>
 //   CHECKPARALLEL-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
 //       CHECKPARALLEL:   store %[[res]], %[[C]][] : memref<f32>
 
 
 func @dot_view(%arg0: memref<?xf32, offset: ?, strides: [1]>, %arg1: memref<?xf32, offset: ?, strides: [1]>, %arg2: memref<f32>) {
   linalg.dot ins(%arg0, %arg1 : memref<?xf32, offset: ?, strides: [1]>,
                                 memref<?xf32, offset: ?, strides: [1]>)
             outs(%arg2:  memref<f32>)
   return
 }
-// CHECKLOOP-LABEL: func @dot_view(
-//       CHECKLOOP:   %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<f32>) {
-//       CHECKLOOP: %[[K:.*]] = memref.dim %arg0, %c0 : memref<?xf32, #[[$strided1D]]>
-//       CHECKLOOP: scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//   CHECKLOOP-DAG:   %[[a:.*]] = memref.load %arg0[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
-//   CHECKLOOP-DAG:   %[[b:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
-//   CHECKLOOP-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
-//   CHECKLOOP-DAG:   %[[c:.*]] = memref.load %{{.*}}[] : memref<f32>
-//   CHECKLOOP-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
-//       CHECKLOOP:   store %[[res]], %{{.*}}[] : memref<f32>
+// CHECK-LABEL: func @dot_view(
+//       CHECK:   %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<f32>) {
+//       CHECK: %[[K:.*]] = memref.dim %arg0, %c0 : memref<?xf32, #[[$strided1D]]>
+//       CHECK: scf.for {{.*}} to %[[K]]
+//   CHECK-DAG:   %[[a:.*]] = memref.load %arg0[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
+//   CHECK-DAG:   %[[b:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
+//   CHECK-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
+//   CHECK-DAG:   %[[c:.*]] = memref.load %{{.*}}[] : memref<f32>
+//   CHECK-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
+//       CHECK:   store %[[res]], %{{.*}}[] : memref<f32>
 
 // CHECKPARALLEL-LABEL: func @dot_view(
 //       CHECKPARALLEL:   %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<f32>) {
 //       CHECKPARALLEL: %[[K:.*]] = memref.dim %arg0, %c0 : memref<?xf32, #[[$strided1D]]>
-//       CHECKPARALLEL: scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
+//       CHECKPARALLEL: scf.for {{.*}} to %[[K]]
 //   CHECKPARALLEL-DAG:   %[[a:.*]] = memref.load %arg0[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 //   CHECKPARALLEL-DAG:   %[[b:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 //   CHECKPARALLEL-DAG:   %[[inc:.*]] = mulf %[[a]], %[[b]] : f32
 //   CHECKPARALLEL-DAG:   %[[c:.*]] = memref.load %{{.*}}[] : memref<f32>
 //   CHECKPARALLEL-DAG:   %[[res:.*]] = addf %[[c]], %[[inc]] : f32
 //       CHECKPARALLEL:   store %[[res]], %{{.*}}[] : memref<f32>
 
 func @fill_view(%arg0: memref<?xf32, offset: ?, strides: [1]>, %arg1: f32) {
   linalg.fill(%arg0, %arg1) : memref<?xf32, offset: ?, strides: [1]>, f32
   return
 }
-// CHECKLOOP-LABEL: func @fill_view(
-//       CHECKLOOP: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: f32) {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:     store %{{.*}}, %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
+// CHECK-LABEL: func @fill_view(
+//       CHECK: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: f32) {
+//       CHECK:   scf.for {{.*}} to %{{.*}}
+//       CHECK:     store %{{.*}}, %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 
 // CHECKPARALLEL-LABEL: func @fill_view(
 //       CHECKPARALLEL: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: f32) {
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}) = (%{{.*}}) to (%{{.*}}) step (%{{.*}}) {
 //       CHECKPARALLEL:     store %{{.*}}, %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 
 func @fill_view0(%arg0: memref<f32>, %arg1: f32) {
   linalg.fill(%arg0, %arg1) : memref<f32>, f32
   return
 }
-// CHECKLOOP-LABEL: func @fill_view0(%{{.*}}: memref<f32>, %{{.*}}: f32) {
-//       CHECKLOOP:   store %{{.*}}, %{{.*}}[] : memref<f32>
+// CHECK-LABEL: func @fill_view0(%{{.*}}: memref<f32>, %{{.*}}: f32) {
+//       CHECK:   store %{{.*}}, %{{.*}}[] : memref<f32>
 
 // CHECKPARALLEL-LABEL: func @fill_view0(%{{.*}}: memref<f32>, %{{.*}}: f32) {
 //       CHECKPARALLEL:   store %{{.*}}, %{{.*}}[] : memref<f32>
 
 func @fill_view3(%arg0: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg1: f32) {
   linalg.fill(%arg0, %arg1) : memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, f32
   return
 }
-// CHECKLOOP-LABEL: func @fill_view3(
-//       CHECKLOOP: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: f32) {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:         store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+// CHECK-LABEL: func @fill_view3(
+//       CHECK: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: f32) {
+//       CHECK:   scf.for {{.*}} to %{{.*}}
+//       CHECK:     scf.for {{.*}} to %{{.*}}
+//       CHECK:       scf.for {{.*}} to %{{.*}}
+//       CHECK:         store %{{.*}}, {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 // CHECKPARALLEL-LABEL: func @fill_view3(
 //       CHECKPARALLEL: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: f32) {
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}, %{{.*}}) to (%{{.*}}, %{{.*}}, %{{.*}}) step (%{{.*}}, %{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:     store %{{.*}}, {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 func @copy_view(%arg0: memref<?xf32, offset: ?, strides: [1]>, %arg1: memref<?xf32, offset: ?, strides: [1]>) {
   linalg.copy(%arg0, %arg1) : memref<?xf32, offset: ?, strides: [1]>, memref<?xf32, offset: ?, strides: [1]>
   return
 }
-// CHECKLOOP-LABEL: func @copy_view(
-//       CHECKLOOP: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>) {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:     %[[L:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
-//       CHECKLOOP:     store %[[L]], %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
+// CHECK-LABEL: func @copy_view(
+//       CHECK: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>) {
+//       CHECK:   scf.for {{.*}} to %{{.*}}
+//       CHECK:     %[[L:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
+//       CHECK:     store %[[L]], %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 
 // CHECKPARALLEL-LABEL: func @copy_view(
 //       CHECKPARALLEL: %{{.*}}: memref<?xf32, #[[$strided1D]]>, %{{.*}}: memref<?xf32, #[[$strided1D]]>) {
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}) = (%{{.*}}) to (%{{.*}}) step (%{{.*}}) {
 //       CHECKPARALLEL:     %[[L:.*]] = memref.load %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 //       CHECKPARALLEL:     store %[[L]], %{{.*}}[%{{.*}}] : memref<?xf32, #[[$strided1D]]>
 
 func @copy_view0(%arg0: memref<f32>, %arg1: memref<f32>) {
   linalg.copy(%arg0, %arg1) : memref<f32>, memref<f32>
   return
 }
-// CHECKLOOP-LABEL: func @copy_view0(%{{.*}}: memref<f32>, %{{.*}}: memref<f32>) {
-//       CHECKLOOP:   %{{.*}} = memref.load %{{.*}}[] : memref<f32>
-//       CHECKLOOP:   store %{{.*}}, %{{.*}}[] : memref<f32>
+// CHECK-LABEL: func @copy_view0(%{{.*}}: memref<f32>, %{{.*}}: memref<f32>) {
+//       CHECK:   memref.load %{{.*}}[] : memref<f32>
+//       CHECK:   store %{{.*}}, %{{.*}}[] : memref<f32>
 
 // CHECKPARALLEL-LABEL: func @copy_view0(%{{.*}}: memref<f32>, %{{.*}}: memref<f32>) {
-//       CHECKPARALLEL:   %{{.*}} = memref.load %{{.*}}[] : memref<f32>
+//       CHECKPARALLEL:   memref.load %{{.*}}[] : memref<f32>
 //       CHECKPARALLEL:   store %{{.*}}, %{{.*}}[] : memref<f32>
 
 func @copy_view3(%arg0: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg1: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>) {
   linalg.copy(%arg0, %arg1) {inputPermutation = affine_map<(i, j, k) -> (i, k, j)>,
                              outputPermutation = affine_map<(i, j, k) -> (k, j, i)>} :
     memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>
   return
 }
-// CHECKLOOP-LABEL: func @copy_view3
-//       CHECKLOOP: (%{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
-//       CHECKLOOP:         %[[L:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:         store %[[L]], %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+// CHECK-LABEL: func @copy_view3
+//       CHECK: (%{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
+//       CHECK:   scf.for {{.*}} to %{{.*}}
+//       CHECK:     scf.for {{.*}} to %{{.*}}
+//       CHECK:       scf.for {{.*}} to %{{.*}}
+//       CHECK:         %[[L:.*]] = memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:         store %[[L]], {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 // CHECKPARALLEL-LABEL: func @copy_view3
 //       CHECKPARALLEL: (%{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}, %{{.*}}) to (%{{.*}}, %{{.*}}, %{{.*}}) step (%{{.*}}, %{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     %[[L:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKPARALLEL:     store %[[L]], %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:     %[[L:.*]] = memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:     store %[[L]], {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 func @conv_view3(%arg0: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg1: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, %arg2: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>) {
   linalg.conv(%arg0, %arg1, %arg2) {strides = [2]}: memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>, memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>
   return
 }
-// CHECKLOOP-LABEL: func @conv_view3(
-//       CHECKLOOP: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
-//       CHECKLOOP:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:   %[[Q:.*]] = memref.dim %arg0, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:   %[[K:.*]] = memref.dim %arg0, %c2 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[B]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[X0]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKLOOP:           scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKLOOP:             %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:             %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[SUM]], %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:             %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:             %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:             %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:             %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:             store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+// CHECK-LABEL: func @conv_view3(
+//       CHECK: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
+//       CHECK:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:   %[[Q:.*]] = memref.dim %arg0, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:   %[[K:.*]] = memref.dim %arg0, %c2 : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:   scf.for {{.*}} to %[[B]]
+//       CHECK:     scf.for {{.*}} to %[[X0]]
+//       CHECK:       scf.for {{.*}} to %[[K]]
+//       CHECK:         scf.for {{.*}} to %[[Q]]
+//       CHECK:           scf.for {{.*}} to %[[Z0]]
+//       CHECK:             %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECK:             memref.load %{{.*}}[%{{.*}}, %[[SUM]], %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:             memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:             mulf
+//       CHECK:             memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:             addf
+//       CHECK:             store %{{.*}}, {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 // CHECKPARALLEL-LABEL: func @conv_view3(
 //       CHECKPARALLEL: %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>, %{{.*}}: memref<?x?x?xf32, #[[$strided3D]]>) {
 //       CHECKPARALLEL:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[Q:.*]] = memref.dim %arg0, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[K:.*]] = memref.dim %arg0, %c2 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}, %{{.*}}) to (%[[B]], %[[X0]], %[[K]]) step (%{{.*}}, %{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[SUM]], %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKPARALLEL:         %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKPARALLEL:         %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:         store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[Q]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[Z0]]
+//       CHECKPARALLEL:         %[[SUM:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECKPARALLEL:         memref.load %{{.*}}[%{{.*}}, %[[SUM]], %{{.*}}] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:         memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:         mulf
+//       CHECKPARALLEL:         memref.load {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECKPARALLEL:         addf
+//       CHECKPARALLEL:         store %{{.*}}, {{.*}} : memref<?x?x?xf32, #[[$strided3D]]>
 
 func @conv_view4(%arg0: memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>, %arg1: memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>, %arg2: memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>) {
   linalg.conv(%arg0, %arg1, %arg2) {dilations = [4, 5], strides = [2, 3]} : memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>, memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>, memref<?x?x?x?xf32, offset: ?, strides: [?, ?, ?, 1]>
   return
 }
-// CHECKLOOP-LABEL: func @conv_view4(
-//       CHECKLOOP: %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>) {
-//       CHECKLOOP:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[Q:.*]] = memref.dim %arg0, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[K:.*]] = memref.dim %arg0, %c3 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[B]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[X0]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[X1]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//       CHECKLOOP:           scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKLOOP:             scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKLOOP:               scf.for %{{.*}} = %{{.*}} to %[[Z1]] step %{{.*}} {
-//       CHECKLOOP:                 %[[SUM0:.*]] = affine.apply #[[$stride2Dilation4]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:                 %[[SUM1:.*]] = affine.apply #[[$stride3Dilation5]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[SUM0]], %[[SUM1]], %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:                 %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKLOOP:                 %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:                 store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
+// CHECK-LABEL: func @conv_view4(
+//       CHECK: %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>) {
+//       CHECK:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[Q:.*]] = memref.dim %arg0, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[K:.*]] = memref.dim %arg0, %c3 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:   scf.for {{.*}} to %[[B]]
+//       CHECK:     scf.for {{.*}} to %[[X0]]
+//       CHECK:       scf.for {{.*}} to %[[X1]]
+//       CHECK:         scf.for {{.*}} to %[[K]]
+//       CHECK:           scf.for {{.*}} to %[[Q]]
+//       CHECK:             scf.for {{.*}} to %[[Z0]]
+//       CHECK:               scf.for {{.*}} to %[[Z1]]
+//       CHECK:                 %[[SUM0:.*]] = affine.apply #[[$stride2Dilation4]]
+//       CHECK:                 %[[SUM1:.*]] = affine.apply #[[$stride3Dilation5]]
+//       CHECK:                 memref.load %{{.*}}[%{{.*}}, %[[SUM0]], %[[SUM1]], %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:                 memref.load {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:                 mulf
+//       CHECK:                 memref.load {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECK:                 addf
+//       CHECK:                 store %{{.*}}, {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
 
 // CHECKPARALLEL-LABEL: func @conv_view4(
 //       CHECKPARALLEL: %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>, %{{.*}}: memref<?x?x?x?xf32, #[[$strided4D]]>) {
 //       CHECKPARALLEL:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[Q:.*]] = memref.dim %arg0, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[K:.*]] = memref.dim %arg0, %c3 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[B:.*]] = memref.dim %arg1, %c0 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32, #[[$strided4D]]>
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) to (%[[B]], %[[X0]], %[[X1]], %[[K]]) step (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKPARALLEL:         scf.for %{{.*}} = %{{.*}} to %[[Z1]] step %{{.*}} {
-//       CHECKPARALLEL:           %[[SUM0:.*]] = affine.apply #[[$stride2Dilation4]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:           %[[SUM1:.*]] = affine.apply #[[$stride3Dilation5]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[SUM0]], %[[SUM1]], %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKPARALLEL:           %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
-//       CHECKPARALLEL:           %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:           store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[Q]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[Z0]]
+//       CHECKPARALLEL:         scf.for {{.*}} to %[[Z1]]
+//       CHECKPARALLEL:           %[[SUM0:.*]] = affine.apply #[[$stride2Dilation4]]
+//       CHECKPARALLEL:           %[[SUM1:.*]] = affine.apply #[[$stride3Dilation5]]
+//       CHECKPARALLEL:           memref.load %{{.*}}[%{{.*}}, %[[SUM0]], %[[SUM1]], %{{.*}}] : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECKPARALLEL:           memref.load {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECKPARALLEL:           mulf
+//       CHECKPARALLEL:           memref.load {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
+//       CHECKPARALLEL:           addf
+//       CHECKPARALLEL:           store %{{.*}}, {{.*}} : memref<?x?x?x?xf32, #[[$strided4D]]>
 
 func @conv_padding(%arg0: memref<?x?x?x?xf32>,
                    %arg1: memref<?x?x?x?xf32>,
                    %arg2: memref<?x?x?x?xf32>) {
   linalg.conv(%arg0, %arg1, %arg2) {dilations = [1, 1],
                                     padding = dense<[[0, 1], [1, 1]]> : tensor<2x2xi64>,
                                     strides = [1, 1]} :
     memref<?x?x?x?xf32>, memref<?x?x?x?xf32>, memref<?x?x?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @conv_padding
-//       CHECKLOOP: %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>) {
-//       CHECKLOOP:   %[[ZERO:.*]] = constant 0.000000e+00 : f32
-//       CHECKLOOP:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[Q:.*]] =  memref.dim %arg0, %c2 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[K:.*]] =  memref.dim %arg0, %c3 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[B:.*]] =  memref.dim %arg1, %c0 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[B]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[X0]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[X1]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[K]] step %{{.*}} {
-//       CHECKLOOP:           scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKLOOP:             scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKLOOP:               scf.for %{{.*}} = %{{.*}} to %[[Z1]] step %{{.*}} {
-//       CHECKLOOP:                 %[[SUM0:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
-//       CHECKLOOP:                 %[[SUM1:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
-//       CHECKLOOP:                 %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[SUM0]])
-//       CHECKLOOP:                 %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[SUM1]])
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKLOOP:                 %{{.*}} = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKLOOP:                 %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:                 %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKLOOP:                 %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:                 store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
+// CHECK-LABEL: func @conv_padding
+//       CHECK: %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>) {
+//       CHECK:   %[[ZERO:.*]] = constant 0.000000e+00 : f32
+//       CHECK:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32>
+//       CHECK:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32>
+//       CHECK:   %[[Q:.*]] =  memref.dim %arg0, %c2 : memref<?x?x?x?xf32>
+//       CHECK:   %[[K:.*]] =  memref.dim %arg0, %c3 : memref<?x?x?x?xf32>
+//       CHECK:   %[[B:.*]] =  memref.dim %arg1, %c0 : memref<?x?x?x?xf32>
+//       CHECK:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32>
+//       CHECK:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[B]]
+//       CHECK:     scf.for {{.*}} to %[[X0]]
+//       CHECK:       scf.for {{.*}} to %[[X1]]
+//       CHECK:         scf.for {{.*}} to %[[K]]
+//       CHECK:           scf.for {{.*}} to %[[Q]]
+//       CHECK:             scf.for {{.*}} to %[[Z0]]
+//       CHECK:               scf.for {{.*}} to %[[Z1]]
+//       CHECK:                 %[[SUM0:.*]] = affine.apply #{{.*}}
+//       CHECK:                 %[[SUM1:.*]] = affine.apply #{{.*}}
+//       CHECK:                 %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[SUM0]])
+//       CHECK:                 %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[SUM1]])
+//       CHECK:                 memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
+//       CHECK:                 select %{{.*}},
+//       CHECK:                 memref.load {{.*}} : memref<?x?x?x?xf32>
+//       CHECK:                 mulf
+//       CHECK:                 memref.load {{.*}} : memref<?x?x?x?xf32>
+//       CHECK:                 addf
+//       CHECK:                 store %{{.*}}, {{.*}} : memref<?x?x?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @conv_padding
 //       CHECKPARALLEL: %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>, %{{.*}}: memref<?x?x?x?xf32>) {
 //       CHECKPARALLEL:   %[[ZERO:.*]] = constant 0.000000e+00 : f32
 //       CHECKPARALLEL:   %[[Z0:.*]] = memref.dim %arg0, %c0 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[Z1:.*]] = memref.dim %arg0, %c1 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[Q:.*]] =  memref.dim %arg0, %c2 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[K:.*]] =  memref.dim %arg0, %c3 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[B:.*]] =  memref.dim %arg1, %c0 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[X0:.*]] = memref.dim %arg2, %c1 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   %[[X1:.*]] = memref.dim %arg2, %c2 : memref<?x?x?x?xf32>
 //       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) to (%[[B]], %[[X0]], %[[X1]], %[[K]]) step (%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[Q]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[Z0]] step %{{.*}} {
-//       CHECKPARALLEL:         scf.for %{{.*}} = %{{.*}} to %[[Z1]] step %{{.*}} {
-//       CHECKPARALLEL:           %[[SUM0:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:           %[[SUM1:.*]] = affine.apply #{{.*}}(%{{.*}}, %{{.*}})
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[Q]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[Z0]]
+//       CHECKPARALLEL:         scf.for {{.*}} to %[[Z1]]
+//       CHECKPARALLEL:           %[[SUM0:.*]] = affine.apply #{{.*}}
+//       CHECKPARALLEL:           %[[SUM1:.*]] = affine.apply #{{.*}}
 //       CHECKPARALLEL:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[SUM0]])
 //       CHECKPARALLEL:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[SUM1]])
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKPARALLEL:           %{{.*}} = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKPARALLEL:           %{{.*}} = mulf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
-//       CHECKPARALLEL:           %{{.*}} = addf %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:           store %{{.*}}, %{{.*}}[%{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}] : memref<?x?x?x?xf32>
+//       CHECKPARALLEL:           memref.load %{{.*}}[%{{.*}}, %[[IDX]], %[[IDY]], %{{.*}}] : memref<?x?x?x?xf32>
+//       CHECKPARALLEL:           select %{{.*}},
+//       CHECKPARALLEL:           memref.load {{.*}} : memref<?x?x?x?xf32>
+//       CHECKPARALLEL:           mulf
+//       CHECKPARALLEL:           memref.load {{.*}} : memref<?x?x?x?xf32>
+//       CHECKPARALLEL:           addf
+//       CHECKPARALLEL:           store %{{.*}}, {{.*}} : memref<?x?x?x?xf32>
 
 func @pooling_max(%arg0: memref<?x?xf32>,
                   %arg1: memref<?x?xi32>,
                   %arg2: memref<?x?xf32>) {
   linalg.pooling_max(%arg0, %arg1, %arg2) { strides = [2, 1] }:
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_max
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %{{.*}} = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_max
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
+//       CHECK:           memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
+//       CHECK:           %[[RES:.*]] = select %{{.*}},
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_max
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
+//       CHECKPARALLEL:         memref.load {{.*}} : memref<?x?xf32>
+//       CHECKPARALLEL:         memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
+//       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}},
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_max_padding(%arg0: memref<?x?xf32>,
                           %arg1: memref<?x?xi32>,
                           %arg2: memref<?x?xf32>) {
   linalg.pooling_max(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_max_padding
-//       CHECKLOOP:   %[[PAD:.*]] = constant 0xFF800000 : f32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
-//       CHECKLOOP:           %[[CMP:.*]] = cmpf ogt, %[[RHS]], %[[SEL]] : f32
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_max_padding
+//       CHECK:   %[[PAD:.*]] = constant 0xFF800000 : f32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
+//       CHECK:           %[[CMP:.*]] = cmpf ogt, %[[RHS]], %[[SEL]] : f32
+//       CHECK:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_max_padding
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant 0xFF800000 : f32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
 //       CHECKPARALLEL:         %[[CMP:.*]] = cmpf ogt, %[[RHS]], %[[SEL]] : f32
 //       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_max_padding_i32(%arg0: memref<?x?xi32>,
                               %arg1: memref<?x?xi32>,
                               %arg2: memref<?x?xi32>) {
   linalg.pooling_max(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xi32>, memref<?x?xi32>, memref<?x?xi32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_max_padding_i32
-//       CHECKLOOP:   %[[PAD:.*]] = constant -2147483648 : i32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
-//       CHECKLOOP:           %[[CMP:.*]] = cmpi sgt, %[[RHS]], %[[SEL]] : i32
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+// CHECK-LABEL: func @pooling_max_padding_i32
+//       CHECK:   %[[PAD:.*]] = constant -2147483648 : i32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
+//       CHECK:           %[[CMP:.*]] = cmpi sgt, %[[RHS]], %[[SEL]] : i32
+//       CHECK:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xi32>
 
 // CHECKPARALLEL-LABEL: func @pooling_max_padding_i32
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant -2147483648 : i32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
 //       CHECKPARALLEL:         %[[CMP:.*]] = cmpi sgt, %[[RHS]], %[[SEL]] : i32
 //       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xi32>
 
 func @pooling_min(%arg0: memref<?x?xf32>,
                   %arg1: memref<?x?xi32>,
                   %arg2: memref<?x?xf32>) {
   linalg.pooling_min(%arg0, %arg1, %arg2) { strides = [2, 1] }:
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_min
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %{{.*}} = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_min
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
+//       CHECK:           memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
+//       CHECK:           %[[RES:.*]] = select %{{.*}},
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_min
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKPARALLEL:         %{{.*}} = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %{{.*}}, %{{.*}} : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
+//       CHECKPARALLEL:         memref.load {{.*}} : memref<?x?xf32>
+//       CHECKPARALLEL:         memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
+//       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}},
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_min_padding(%arg0: memref<?x?xf32>,
                           %arg1: memref<?x?xi32>,
                           %arg2: memref<?x?xf32>) {
   linalg.pooling_min(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_min_padding
-//       CHECKLOOP:   %[[PAD:.*]] = constant 0x7F800000 : f32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
-//       CHECKLOOP:           %[[CMP:.*]] = cmpf olt, %[[RHS]], %[[SEL]] : f32
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_min_padding
+//       CHECK:   %[[PAD:.*]] = constant 0x7F800000 : f32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
+//       CHECK:           %[[CMP:.*]] = cmpf olt, %[[RHS]], %[[SEL]] : f32
+//       CHECK:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_min_padding
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant 0x7F800000 : f32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
 //       CHECKPARALLEL:         %[[CMP:.*]] = cmpf olt, %[[RHS]], %[[SEL]] : f32
 //       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_min_padding_i32(%arg0: memref<?x?xi32>,
                               %arg1: memref<?x?xi32>,
                               %arg2: memref<?x?xi32>) {
   linalg.pooling_min(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xi32>, memref<?x?xi32>, memref<?x?xi32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_min_padding_i32
-//       CHECKLOOP:   %[[PAD:.*]] = constant 2147483647 : i32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
-//       CHECKLOOP:           %[[CMP:.*]] = cmpi slt, %[[RHS]], %[[SEL]] : i32
-//       CHECKLOOP:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+// CHECK-LABEL: func @pooling_min_padding_i32
+//       CHECK:   %[[PAD:.*]] = constant 2147483647 : i32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
+//       CHECK:           %[[CMP:.*]] = cmpi slt, %[[RHS]], %[[SEL]] : i32
+//       CHECK:           %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xi32>
 
 // CHECKPARALLEL-LABEL: func @pooling_min_padding_i32
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant 2147483647 : i32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
 //       CHECKPARALLEL:         %[[CMP:.*]] = cmpi slt, %[[RHS]], %[[SEL]] : i32
 //       CHECKPARALLEL:         %[[RES:.*]] = select %{{.*}}, %[[RHS]], %[[SEL]] : i32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xi32>
 
 func @pooling_sum(%arg0: memref<?x?xf32>,
                   %arg1: memref<?x?xi32>,
                   %arg2: memref<?x?xf32>) {
   linalg.pooling_sum(%arg0, %arg1, %arg2) { strides = [2, 1] }:
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_sum
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %[[RES:.*]] = addf %[[LHS]], %[[RHS]] : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_sum
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
+//       CHECK:           %[[RHS:.*]] = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
+//       CHECK:           %[[LHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           %[[RES:.*]] = addf %[[LHS]], %[[RHS]] : f32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_sum
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]](%{{.*}}, %{{.*}})
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride2Dilation1]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1]]
 //       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%[[IX]], %[[IY]]] : memref<?x?xf32>
-//       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         %[[LHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[RES:.*]] = addf %[[LHS]], %[[RHS]] : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_sum_padding(%arg0: memref<?x?xf32>,
                           %arg1: memref<?x?xi32>,
                           %arg2: memref<?x?xf32>) {
   linalg.pooling_sum(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xf32>, memref<?x?xi32>, memref<?x?xf32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_sum_padding
-//       CHECKLOOP:   %[[PAD:.*]] = constant 0.000000e+00 : f32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
-//       CHECKLOOP:           %[[RES:.*]] = addf %[[RHS]], %[[SEL]] : f32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+// CHECK-LABEL: func @pooling_sum_padding
+//       CHECK:   %[[PAD:.*]] = constant 0.000000e+00 : f32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
+//       CHECK:           %[[RES:.*]] = addf %[[RHS]], %[[SEL]] : f32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: func @pooling_sum_padding
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant 0.000000e+00 : f32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xf32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xf32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : f32
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xf32>
 //       CHECKPARALLEL:         %[[RES:.*]] = addf %[[RHS]], %[[SEL]] : f32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xf32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xf32>
 
 func @pooling_sum_padding_i32(%arg0: memref<?x?xi32>,
                               %arg1: memref<?x?xi32>,
                               %arg2: memref<?x?xi32>) {
   linalg.pooling_sum(%arg0, %arg1, %arg2) { padding = dense<[[2, 2], [1, 1]]> : tensor<2x2xi64> } :
     memref<?x?xi32>, memref<?x?xi32>, memref<?x?xi32>
   return
 }
-// CHECKLOOP-LABEL: func @pooling_sum_padding_i32
-//       CHECKLOOP:   %[[PAD:.*]] = constant 0 : i32
-//       CHECKLOOP:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
-//       CHECKLOOP:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKLOOP:   scf.for %{{.*}} = %{{.*}} to %[[OX]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %{{.*}} = %{{.*}} to %[[OY]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKLOOP:         scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKLOOP:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
-//       CHECKLOOP:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
-//       CHECKLOOP:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
-//       CHECKLOOP:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
-//       CHECKLOOP:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
-//       CHECKLOOP:           %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
-//       CHECKLOOP:           %[[RES:.*]] = addi %[[RHS]], %[[SEL]] : i32
-//       CHECKLOOP:           store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+// CHECK-LABEL: func @pooling_sum_padding_i32
+//       CHECK:   %[[PAD:.*]] = constant 0 : i32
+//       CHECK:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
+//       CHECK:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
+//       CHECK:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
+//       CHECK:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
+//       CHECK:   scf.for {{.*}} to %[[OX]]
+//       CHECK:     scf.for {{.*}} to %[[OY]]
+//       CHECK:       scf.for {{.*}} to %[[WX]]
+//       CHECK:         scf.for {{.*}} to %[[WY]]
+//       CHECK:           %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECK:           %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
+//       CHECK:           %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
+//       CHECK:           %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
+//       CHECK:           %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
+//       CHECK:           %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
+//       CHECK:           %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
+//       CHECK:           %[[RES:.*]] = addi %[[RHS]], %[[SEL]] : i32
+//       CHECK:           store %[[RES]], {{.*}} : memref<?x?xi32>
 
 // CHECKPARALLEL-LABEL: func @pooling_sum_padding_i32
 //       CHECKPARALLEL:   %[[PAD:.*]] = constant 0 : i32
 //       CHECKPARALLEL:   %[[WX:.*]] = memref.dim %arg1, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[WY:.*]] = memref.dim %arg1, %c1 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OX:.*]] = memref.dim %arg2, %c0 : memref<?x?xi32>
 //       CHECKPARALLEL:   %[[OY:.*]] = memref.dim %arg2, %c1 : memref<?x?xi32>
-//       CHECKPARALLEL:   scf.parallel (%{{.*}}, %{{.*}}) = (%{{.*}}, %{{.*}}) to (%[[OX]], %[[OY]]) step (%{{.*}}, %{{.*}}) {
-//       CHECKPARALLEL:     scf.for %{{.*}} = %{{.*}} to %[[WX]] step %{{.*}} {
-//       CHECKPARALLEL:       scf.for %{{.*}} = %{{.*}} to %[[WY]] step %{{.*}} {
-//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]](%{{.*}}, %{{.*}})
-//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]](%{{.*}}, %{{.*}})
+//       CHECKPARALLEL:   scf.parallel {{.*}} to (%[[OX]], %[[OY]])
+//       CHECKPARALLEL:     scf.for {{.*}} to %[[WX]]
+//       CHECKPARALLEL:       scf.for {{.*}} to %[[WY]]
+//       CHECKPARALLEL:         %[[IX:.*]] = affine.apply #[[$stride1Dilation1Padding2]]
+//       CHECKPARALLEL:         %[[IY:.*]] = affine.apply #[[$stride1Dilation1Padding1]]
 //       CHECKPARALLEL:         %[[IDX:.*]] = affine.max #[[$clampMinMap]](%[[IX]])
 //       CHECKPARALLEL:         %[[IDY:.*]] = affine.max #[[$clampMinMap]](%[[IY]])
 //       CHECKPARALLEL:         %[[LHS:.*]] = memref.load %{{.*}}[%[[IDX]], %[[IDY]]] : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[SEL:.*]] = select %{{.*}}, %[[PAD]], %[[LHS]] : i32
-//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:         %[[RHS:.*]] = memref.load {{.*}} : memref<?x?xi32>
 //       CHECKPARALLEL:         %[[RES:.*]] = addi %[[RHS]], %[[SEL]] : i32
-//       CHECKPARALLEL:         store %[[RES]], %{{.*}}[%{{.*}}, %{{.*}}] : memref<?x?xi32>
+//       CHECKPARALLEL:         store %[[RES]], {{.*}} : memref<?x?xi32>
 
 #accesses = [
   affine_map<(i, j, k) -> (i, j)>,
   affine_map<(i, j, k) -> (i, j, k)>,
   affine_map<(i, j, k) -> (i, k, j)>
 ]
 #trait2 = {
   args_in = 1,
outs(%arg1, %arg2 : memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>,
                        memref<?x?x?xf32, offset: ?, strides: [?, ?, 1]>) {
     ^bb0(%a: f32, %b: f32, %c: f32):
       %d = mulf %a, %b : f32
       %e = addf %c, %d : f32
       linalg.yield %d, %e : f32, f32
   }
   return
 }
-// CHECKLOOP-LABEL: @generic_region
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   scf.for %[[j:.*]] = {{.*}}
-//       CHECKLOOP:     scf.for %[[k:.*]] = {{.*}}
-//       CHECKLOOP:       %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]] : memref<?x?xf32, #[[$strided2D]]>
-//       CHECKLOOP:       %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:       %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:       %[[d:.*]] = mulf %[[a]], %[[b]] : f32
-//       CHECKLOOP:       %[[e:.*]] = addf %[[c]], %[[d]] : f32
-//       CHECKLOOP:       store %[[d]], %{{.*}}[%[[i]], %[[j]], %[[k]]] : memref<?x?x?xf32, #[[$strided3D]]>
-//       CHECKLOOP:       store %[[e]], %{{.*}}[%[[i]], %[[k]], %[[j]]] : memref<?x?x?xf32, #[[$strided3D]]>
+// CHECK-LABEL: @generic_region
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   scf.for %[[j:.*]] = {{.*}}
+//       CHECK:     scf.for %[[k:.*]] = {{.*}}
+//       CHECK:       %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]] : memref<?x?xf32, #[[$strided2D]]>
+//       CHECK:       %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:       %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:       %[[d:.*]] = mulf %[[a]], %[[b]] : f32
+//       CHECK:       %[[e:.*]] = addf %[[c]], %[[d]] : f32
+//       CHECK:       store %[[d]], %{{.*}}[%[[i]], %[[j]], %[[k]]] : memref<?x?x?xf32, #[[$strided3D]]>
+//       CHECK:       store %[[e]], %{{.*}}[%[[i]], %[[k]], %[[j]]] : memref<?x?x?xf32, #[[$strided3D]]>
 
 // CHECKPARALLEL-LABEL: @generic_region
 //       CHECKPARALLEL: scf.parallel (%[[i:[a-zA-Z0-9_]*]], %[[j:[a-zA-Z0-9_]*]], %[[k:[a-zA-Z0-9_]*]])
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]] : memref<?x?xf32, #[[$strided2D]]>
 //       CHECKPARALLEL:   %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]] : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]] : memref<?x?x?xf32, #[[$strided3D]]>
 //       CHECKPARALLEL:   %[[d:.*]] = mulf %[[a]], %[[b]] : f32
 //       CHECKPARALLEL:   %[[e:.*]] = addf %[[c]], %[[d]] : f32
^bb0(%a: f32, %b: f32, %c: f32):
       %ijk_float = sitofp %ijk_int : i32 to f32
 
       %result_2 = addf %c, %ijk_float : f32
       linalg.yield %result_1, %result_2 : f32, f32
   }
   return
 }
 
-// CHECKLOOP-LABEL: @generic_index_region
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   scf.for %[[j:.*]] = {{.*}}
-//       CHECKLOOP:     scf.for %[[k:.*]] = {{.*}}
-//       CHECKLOOP:       %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]]
-//       CHECKLOOP:       %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]]
-//       CHECKLOOP:       %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]]
-//       CHECKLOOP:       %[[result_1:.*]] = mulf %[[a]], %[[b]] : f32
-//       CHECKLOOP:       %[[ij:.*]] = addi %[[i]], %[[j]] : index
-//       CHECKLOOP:       %[[ijk:.*]] = addi %[[ij]], %[[k]] : index
-//       CHECKLOOP:       %[[ijk_int:.*]] = index_cast %[[ijk]] : index to i32
-//       CHECKLOOP:       %[[ijk_float:.*]] = sitofp %[[ijk_int]] : i32 to f32
-//       CHECKLOOP:       %[[result_2:.*]] = addf %[[c]], %[[ijk_float]] : f32
-//       CHECKLOOP:       store %[[result_1]], %{{.*}}[%[[i]], %[[j]], %[[k]]]
-//       CHECKLOOP:       store %[[result_2]], %{{.*}}[%[[i]], %[[k]], %[[j]]]
+// CHECK-LABEL: @generic_index_region
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   scf.for %[[j:.*]] = {{.*}}
+//       CHECK:     scf.for %[[k:.*]] = {{.*}}
+//       CHECK:       %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]]
+//       CHECK:       %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]]
+//       CHECK:       %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]]
+//       CHECK:       %[[result_1:.*]] = mulf %[[a]], %[[b]] : f32
+//       CHECK:       %[[ij:.*]] = addi %[[i]], %[[j]] : index
+//       CHECK:       %[[ijk:.*]] = addi %[[ij]], %[[k]] : index
+//       CHECK:       %[[ijk_int:.*]] = index_cast %[[ijk]] : index to i32
+//       CHECK:       %[[ijk_float:.*]] = sitofp %[[ijk_int]] : i32 to f32
+//       CHECK:       %[[result_2:.*]] = addf %[[c]], %[[ijk_float]] : f32
+//       CHECK:       store %[[result_1]], %{{.*}}[%[[i]], %[[j]], %[[k]]]
+//       CHECK:       store %[[result_2]], %{{.*}}[%[[i]], %[[k]], %[[j]]]
 
 // CHECKPARALLEL-LABEL: @generic_index_region
 //       CHECKPARALLEL: scf.parallel (%[[i:[a-zA-Z0-9_]*]], %[[j:[a-zA-Z0-9_]*]], %[[k:[a-zA-Z0-9_]*]])
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %{{.*}}[%[[i]], %[[j]]]
 //       CHECKPARALLEL:   %[[b:.*]] = memref.load %{{.*}}[%[[i]], %[[j]], %[[k]]]
 //       CHECKPARALLEL:   %[[c:.*]] = memref.load %{{.*}}[%[[i]], %[[k]], %[[j]]]
 //       CHECKPARALLEL:   %[[result_1:.*]] = mulf %[[a]], %[[b]] : f32
 //       CHECKPARALLEL:   %[[ij:.*]] = addi %[[i]], %[[j]] : index
linalg.generic #trait_broadcast
       ins(%arg0 : memref<f32>)
      outs(%arg1 : memref<3x4xf32>) {
     ^bb(%a: f32, %b: f32) :
       linalg.yield %a : f32
   }
   return
 }
 
-// CHECKLOOP-LABEL: @generic_op_zero_rank
-//  CHECKLOOP-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
-//  CHECKLOOP-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xf32>
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   scf.for %[[j:.*]] = {{.*}}
-//       CHECKLOOP:     %[[a:.*]] = memref.load %[[ARG0]][]
-//       CHECKLOOP:     store %[[a]], %[[ARG1]][%[[i]], %[[j]]]
+// CHECK-LABEL: @generic_op_zero_rank
+//  CHECK-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
+//  CHECK-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xf32>
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   scf.for %[[j:.*]] = {{.*}}
+//       CHECK:     %[[a:.*]] = memref.load %[[ARG0]][]
+//       CHECK:     store %[[a]], %[[ARG1]][%[[i]], %[[j]]]
 
 // CHECKPARALLEL-LABEL: @generic_op_zero_rank
 //  CHECKPARALLEL-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
 //  CHECKPARALLEL-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xf32>
 //       CHECKPARALLEL: scf.parallel (%[[i:[a-zA-Z0-9_]*]], %[[j:[a-zA-Z0-9_]*]])
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %[[ARG0]][]
 //       CHECKPARALLEL:   store %[[a]], %[[ARG1]][%[[i]], %[[j]]]
 
 func @generic_index_op_zero_rank(%arg0: memref<i32>, %arg1: memref<3x4xi32>)
 {
   linalg.generic #trait_broadcast
       ins(%arg0 : memref<i32>)
      outs(%arg1 : memref<3x4xi32>) {
     ^bb(%a: i32, %b: i32) :
       %i = linalg.index 0 : index
       %j = linalg.index 1 : index
       %ij = addi %i, %j : index
       %ij_int = index_cast %ij : index to i32
       %result = addi %a, %ij_int : i32
       linalg.yield %result : i32
   }
   return
 }
 
-// CHECKLOOP-LABEL: @generic_index_op_zero_rank
-//  CHECKLOOP-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<i32>
-//  CHECKLOOP-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xi32>
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   scf.for %[[j:.*]] = {{.*}}
-//       CHECKLOOP:     %[[a:.*]] = memref.load %[[ARG0]][
-//       CHECKLOOP:     %[[ij:.*]] = addi %[[i]], %[[j]] : index
-//       CHECKLOOP:     %[[ij_int:.*]] = index_cast %[[ij]] : index to i32
-//       CHECKLOOP:     %[[result:.*]] = addi %[[a]], %[[ij_int]] : i32
-//       CHECKLOOP:     store %[[result]], %[[ARG1]][%[[i]], %[[j]]]
+// CHECK-LABEL: @generic_index_op_zero_rank
+//  CHECK-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<i32>
+//  CHECK-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xi32>
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   scf.for %[[j:.*]] = {{.*}}
+//       CHECK:     %[[a:.*]] = memref.load %[[ARG0]][
+//       CHECK:     %[[ij:.*]] = addi %[[i]], %[[j]] : index
+//       CHECK:     %[[ij_int:.*]] = index_cast %[[ij]] : index to i32
+//       CHECK:     %[[result:.*]] = addi %[[a]], %[[ij_int]] : i32
+//       CHECK:     store %[[result]], %[[ARG1]][%[[i]], %[[j]]]
 
 // CHECKPARALLEL-LABEL: @generic_index_op_zero_rank
 //  CHECKPARALLEL-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<i32>
 //  CHECKPARALLEL-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<3x4xi32>
 //       CHECKPARALLEL: scf.parallel (%[[i:[a-zA-Z0-9_]*]], %[[j:[a-zA-Z0-9_]*]])
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %[[ARG0]][
 //       CHECKPARALLEL:   %[[ij:.*]] = addi %[[i]], %[[j]] : index
 //       CHECKPARALLEL:   %[[ij_int:.*]] = index_cast %[[ij]] : index to i32
linalg.generic #trait_reduce_1D
       ins(%arg0 : memref<?xf32>)
      outs(%arg1 : memref<f32>) {
     ^bb(%a: f32, %b: f32) :
       %0 = addf %a, %b : f32
       linalg.yield %0 : f32
   }
   return
 }
-// CHECKLOOP-LABEL: @generic_op_1D_reduce
-//  CHECKLOOP-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
-//  CHECKLOOP-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
-//       CHECKLOOP:   %[[b:.*]] = memref.load %[[ARG1]][]
-//       CHECKLOOP:   %[[c:.*]] = addf %[[a]], %[[b]] : f32
-//       CHECKLOOP:   store %[[c]], %[[ARG1]][]
+// CHECK-LABEL: @generic_op_1D_reduce
+//  CHECK-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
+//  CHECK-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
+//       CHECK:   %[[b:.*]] = memref.load %[[ARG1]][]
+//       CHECK:   %[[c:.*]] = addf %[[a]], %[[b]] : f32
+//       CHECK:   store %[[c]], %[[ARG1]][]
 
 // CHECKPARALLEL-LABEL: @generic_op_1D_reduce
 //  CHECKPARALLEL-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
 //  CHECKPARALLEL-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
 //       CHECKPARALLEL: scf.for %[[i:.*]] = {{.*}}
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
 //       CHECKPARALLEL:   %[[b:.*]] = memref.load %[[ARG1]][]
 //       CHECKPARALLEL:   %[[c:.*]] = addf %[[a]], %[[b]] : f32
^bb(%a: f32, %b: f32, %c: f32) :
       %0 = constant 0 : index
       %1 = cmpi eq, %0, %i : index
       %2 = select %1, %b, %c : f32
       %3 = addf %a, %2 : f32
       linalg.yield %3 : f32
   }
   return
 }
-// CHECKLOOP-LABEL: @generic_index_op_1D_reduce
-//  CHECKLOOP-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
-//  CHECKLOOP-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
-//  CHECKLOOP-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
-//       CHECKLOOP:   %[[b:.*]] = memref.load %[[ARG1]][]
-//       CHECKLOOP:   %[[c:.*]] = memref.load %[[ARG2]][]
-//       CHECKLOOP:   %[[d:.*]] = select %{{.*}}, %[[b]], %[[c]]
-//       CHECKLOOP:   %[[e:.*]] = addf %[[a]], %[[d]]
-//       CHECKLOOP:   store %[[e]], %[[ARG2]][]
+// CHECK-LABEL: @generic_index_op_1D_reduce
+//  CHECK-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
+//  CHECK-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
+//  CHECK-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
+//       CHECK:   %[[b:.*]] = memref.load %[[ARG1]][]
+//       CHECK:   %[[c:.*]] = memref.load %[[ARG2]][]
+//       CHECK:   %[[d:.*]] = select %{{.*}}, %[[b]], %[[c]]
+//       CHECK:   %[[e:.*]] = addf %[[a]], %[[d]]
+//       CHECK:   store %[[e]], %[[ARG2]][]
 
 // CHECKPARALLEL-LABEL: @generic_index_op_1D_reduce
 //  CHECKPARALLEL-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<?xf32>
 //  CHECKPARALLEL-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
 //  CHECKPARALLEL-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
 //       CHECKPARALLEL: scf.for %[[i:.*]] = {{.*}}
 //       CHECKPARALLEL:   %[[a:.*]] = memref.load %[[ARG0]][%[[i]]]
 //       CHECKPARALLEL:   %[[b:.*]] = memref.load %[[ARG1]][]

 func @generic_const_init(%arg0: memref<?xf32>) {
         %cst = constant 1.0 : f32
   linalg.generic #trait_const_fill outs(%arg0 : memref<?xf32>) {
     ^bb0(%arg1: f32):   // no predecessors
       linalg.yield %cst : f32
     }
     return
 }
-// CHECKLOOP-LABEL: @generic_const_init
-//  CHECKLOOP-SAME: %[[ARG0:.*]]: memref<?xf32>
-//       CHECKLOOP: %[[CONST:.*]] = constant 1.000000e+00 : f32
-//       CHECKLOOP: scf.for %[[i:.*]] = {{.*}}
-//       CHECKLOOP:   store %[[CONST]], %[[ARG0]]
+// CHECK-LABEL: @generic_const_init
+//  CHECK-SAME: %[[ARG0:.*]]: memref<?xf32>
+//       CHECK: %[[CONST:.*]] = constant 1.000000e+00 : f32
+//       CHECK: scf.for %[[i:.*]] = {{.*}}
+//       CHECK:   store %[[CONST]], %[[ARG0]]
 
 // CHECKPARALLEL-LABEL: @generic_const_init
 //  CHECKPARALLEL-SAME: %[[ARG0:.*]]: memref<?xf32>
 //       CHECKPARALLEL: %[[CONST:.*]] = constant 1.000000e+00 : f32
 //       CHECKPARALLEL: scf.parallel (%[[i:.*]])
 //       CHECKPARALLEL:   store %[[CONST]], %[[ARG0]]
 
 #scalar_access = [
%result = scf.if %arg3 -> (f32) {
       scf.yield %a : f32
     } else {
       scf.yield %b : f32
     }
     linalg.yield %result : f32
   }
   return
 }
-// CHECKLOOP-LABEL: @scalar_code
-//  CHECKLOOP-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
-//  CHECKLOOP-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
-//  CHECKLOOP-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
-//   CHECKLOOP-NOT: scf.for
-//       CHECKLOOP: memref.load %[[ARG0]][]
-//       CHECKLOOP: memref.load %[[ARG1]][]
-//       CHECKLOOP: scf.if
-//       CHECKLOOP: scf.yield
-//       CHECKLOOP: else
-//       CHECKLOOP: scf.yield
-//       CHECKLOOP: store %{{.*}}, %[[ARG2]][]
+// CHECK-LABEL: @scalar_code
+//  CHECK-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
+//  CHECK-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
+//  CHECK-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
+//   CHECK-NOT: scf.for
+//       CHECK: memref.load %[[ARG0]][]
+//       CHECK: memref.load %[[ARG1]][]
+//       CHECK: scf.if
+//       CHECK: scf.yield
+//       CHECK: else
+//       CHECK: scf.yield
+//       CHECK: store %{{.*}}, %[[ARG2]][]
 
 // CHECKPARALLEL-LABEL: @scalar_code
 //  CHECKPARALLEL-SAME: %[[ARG0:[a-zA-Z0-9_]*]]: memref<f32>
 //  CHECKPARALLEL-SAME: %[[ARG1:[a-zA-Z0-9_]*]]: memref<f32>
 //  CHECKPARALLEL-SAME: %[[ARG2:[a-zA-Z0-9_]*]]: memref<f32>
 //   CHECKPARALLEL-NOT: scf.for
 //       CHECKPARALLEL: memref.load %[[ARG0]][]
 //       CHECKPARALLEL: memref.load %[[ARG1]][]
 //       CHECKPARALLEL: scf.if
 //       CHECKPARALLEL: scf.yield
 //       CHECKPARALLEL: else
 //       CHECKPARALLEL: scf.yield
 //       CHECKPARALLEL: store %{{.*}}, %[[ARG2]][]
 
 //----------------------------------------------------------------------------//
 // Named ops to loops.
 //----------------------------------------------------------------------------//
 func @named_batch_matmul(%A: memref<?x?x?xf32>, %B: memref<?x?x?xf32>, %C: memref<?x?x?xf32>) {
   linalg.batch_matmul ins(%A, %B : memref<?x?x?xf32>, memref<?x?x?xf32>)
                      outs(%C : memref<?x?x?xf32>)
   return
 }
-// CHECKLOOP-LABEL: @named_batch_matmul
-//  CHECKLOOP-SAME: %[[mA:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//  CHECKLOOP-SAME: %[[mB:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//  CHECKLOOP-SAME: %[[mC:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//       CHECKLOOP: %[[B:.*]] = memref.dim %[[mA]], %c0 : memref<?x?x?xf32>
-//       CHECKLOOP: %[[M:.*]] = memref.dim %[[mA]], %c1 : memref<?x?x?xf32>
-//       CHECKLOOP: %[[K:.*]] = memref.dim %[[mA]], %c2 : memref<?x?x?xf32>
-//       CHECKLOOP: %[[N:.*]] = memref.dim %[[mB]], %c2 : memref<?x?x?xf32>
-//       CHECKLOOP: scf.for %[[b:.*]] = %{{.*}} to %[[B]] step %{{.*}} {
-//       CHECKLOOP:   scf.for %[[m:.*]] = %{{.*}} to %[[M]] step %{{.*}} {
-//       CHECKLOOP:     scf.for %[[n:.*]] = %{{.*}} to %[[N]] step %{{.*}} {
-//       CHECKLOOP:       scf.for %[[k:.*]] = %{{.*}} to %[[K]] step %{{.*}} {
-//       CHECKLOOP:       %[[va:.*]] = memref.load %[[mA]][%[[b]], %[[m]], %[[k]]] : memref<?x?x?xf32>
-//       CHECKLOOP:       %[[vb:.*]] = memref.load %[[mB]][%[[b]], %[[k]], %[[n]]] : memref<?x?x?xf32>
-//       CHECKLOOP:       %[[vc:.*]] = memref.load %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
-//       CHECKLOOP:       %[[inc:.*]] = mulf %[[va]], %[[vb]] : f32
-//       CHECKLOOP:       %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
-//       CHECKLOOP:       store %[[res]], %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
+// CHECK-LABEL: @named_batch_matmul
+//  CHECK-SAME: %[[mA:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//  CHECK-SAME: %[[mB:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//  CHECK-SAME: %[[mC:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//       CHECK: %[[B:.*]] = memref.dim %[[mA]], %c0 : memref<?x?x?xf32>
+//       CHECK: %[[M:.*]] = memref.dim %[[mA]], %c1 : memref<?x?x?xf32>
+//       CHECK: %[[K:.*]] = memref.dim %[[mA]], %c2 : memref<?x?x?xf32>
+//       CHECK: %[[N:.*]] = memref.dim %[[mB]], %c2 : memref<?x?x?xf32>
+//       CHECK: scf.for %[[b:.*]] = %{{.*}} to %[[B]]
+//       CHECK:   scf.for %[[m:.*]] = %{{.*}} to %[[M]]
+//       CHECK:     scf.for %[[n:.*]] = %{{.*}} to %[[N]]
+//       CHECK:       scf.for %[[k:.*]] = %{{.*}} to %[[K]]
+//       CHECK:       %[[va:.*]] = memref.load %[[mA]][%[[b]], %[[m]], %[[k]]] : memref<?x?x?xf32>
+//       CHECK:       %[[vb:.*]] = memref.load %[[mB]][%[[b]], %[[k]], %[[n]]] : memref<?x?x?xf32>
+//       CHECK:       %[[vc:.*]] = memref.load %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
+//       CHECK:       %[[inc:.*]] = mulf %[[va]], %[[vb]] : f32
+//       CHECK:       %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
+//       CHECK:       store %[[res]], %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
 
 // CHECKPARALLEL-LABEL: @named_batch_matmul
 //  CHECKPARALLEL-SAME: %[[mA:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECKPARALLEL-SAME: %[[mB:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECKPARALLEL-SAME: %[[mC:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //       CHECKPARALLEL: %[[B:.*]] = memref.dim %[[mA]], %c0 : memref<?x?x?xf32>
 //       CHECKPARALLEL: %[[M:.*]] = memref.dim %[[mA]], %c1 : memref<?x?x?xf32>
 //       CHECKPARALLEL: %[[K:.*]] = memref.dim %[[mA]], %c2 : memref<?x?x?xf32>
 //       CHECKPARALLEL: %[[N:.*]] = memref.dim %[[mB]], %c2 : memref<?x?x?xf32>
 //       CHECKPARALLEL: scf.parallel (%[[b:.*]], %[[m:.*]], %[[n:.*]]) = ({{.*}}) to (%[[B]], %[[M]], %[[N]]) step ({{.*}}) {
-//       CHECKPARALLEL:   scf.for %[[k:.*]] = %{{.*}} to %[[K]] step %{{.*}} {
+//       CHECKPARALLEL:   scf.for %[[k:.*]] = %{{.*}} to %[[K]]
 //       CHECKPARALLEL:       %[[va:.*]] = memref.load %[[mA]][%[[b]], %[[m]], %[[k]]] : memref<?x?x?xf32>
 //       CHECKPARALLEL:       %[[vb:.*]] = memref.load %[[mB]][%[[b]], %[[k]], %[[n]]] : memref<?x?x?xf32>
 //       CHECKPARALLEL:       %[[vc:.*]] = memref.load %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
 //       CHECKPARALLEL:       %[[inc:.*]] = mulf %[[va]], %[[vb]] : f32
 //       CHECKPARALLEL:       %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
 //       CHECKPARALLEL:       store %[[res]], %[[mC]][%[[b]], %[[m]], %[[n]]] : memref<?x?x?xf32>
 
 
 func @conv1d_no_symbols(%in : memref<?xf32>, %filter : memref<?xf32>, %out : memref<?xf32>) -> () {
   linalg.conv_1d ins(%in, %filter : memref<?xf32>, memref<?xf32>)
                 outs(%out : memref<?xf32>)
   return
 }
 
-// CHECKLOOP-LABEL: @conv1d_no_symbols
-//  CHECKLOOP-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?xf32>
-//  CHECKLOOP-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?xf32>
-//  CHECKLOOP-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?xf32>
-//       CHECKLOOP: %[[c0:.*]] = constant 0 : index
-//       CHECKLOOP: %[[c1:.*]] = constant 1 : index
-//       CHECKLOOP: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?xf32>
-//       CHECKLOOP: %[[dim1:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?xf32>
-//       CHECKLOOP: scf.for %[[b:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
-//       CHECKLOOP:   scf.for %[[m:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
-//       CHECKLOOP:     %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[b]], %[[m]])
-//       CHECKLOOP:     %[[vb:.*]] = memref.load %[[arg0]][%[[aff]]] : memref<?xf32>
-//       CHECKLOOP:     %[[va:.*]] = memref.load %[[arg1]][%[[m]]] : memref<?xf32>
-//       CHECKLOOP:     %[[vc:.*]] = memref.load %[[arg2]][%[[b]]] : memref<?xf32>
-//       CHECKLOOP:     %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
-//       CHECKLOOP:     %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
-//       CHECKLOOP:     store %[[res]], %[[arg2]][%[[b]]] : memref<?xf32>
+// CHECK-LABEL: @conv1d_no_symbols
+//  CHECK-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?xf32>
+//  CHECK-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?xf32>
+//  CHECK-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?xf32>
+//       CHECK: %[[c0:.*]] = constant 0 : index
+//       CHECK: %[[c1:.*]] = constant 1 : index
+//       CHECK: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?xf32>
+//       CHECK: %[[dim1:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?xf32>
+//       CHECK: scf.for %[[b:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
+//       CHECK:   scf.for %[[m:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
+//       CHECK:     %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[b]], %[[m]])
+//       CHECK:     %[[vb:.*]] = memref.load %[[arg0]][%[[aff]]] : memref<?xf32>
+//       CHECK:     %[[va:.*]] = memref.load %[[arg1]][%[[m]]] : memref<?xf32>
+//       CHECK:     %[[vc:.*]] = memref.load %[[arg2]][%[[b]]] : memref<?xf32>
+//       CHECK:     %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
+//       CHECK:     %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
+//       CHECK:     store %[[res]], %[[arg2]][%[[b]]] : memref<?xf32>
 
 // CHECKPARALLEL-LABEL: @conv1d_no_symbols
 //  CHECKPARALLEL-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?xf32>
 //  CHECKPARALLEL-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?xf32>
 //  CHECKPARALLEL-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?xf32>
 //       CHECKPARALLEL: %[[c0:.*]] = constant 0 : index
 //       CHECKPARALLEL: %[[c1:.*]] = constant 1 : index
 //       CHECKPARALLEL: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?xf32>

 //       CHECKPARALLEL:     store %[[res]], %[[arg2]][%[[b]]] : memref<?xf32>
 
 
 func @conv2d_no_symbols(%in : memref<?x?xf32>, %filter : memref<?x?xf32>, %out : memref<?x?xf32>) -> () {
   linalg.conv_2d ins(%in, %filter : memref<?x?xf32>, memref<?x?xf32>)
                 outs(%out: memref<?x?xf32>)
   return
 }
-// CHECKLOOP-LABEL: @conv2d_no_symbols
-//  CHECKLOOP-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?xf32>
-//  CHECKLOOP-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?xf32>
-//  CHECKLOOP-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?xf32>
-//       CHECKLOOP: %[[c0:.*]] = constant 0 : index
-//       CHECKLOOP: %[[c1:.*]] = constant 1 : index
-//       CHECKLOOP: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?x?xf32>
-//       CHECKLOOP: %[[dim1:.*]] = memref.dim %[[arg1]], %[[c1]] : memref<?x?xf32>
-//       CHECKLOOP: %[[dim2:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?x?xf32>
-//       CHECKLOOP: %[[dim3:.*]] = memref.dim %[[arg2]], %[[c1]] : memref<?x?xf32>
-//       CHECKLOOP: scf.for %[[arg3:.*]] = %[[c0]] to %[[dim2]] step %[[c1]] {
-//       CHECKLOOP:   scf.for %[[arg4:.*]] = %[[c0]] to %[[dim3]] step %[[c1]] {
-//       CHECKLOOP:     scf.for %[[arg5:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
-//       CHECKLOOP:       scf.for %[[arg6:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
-//       CHECKLOOP:         %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg3]], %[[arg5]])
-//       CHECKLOOP:         %[[aff2:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg4]], %[[arg6]])
-//       CHECKLOOP:         %[[vb:.*]] = memref.load %[[arg0]][%[[aff]], %[[aff2]]] : memref<?x?xf32>
-
-//       CHECKLOOP:         %[[va:.*]] = memref.load %[[arg1]][%[[arg5]], %[[arg6]]] : memref<?x?xf32>
-//       CHECKLOOP:         %[[vc:.*]] = memref.load %[[arg2]][%[[arg3]], %[[arg4]]] : memref<?x?xf32>
-
-//       CHECKLOOP:         %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
-//       CHECKLOOP:         %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
-//       CHECKLOOP:         store %[[res]], %[[arg2]][%[[arg3]], %[[arg4]]] : memref<?x?xf32>
+// CHECK-LABEL: @conv2d_no_symbols
+//  CHECK-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?xf32>
+//  CHECK-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?xf32>
+//  CHECK-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?xf32>
+//       CHECK: %[[c0:.*]] = constant 0 : index
+//       CHECK: %[[c1:.*]] = constant 1 : index
+//       CHECK: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?x?xf32>
+//       CHECK: %[[dim1:.*]] = memref.dim %[[arg1]], %[[c1]] : memref<?x?xf32>
+//       CHECK: %[[dim2:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?x?xf32>
+//       CHECK: %[[dim3:.*]] = memref.dim %[[arg2]], %[[c1]] : memref<?x?xf32>
+//       CHECK: scf.for %[[arg3:.*]] = %[[c0]] to %[[dim2]] step %[[c1]] {
+//       CHECK:   scf.for %[[arg4:.*]] = %[[c0]] to %[[dim3]] step %[[c1]] {
+//       CHECK:     scf.for %[[arg5:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
+//       CHECK:       scf.for %[[arg6:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
+//       CHECK:         %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg3]], %[[arg5]])
+//       CHECK:         %[[aff2:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg4]], %[[arg6]])
+//       CHECK:         %[[vb:.*]] = memref.load %[[arg0]][%[[aff]], %[[aff2]]] : memref<?x?xf32>
+
+//       CHECK:         %[[va:.*]] = memref.load %[[arg1]][%[[arg5]], %[[arg6]]] : memref<?x?xf32>
+//       CHECK:         %[[vc:.*]] = memref.load %[[arg2]][%[[arg3]], %[[arg4]]] : memref<?x?xf32>
+
+//       CHECK:         %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
+//       CHECK:         %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
+//       CHECK:         store %[[res]], %[[arg2]][%[[arg3]], %[[arg4]]] : memref<?x?xf32>
 
 // CHECKPARALLEL-LABEL: @conv2d_no_symbols
 //  CHECKPARALLEL-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?xf32>
 //  CHECKPARALLEL-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?xf32>
 //  CHECKPARALLEL-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?xf32>
 //       CHECKPARALLEL: %[[c0:.*]] = constant 0 : index
 //       CHECKPARALLEL: %[[c1:.*]] = constant 1 : index
 //       CHECKPARALLEL: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?x?xf32>

 
 
 func @conv3d_no_symbols(%in : memref<?x?x?xf32>, %filter : memref<?x?x?xf32>, %out : memref<?x?x?xf32>) -> () {
   linalg.conv_3d ins(%in, %filter : memref<?x?x?xf32>, memref<?x?x?xf32>)
                 outs(%out : memref<?x?x?xf32>)
   return
 }
 
-// CHECKLOOP-LABEL: @conv3d_no_symbols
-//  CHECKLOOP-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//  CHECKLOOP-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//  CHECKLOOP-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
-//       CHECKLOOP: %[[c2:.*]] = constant 2 : index
-//       CHECKLOOP: %[[c0:.*]] = constant 0 : index
-//       CHECKLOOP: %[[c1:.*]] = constant 1 : index
-//       CHECKLOOP: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?x?x?xf32>
-//       CHECKLOOP: %[[dim1:.*]] = memref.dim %[[arg1]], %[[c1]] : memref<?x?x?xf32>
-//       CHECKLOOP: %[[dim2:.*]] = memref.dim %[[arg1]], %[[c2]] : memref<?x?x?xf32>
-//       CHECKLOOP: %[[dim3:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?x?x?xf32>
-//       CHECKLOOP: %[[dim4:.*]] = memref.dim %[[arg2]], %[[c1]] : memref<?x?x?xf32>
-//       CHECKLOOP: %[[dim5:.*]] = memref.dim %[[arg2]], %[[c2]] : memref<?x?x?xf32>
-//       CHECKLOOP: scf.for %[[arg3:.*]] = %[[c0]] to %[[dim3]] step %[[c1]] {
-//       CHECKLOOP:   scf.for %[[arg4:.*]] = %[[c0]] to %[[dim4]] step %[[c1]] {
-//       CHECKLOOP:     scf.for %[[arg5:.*]] = %[[c0]] to %[[dim5]] step %[[c1]] {
-//       CHECKLOOP:       scf.for %[[arg6:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
-//       CHECKLOOP:         scf.for %[[arg7:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
-//       CHECKLOOP:           scf.for %[[arg8:.*]] = %[[c0]] to %[[dim2]] step %[[c1]] {
-//       CHECKLOOP:             %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg3]], %[[arg6]])
-//       CHECKLOOP:             %[[aff2:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg4]], %[[arg7]])
-//       CHECKLOOP:             %[[aff3:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg5]], %[[arg8]])
-//       CHECKLOOP:             %[[vb:.*]] = memref.load %[[arg0]][%[[aff]], %[[aff2]], %[[aff3]]] : memref<?x?x?xf32>
-
-//       CHECKLOOP:             %[[va:.*]] = memref.load %[[arg1]][%[[arg6]], %[[arg7]], %[[arg8]]] : memref<?x?x?xf32>
-//       CHECKLOOP:             %[[vc:.*]] = memref.load %[[arg2]][%[[arg3]], %[[arg4]], %[[arg5]]] : memref<?x?x?xf32>
-
-//       CHECKLOOP:             %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
-//       CHECKLOOP:             %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
-//       CHECKLOOP:             store %[[res]], %[[arg2]][%[[arg3]], %[[arg4]], %[[arg5]]] : memref<?x?x?xf32>
+// CHECK-LABEL: @conv3d_no_symbols
+//  CHECK-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//  CHECK-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//  CHECK-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
+//       CHECK: %[[c2:.*]] = constant 2 : index
+//       CHECK: %[[c0:.*]] = constant 0 : index
+//       CHECK: %[[c1:.*]] = constant 1 : index
+//       CHECK: %[[dim0:.*]] = memref.dim %[[arg1]], %[[c0]] : memref<?x?x?xf32>
+//       CHECK: %[[dim1:.*]] = memref.dim %[[arg1]], %[[c1]] : memref<?x?x?xf32>
+//       CHECK: %[[dim2:.*]] = memref.dim %[[arg1]], %[[c2]] : memref<?x?x?xf32>
+//       CHECK: %[[dim3:.*]] = memref.dim %[[arg2]], %[[c0]] : memref<?x?x?xf32>
+//       CHECK: %[[dim4:.*]] = memref.dim %[[arg2]], %[[c1]] : memref<?x?x?xf32>
+//       CHECK: %[[dim5:.*]] = memref.dim %[[arg2]], %[[c2]] : memref<?x?x?xf32>
+//       CHECK: scf.for %[[arg3:.*]] = %[[c0]] to %[[dim3]] step %[[c1]] {
+//       CHECK:   scf.for %[[arg4:.*]] = %[[c0]] to %[[dim4]] step %[[c1]] {
+//       CHECK:     scf.for %[[arg5:.*]] = %[[c0]] to %[[dim5]] step %[[c1]] {
+//       CHECK:       scf.for %[[arg6:.*]] = %[[c0]] to %[[dim0]] step %[[c1]] {
+//       CHECK:         scf.for %[[arg7:.*]] = %[[c0]] to %[[dim1]] step %[[c1]] {
+//       CHECK:           scf.for %[[arg8:.*]] = %[[c0]] to %[[dim2]] step %[[c1]] {
+//       CHECK:             %[[aff:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg3]], %[[arg6]])
+//       CHECK:             %[[aff2:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg4]], %[[arg7]])
+//       CHECK:             %[[aff3:.*]] = affine.apply #[[$stride1Dilation1]](%[[arg5]], %[[arg8]])
+//       CHECK:             %[[vb:.*]] = memref.load %[[arg0]][%[[aff]], %[[aff2]], %[[aff3]]] : memref<?x?x?xf32>
+
+//       CHECK:             %[[va:.*]] = memref.load %[[arg1]][%[[arg6]], %[[arg7]], %[[arg8]]] : memref<?x?x?xf32>
+//       CHECK:             %[[vc:.*]] = memref.load %[[arg2]][%[[arg3]], %[[arg4]], %[[arg5]]] : memref<?x?x?xf32>
+
+//       CHECK:             %[[inc:.*]] = mulf %[[vb]], %[[va]] : f32
+//       CHECK:             %[[res:.*]] = addf %[[vc]], %[[inc]] : f32
+//       CHECK:             store %[[res]], %[[arg2]][%[[arg3]], %[[arg4]], %[[arg5]]] : memref<?x?x?xf32>
 
 // CHECKPARALLEL-LABEL: @conv3d_no_symbols
 //  CHECKPARALLEL-SAME: %[[arg0:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECKPARALLEL-SAME: %[[arg1:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //  CHECKPARALLEL-SAME: %[[arg2:[a-zA-Z0-9]+]]: memref<?x?x?xf32>
 //       CHECKPARALLEL: %[[c2:.*]] = constant 2 : index
 //       CHECKPARALLEL: %[[c0:.*]] = constant 0 : index
 //       CHECKPARALLEL: %[[c1:.*]] = constant 1 : index