[mlir] Move memref.[tensor_load|buffer_cast|clone] to "bufferization" dialect.

[pifon2a]

Summary

https://llvm.discourse.group/t/rfc-dialect-for-bufferization-related-ops/4712

[pifon2a]

Clean-up

[mehdi_amini]

mlir/include/mlir/Dialect/Bufferization/IR/BufferizationBase.td

21

This is what will show up on the website, I think we need a more extensive doc here, to set up the context and also linking to the other existing docs.

[springerm]

mlir/include/mlir/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h

386 ↗

(On Diff #389572)

This should probably not be part of this change.

[herhut]

Thank you for cleaning this up. Just some nits and please address the comments others already had.

I would be OK with improving documentation in a follow up, just to avoid merge conflicts and rebase.

mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp

186

This comment is now off.

269

This is not dim of cast but dim of ToMemref.

mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ComprehensiveBufferize.cpp

939–940

nit: variable name

949–950

nit: variable name.

[pifon2a]

mlir/include/mlir/Dialect/Bufferization/IR/BufferizationBase.td

21

Good point, also updated the names of the ops in Bufferization.md and BufferDeallocationInternals.md

[pifon2a]

address the comments.

[aartbik]

LGTM for the sparse part (there are some dangling references in the doc now to "memref" but since this is submitted, I will fix that in a follow up).

[pifon2a]

In D114552#3158806, @aartbik wrote:

LGTM for the sparse part (there are some dangling references in the doc now to "memref" but since this is submitted, I will fix that in a follow up).

Thank you, Aart!

mlir/docs/BufferDeallocationInternals.md

 result:
 
 ```mlir
 func @branch(%arg0: i1) {
   %0 = memref.alloc() : memref<2xf32>
   cond_br %arg0, ^bb1, ^bb2
 ^bb1:
   %1 = memref.alloc() : memref<2xf32>
-  %3 = memref.clone %1 : (memref<2xf32>) -> (memref<2xf32>)
+  %3 = bufferization.clone %1 : (memref<2xf32>) -> (memref<2xf32>)
   memref.dealloc %1 : memref<2xf32> // %1 can be safely freed here
   br ^bb3(%3 : memref<2xf32>)
 ^bb2:
   use(%0)
-  %4 = memref.clone %0 : (memref<2xf32>) -> (memref<2xf32>)
+  %4 = bufferization.clone %0 : (memref<2xf32>) -> (memref<2xf32>)
   br ^bb3(%4 : memref<2xf32>)
 ^bb3(%2: memref<2xf32>):
   …
   memref.dealloc %2 : memref<2xf32> // free temp buffer %2
   memref.dealloc %0 : memref<2xf32> // %0 can be safely freed here
   return
 }
 ```

 func @condBranchDynamicTypeNested(
   %arg0: i1,
   %arg1: memref<?xf32>,
   %arg2: memref<?xf32>,
   %arg3: index) {
   cond_br %arg0, ^bb1, ^bb2(%arg3 : index)
 ^bb1:
   // temp buffer required due to alias %3
-  %5 = memref.clone %arg1 : (memref<?xf32>) -> (memref<?xf32>)
+  %5 = bufferization.clone %arg1 : (memref<?xf32>) -> (memref<?xf32>)
   br ^bb6(%5 : memref<?xf32>)
 ^bb2(%0: index):
   %1 = memref.alloc(%0) : memref<?xf32>
   use(%1)
   cond_br %arg0, ^bb3, ^bb4
 ^bb3:
   br ^bb5(%1 : memref<?xf32>)
 ^bb4:
   br ^bb5(%1 : memref<?xf32>)
 ^bb5(%2: memref<?xf32>):
-  %6 = memref.clone %1 : (memref<?xf32>) -> (memref<?xf32>)
+  %6 = bufferization.clone %1 : (memref<?xf32>) -> (memref<?xf32>)
   memref.dealloc %1 : memref<?xf32>
   br ^bb6(%6 : memref<?xf32>)
 ^bb6(%3: memref<?xf32>):
   br ^bb7(%3 : memref<?xf32>)
 ^bb7(%4: memref<?xf32>):
   test.copy(%4, %arg2) : (memref<?xf32>, memref<?xf32>) -> ()
   memref.dealloc %3 : memref<?xf32>  // free %3, since %4 is a non-crit. alias of %3
   return

 
 ```mlir
 func @inner_region_control_flow_div(
   %arg0 : index,
   %arg1 : index) -> memref<?x?xf32> {
   %0 = memref.alloc(%arg0, %arg0) : memref<?x?xf32>
   %1 = custom.region_if %0 : memref<?x?xf32> -> (memref<?x?xf32>)
    then(%arg2 : memref<?x?xf32>) {
-    %4 = memref.clone %arg2 : (memref<?x?xf32>) -> (memref<?x?xf32>)
+    %4 = bufferization.clone %arg2 : (memref<?x?xf32>) -> (memref<?x?xf32>)
     custom.region_if_yield %4 : memref<?x?xf32>
    } else(%arg3 : memref<?x?xf32>) {
     %2 = memref.alloc(%arg0, %arg1) : memref<?x?xf32>
-    %5 = memref.clone %2 : (memref<?x?xf32>) -> (memref<?x?xf32>)
+    %5 = bufferization.clone %2 : (memref<?x?xf32>) -> (memref<?x?xf32>)
     memref.dealloc %2 : memref<?x?xf32>
     custom.region_if_yield %5 : memref<?x?xf32>
    } join(%arg4: memref<?x?xf32>) {
-    %4 = memref.clone %arg4 : (memref<?x?xf32>) -> (memref<?x?xf32>)
+    %4 = bufferization.clone %arg4 : (memref<?x?xf32>) -> (memref<?x?xf32>)
     memref.dealloc %arg4 : memref<?x?xf32>
     custom.region_if_yield %4 : memref<?x?xf32>
    }
   memref.dealloc %0 : memref<?x?xf32>  // %0 can be safely freed here
   return %1 : memref<?x?xf32>
 }
 ```
 

 
 ```mlir
 func @loop_nested_if(
   %lb: index,
   %ub: index,
   %step: index,
   %buf: memref<2xf32>,
   %res: memref<2xf32>) {
-  %4 = memref.clone %buf : (memref<2xf32>) -> (memref<2xf32>)
+  %4 = bufferization.clone %buf : (memref<2xf32>) -> (memref<2xf32>)
   %0 = scf.for %i = %lb to %ub step %step
     iter_args(%iterBuf = %4) -> memref<2xf32> {
     %1 = arith.cmpi "eq", %i, %ub : index
     %2 = scf.if %1 -> (memref<2xf32>) {
       %3 = memref.alloc() : memref<2xf32> // makes %2 a critical alias
       use(%3)
-      %5 = memref.clone %3 : (memref<2xf32>) -> (memref<2xf32>)
+      %5 = bufferization.clone %3 : (memref<2xf32>) -> (memref<2xf32>)
       memref.dealloc %3 : memref<2xf32>
       scf.yield %5 : memref<2xf32>
     } else {
-      %6 = memref.clone %iterBuf : (memref<2xf32>) -> (memref<2xf32>)
+      %6 = bufferization.clone %iterBuf : (memref<2xf32>) -> (memref<2xf32>)
       scf.yield %6 : memref<2xf32>
     }
-    %7 = memref.clone %2 : (memref<2xf32>) -> (memref<2xf32>)
+    %7 = bufferization.clone %2 : (memref<2xf32>) -> (memref<2xf32>)
     memref.dealloc %2 : memref<2xf32>
     memref.dealloc %iterBuf : memref<2xf32> // free backedge iteration variable
     scf.yield %7 : memref<2xf32>
   }
   test.copy(%0, %res) : (memref<2xf32>, memref<2xf32>) -> ()
   memref.dealloc %0 : memref<2xf32> // free temp copy %0
   return
 }

 this clone operation are also removed. Here is a working example generated by
 the BufferDeallocation pass that allocates a buffer with dynamic size. A deeper
 analysis of this sample reveals that the highlighted operations are redundant
 and can be removed.
 
 ```mlir
 func @dynamic_allocation(%arg0: index, %arg1: index) -> memref<?x?xf32> {
   %1 = memref.alloc(%arg0, %arg1) : memref<?x?xf32>
-  %2 = memref.clone %1 : (memref<?x?xf32>) -> (memref<?x?xf32>)
+  %2 = bufferization.clone %1 : (memref<?x?xf32>) -> (memref<?x?xf32>)
   memref.dealloc %1 : memref<?x?xf32>
   return %2 : memref<?x?xf32>
 }
 ```
 
 Will be transformed to:
 
 ```mlir
test.generic {
     args_in = 1 : i64,
     args_out = 1 : i64,
     indexing_maps = [#map0, #map0],
     iterator_types = ["parallel"]} %arg0, %temp {
   ^bb0(%gen2_arg0: f32, %gen2_arg1: f32):
     %tmp2 = math.exp %gen2_arg0 : f32
     test.yield %tmp2 : f32
   }: memref<2xf32>, memref<2xf32>
-  %result = memref.clone %temp : (memref<2xf32>) -> (memref<2xf32>)
+  %result = bufferization.clone %temp : (memref<2xf32>) -> (memref<2xf32>)
   memref.dealloc %temp : memref<2xf32>
   return
 }
 ```
 
 Will be transformed to:
 
 ```mlir
func @reuseTarget(%arg0: memref<2xf32>, %result: memref<2xf32>){
   }: memref<2xf32>, memref<2xf32>
   return
 }
 ```
 
 ## Known Limitations
 
 BufferDeallocation introduces additional clones from “memref” dialect
-(“memref.clone”). Analogous, all deallocations use the “memref” dialect-free
-operation “memref.dealloc”. The actual copy process is realized using
-“test.copy”. Furthermore, buffers are essentially immutable after their creation
-in a block. Another limitations are known in the case using unstructered control
-flow.
+(“bufferization.clone”). Analogous, all deallocations use the “memref”
+dialect-free operation “memref.dealloc”. The actual copy process is realized
+using “test.copy”. Furthermore, buffers are essentially immutable after their
+creation in a block. Another limitations are known in the case using
+unstructered control flow.

mlir/docs/Bufferization.md

 `populateTensorBufferizePatterns` is separated, so that power users can use the
 patterns independently, if necessary (such as to combine multiple sets of
 conversion patterns into a single conversion call, for performance).
 
 One convenient utility provided by the MLIR bufferization infrastructure is the
 `BufferizeTypeConverter`, which comes pre-loaded with the necessary conversions
 and materializations between `tensor` and `memref`.
 
-In this case, the `MemRefOpsDialect` is marked as legal, so the
-`memref.tensor_load` and `memref.buffer_cast` ops, which are inserted
+In this case, the `BufferizationOpsDialect` is marked as legal, so the
+`bufferization.to_tensor` and `bufferization.to_memref` ops, which are inserted
 automatically by the dialect conversion framework as materializations, are
 legal. There is a helper `populateBufferizeMaterializationLegality`
 ([code](https://github.com/llvm/llvm-project/blob/a0b65a7bcd6065688189b3d678c42ed6af9603db/mlir/include/mlir/Transforms/Bufferize.h#L53))
 which helps with this in general.
 
 ### Other partial bufferization examples
 
 -   `linalg-bufferize`

 
 ## How to write a finalizing bufferization pass
 
 The contract of a finalizing bufferization pass is that all tensors are gone
 from the program.
 
 The easiest way to write a finalizing bufferize pass is to not write one at all!
 MLIR provides a pass `finalizing-bufferize` which eliminates the
-`memref.tensor_load` / `memref.buffer_cast` materialization ops inserted by
-partial bufferization passes and emits an error if that is not sufficient to
-remove all tensors from the program.
+`bufferization.to_tensor` / `bufferization.to_memref` materialization ops
+inserted by partial bufferization passes and emits an error if that is not
+sufficient to remove all tensors from the program.
 
 This pass is sufficient when partial bufferization passes have bufferized all
 the ops in the program, leaving behind only the materializations. When possible,
 it is recommended to structure your pass pipeline this way, as this has the
 significant advantage that if an op does not get bufferized (due to a missing
 pattern, bug in the code, etc.), `finalizing-bufferize` will emit a nice clean
 error, and the IR seen by `finalizing-bufferize` will only contain only one
 unbufferized op.
 
 However, before the current bufferization infrastructure was put in place,
 bufferization could only be done as a single finalizing bufferization mega-pass
 that used the `populate*BufferizePatterns` functions from multiple dialects to
 simultaneously bufferize everything at once. Thus, one might see code in
 downstream projects structured this way. This structure is not recommended in
 new code. A helper, `populateEliminateBufferizeMaterializationsPatterns`
 ([code](https://github.com/llvm/llvm-project/blob/a0b65a7bcd6065688189b3d678c42ed6af9603db/mlir/include/mlir/Transforms/Bufferize.h#L58))
 is available for such passes to provide patterns that eliminate
-`memref.tensor_load` and `memref.buffer_cast`.
+`bufferization.to_tensor` and `bufferization.to_memref`.
 
 ## Changes since [the talk](#the-talk)
 
 -   `func-bufferize` was changed to be a partial conversion pass, and there is a
     new `finalizing-bufferize` which serves as a general finalizing
     bufferization pass.

mlir/include/mlir/Dialect/Arithmetic/Transforms/Passes.td

 //===-- Passes.td - Arithmetic pass definition file --------*- tablegen -*-===//
 //
 // 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_ARITHMETIC_TRANSFORMS_PASSES
 #define MLIR_DIALECT_ARITHMETIC_TRANSFORMS_PASSES
 
 include "mlir/Pass/PassBase.td"
 
 def ArithmeticBufferize : FunctionPass<"arith-bufferize"> {
   let summary = "Bufferize Arithmetic dialect ops.";
   let constructor = "mlir::arith::createArithmeticBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect"];
 }
 
 def ArithmeticExpandOps : FunctionPass<"arith-expand"> {
   let summary = "Legalize Arithmetic ops to be convertible to LLVM.";
   let constructor = "mlir::arith::createArithmeticExpandOpsPass()";
   let dependentDialects = ["StandardOpsDialect"];
 }
 
 #endif // MLIR_DIALECT_ARITHMETIC_TRANSFORMS_PASSES

mlir/include/mlir/Dialect/Bufferization/IR/AllocationOpInterface.h

 //===- AllocationOpInterface.h - Allocation op interface ------------------===//
 //
 // 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 the operation interface for allocation ops.
 //
 //===----------------------------------------------------------------------===//
 
 #ifndef MLIR_DIALECT_BUFFERIZATION_IR_ALLOCATIONOPINTERFACE_H_
 #define MLIR_DIALECT_BUFFERIZATION_IR_ALLOCATIONOPINTERFACE_H_
 
-#include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/IR/Builders.h"
 
 #include "mlir/Dialect/Bufferization/IR/AllocationOpInterface.h.inc"
 
 #endif // MLIR_DIALECT_BUFFERIZATION_IR_ALLOCATIONOPINTERFACE_H_

mlir/include/mlir/Dialect/Bufferization/IR/AllocationOpInterface.td

StaticInterfaceMethod<[{
         Builds a clone operation using the provided builder and the current
         allocation value (which refers to the current Op implementing this
         interface). The allocation value is a result of the current operation
         implementing this interface. If there is no compatible clone operation,
         this method can return ::llvm::None.
       }],
       "::mlir::Optional<::mlir::Value>", "buildClone",
       (ins "::mlir::OpBuilder&":$builder, "::mlir::Value":$alloc), [{}],
-      /*defaultImplementation=*/[{
-          return builder.create<memref::CloneOp>(alloc.getLoc(), alloc)
-            .getResult();
-      }]
+      /*defaultImplementation=*/[{ return llvm::None; }]
     >
   ];
 }
 
 #endif  // ALLOCATION_OP_INTERFACE

mlir/include/mlir/Dialect/Bufferization/IR/Bufferization.h

+//===- Bufferization.h - Bufferization dialect ------------------*- 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_BUFFERIZATION_IR_BUFFERIZATION_H_
+#define MLIR_DIALECT_BUFFERIZATION_IR_BUFFERIZATION_H_
+
+#include "mlir/Dialect/Bufferization/IR/AllocationOpInterface.h"
+#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Tensor/IR/Tensor.h"
+
+//===----------------------------------------------------------------------===//
+// Bufferization Dialect
+//===----------------------------------------------------------------------===//
+
+#include "mlir/Dialect/Bufferization/IR/BufferizationOpsDialect.h.inc"
+
+//===----------------------------------------------------------------------===//
+// Bufferization Dialect Operations
+//===----------------------------------------------------------------------===//
+
+#define GET_OP_CLASSES
+#include "mlir/Dialect/Bufferization/IR/BufferizationOps.h.inc"
+
+#endif // MLIR_DIALECT_BUFFERIZATION_IR_BUFFERIZATION_H_

mlir/include/mlir/Dialect/Bufferization/IR/BufferizationBase.td

+//===- BufferizationBase.td - Bufferization dialect base ---*- tablegen -*-===//
+//
+// 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 BUFFERIZATION_BASE
+#define BUFFERIZATION_BASE
+
+include "mlir/IR/OpBase.td"
+
+def Bufferization_Dialect : Dialect {
+  let name = "bufferization";
+  let cppNamespace = "::mlir::bufferization";
+  let description = [{
+    Bufferization in MLIR is the process of converting the `tensor` type to the
+    `memref` type.
+    The `bufferization` dialect is intended to collect operations/interfaces
+    specific to the bufferization passes.
+
+    Overview of the bufferization infrastructure and important conceptual
+    details related to using the MLIR dialect conversion infrastructure can be
+    found in [bufferization](Bufferization.md) and [buffer
+    deallocation](BufferDeallocationInternals.md).
+  }];
+  let dependentDialects = ["memref::MemRefDialect", "tensor::TensorDialect"];
+}
+
+#endif // BUFFERIZATION_BASE

mlir/include/mlir/Dialect/Bufferization/IR/BufferizationOps.td

+//===- BufferizationOps.td - Bufferization op definitions ----------*- tablegen -*-===//
+//
+// 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 BUFFERIZATION_OPS
+#define BUFFERIZATION_OPS
+
+include "mlir/Dialect/Bufferization/IR/AllocationOpInterface.td"
+include "mlir/Dialect/Bufferization/IR/BufferizationBase.td"
+include "mlir/Interfaces/SideEffectInterfaces.td"
+include "mlir/Interfaces/CopyOpInterface.td"
+
+class Bufferization_Op<string mnemonic, list<OpTrait> traits = []>
+    : Op<Bufferization_Dialect, mnemonic, traits>;
+
+//===----------------------------------------------------------------------===//
+// CloneOp
+//===----------------------------------------------------------------------===//
+
+def Bufferization_CloneOp : Bufferization_Op<"clone", [
+    CopyOpInterface,
+    DeclareOpInterfaceMethods<MemoryEffectsOpInterface>,
+    DeclareOpInterfaceMethods<AllocationOpInterface, ["buildDealloc", "buildClone"]>
+  ]> {
+  let builders = [
+    OpBuilder<(ins "Value":$value), [{
+      return build($_builder, $_state, value.getType(), value);
+    }]>];
+
+  let description = [{
+    Clones the data in the input view into an implicitly defined output view.
+
+    Usage:
+
+    ```mlir
+    %arg1 = bufferization.clone %arg0 : memref<?xf32> to memref<?xf32>
+    ```
+
+    Valid implementations of this operation may alias the input and output
+    views or create an actual copy. Mutating the source or result
+    of the clone operation after the clone operation thus leads to undefined
+    behavior.
+  }];
+
+  let arguments = (ins Arg<AnyRankedOrUnrankedMemRef, "", []>:$input);
+  let results = (outs Arg<AnyRankedOrUnrankedMemRef, "", []>:$output);
+
+  let extraClassDeclaration = [{
+    Value getSource() { return input(); }
+    Value getTarget() { return output(); }
+  }];
+
+  let assemblyFormat = "$input attr-dict `:` type($input) `to` type($output)";
+
+  let hasFolder = 1;
+  let verifier = ?;
+  let hasCanonicalizer = 1;
+}
+//===----------------------------------------------------------------------===//
+// ToTensorOp
+//===----------------------------------------------------------------------===//
+
+def Bufferization_ToTensorOp : Bufferization_Op<"to_tensor",
+    [SameOperandsAndResultShape, SameOperandsAndResultElementType,
+     TypesMatchWith<"result type matches tensor equivalent of 'memref'",
+                    "memref", "result",
+                    "memref::getTensorTypeFromMemRefType($_self)">]> {
+  let summary = "memref to tensor operation";
+  let description = [{
+    Create a tensor from a memref, making an independent copy of the element
+    data. The result value is a tensor whose shape and element type match the
+    memref operand.
+
+    The opposite of this op is to_memref. Together, these two ops are
+    useful for source/target materializations when doing type conversions
+    involving tensors and memrefs.
+
+    Example:
+
+    ```mlir
+    // Produces a value of tensor<4x?xf32> type.
+    %12 = bufferization.to_tensor %10 : memref<4x?xf32, #layout, memspace0>
+    ```
+
+    If tensor load is used in the bufferization steps, mutating the source
+    buffer after loading leads to undefined behavior.
+  }];
+
+  let arguments = (ins Arg<AnyRankedOrUnrankedMemRef,
+                       "the reference to load from", [MemRead]>:$memref);
+  let results = (outs AnyTensor:$result);
+  // MemrefToTensor is fully verified by traits.
+  let verifier = ?;
+
+  let builders = [
+    OpBuilder<(ins "Value":$memref), [{
+      $_state.addOperands(memref);
+      $_state.addTypes(memref::getTensorTypeFromMemRefType(memref.getType()));
+    }]>];
+
+  let extraClassDeclaration = [{
+    /// The result of a to_tensor is always a tensor.
+    TensorType getType() {
+      Type resultType = getResult().getType();
+      if (resultType.isa<TensorType>())
+        return resultType.cast<TensorType>();
+      return {};
+    }
+  }];
+
+  let assemblyFormat = "$memref attr-dict `:` type($memref)";
+
+  let hasCanonicalizer = 1;
+  let hasFolder = 1;
+}
+
+
+//===----------------------------------------------------------------------===//
+// ToMemrefOp
+//===----------------------------------------------------------------------===//
+
+def Bufferization_ToMemrefOp : Bufferization_Op<"to_memref",
+    [SameOperandsAndResultShape, SameOperandsAndResultElementType, NoSideEffect,
+     TypesMatchWith<"type of 'tensor' is the tensor equivalent of 'memref'",
+                    "memref", "tensor",
+                    "memref::getTensorTypeFromMemRefType($_self)">]> {
+  let summary = "tensor to memref cast operation";
+  let description = [{
+    Casts a tensor to a memref.
+
+    ```mlir
+    // Result type is tensor<4x?xf32>
+    %12 = bufferization.to_memref %10 : memref<4x?xf32, #map0, 42>
+    ```
+
+    Note, that mutating the result of the to_buffer operation leads to
+    undefined behavior.
+
+    This operation is a specialized variant of the built-in
+    unrealized_conversion_cast and is intended for use in the context of
+    gradual bufferization.
+  }];
+
+  let arguments = (ins AnyTensor:$tensor);
+  let results = (outs AnyRankedOrUnrankedMemRef:$memref);
+  // This op is fully verified by traits.
+  let verifier = ?;
+
+  let assemblyFormat = "$tensor attr-dict `:` type($memref)";
+
+  let hasFolder = 1;
+  let hasCanonicalizer = 1;
+}
+
+#endif // BUFFERIZATION_OPS

mlir/include/mlir/Dialect/Bufferization/IR/CMakeLists.txt

+add_mlir_dialect(BufferizationOps bufferization)
+add_mlir_doc(BufferizationOps BufferizationOps Dialects/ -gen-dialect-doc)
 add_mlir_interface(AllocationOpInterface)

mlir/include/mlir/Dialect/Linalg/Passes.td

let dependentDialects = [
     "scf::SCFDialect"
   ];
 }
 
 def LinalgBufferize : Pass<"linalg-bufferize", "FuncOp"> {
   let summary = "Bufferize the linalg dialect";
   let constructor = "mlir::createLinalgBufferizePass()";
   let dependentDialects = [
-    "linalg::LinalgDialect",
     "AffineDialect",
+    "bufferization::BufferizationDialect",
+    "linalg::LinalgDialect",
     "memref::MemRefDialect",
   ];
 }
 
 def LinalgPromotion : FunctionPass<"linalg-promote-subviews"> {
   let summary = "Promote subview ops to local buffers";
   let constructor = "mlir::createLinalgPromotionPass()";
   let options = [

mlir/include/mlir/Dialect/MemRef/IR/MemRef.h

 //===- MemRef.h - MemRef dialect --------------------------------*- 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_MEMREF_IR_MEMREF_H_
 #define MLIR_DIALECT_MEMREF_IR_MEMREF_H_
 
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
-#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Utils/ReshapeOpsUtils.h"
 #include "mlir/IR/Dialect.h"
 #include "mlir/Interfaces/CallInterfaces.h"
 #include "mlir/Interfaces/CastInterfaces.h"
 #include "mlir/Interfaces/ControlFlowInterfaces.h"
 #include "mlir/Interfaces/CopyOpInterface.h"
 #include "mlir/Interfaces/SideEffectInterfaces.h"
 #include "mlir/Interfaces/ViewLikeInterface.h"
 
 namespace mlir {
 
 class Location;
 class OpBuilder;
 
 raw_ostream &operator<<(raw_ostream &os, const Range &range);
 
 /// Return the list of Range (i.e. offset, size, stride). Each Range
 /// entry contains either the dynamic value or a ConstantIndexOp constructed
 /// with `b` at location `loc`.
 SmallVector<Range, 8> getOrCreateRanges(OffsetSizeAndStrideOpInterface op,
                                         OpBuilder &b, Location loc);
+
+namespace memref {
+
+/// This is a common utility used for patterns of the form
+/// "someop(memref.cast) -> someop". It folds the source of any memref.cast
+/// into the root operation directly.
+LogicalResult foldMemRefCast(Operation *op, Value inner = nullptr);
+
+/// Return an unranked/ranked tensor type for the given unranked/ranked memref
+/// type.
+Type getTensorTypeFromMemRefType(Type type);
+
+} // namespace memref
 } // namespace mlir
 
 //===----------------------------------------------------------------------===//
 // MemRef Dialect
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/MemRef/IR/MemRefOpsDialect.h.inc"
 
 //===----------------------------------------------------------------------===//
 // MemRef Dialect Operations
 //===----------------------------------------------------------------------===//
 
 #define GET_OP_CLASSES
 #include "mlir/Dialect/MemRef/IR/MemRefOps.h.inc"
 
 #endif // MLIR_DIALECT_MEMREF_IR_MEMREF_H_

mlir/include/mlir/Dialect/MemRef/IR/MemRefBase.td

 def MemRef_Dialect : Dialect {
   let name = "memref";
   let cppNamespace = "::mlir::memref";
   let description = [{
     The `memref` dialect is intended to hold core memref creation and
     manipulation ops, which are not strongly associated with any particular
     other dialect or domain abstraction.
   }];
-  let dependentDialects = ["arith::ArithmeticDialect", "tensor::TensorDialect"];
+  let dependentDialects = ["arith::ArithmeticDialect"];
   let hasConstantMaterializer = 1;
 }
 
 #endif // MEMREF_BASE

mlir/include/mlir/Dialect/MemRef/IR/MemRefOps.td

def MemRef_AllocaScopeReturnOp : MemRef_Op<"alloca_scope.return",
   let assemblyFormat =
       [{ attr-dict ($results^ `:` type($results))? }];
 
   // No custom verification needed.
   let verifier = ?;
 }
 
 //===----------------------------------------------------------------------===//
-// BufferCastOp
-//===----------------------------------------------------------------------===//
-
-def MemRef_BufferCastOp : MemRef_Op<"buffer_cast",
-    [SameOperandsAndResultShape, SameOperandsAndResultElementType, NoSideEffect,
-     TypesMatchWith<"type of 'tensor' is the tensor equivalent of 'memref'",
-                    "memref", "tensor",
-                    "getTensorTypeFromMemRefType($_self)">]> {
-  let summary = "tensor to memref cast operation";
-  let description = [{
-    Casts a tensor to a memref.
-
-    ```mlir
-    // Result type is tensor<4x?xf32>
-    %12 = memref.buffer_cast %10 : memref<4x?xf32, #map0, 42>
-    ```
-
-    Note, that mutating the result of the buffer cast operation leads to
-    undefined behavior.
-
-    This operation is a specialized variant of the built-in
-    unrealized_conversion_cast and is intended for use in the context of
-    gradual bufferization.
-  }];
-
-  let arguments = (ins AnyTensor:$tensor);
-  let results = (outs AnyRankedOrUnrankedMemRef:$memref);
-  // This op is fully verified by traits.
-  let verifier = ?;
-
-  let assemblyFormat = "$tensor attr-dict `:` type($memref)";
-
-  let hasFolder = 1;
-  let hasCanonicalizer = 1;
-}
-
-//===----------------------------------------------------------------------===//
 // CastOp
 //===----------------------------------------------------------------------===//
 
 def MemRef_CastOp : MemRef_Op<"cast", [
       NoSideEffect, SameOperandsAndResultShape,
       DeclareOpInterfaceMethods<CastOpInterface>,
       ViewLikeOpInterface,
       MemRefsNormalizable
let extraClassDeclaration = [{
 
     Value getViewSource() { return source(); }
   }];
 
   let hasFolder = 1;
 }
 
 //===----------------------------------------------------------------------===//
-// CloneOp
-//===----------------------------------------------------------------------===//
-
-def CloneOp : MemRef_Op<"clone", [
-    CopyOpInterface,
-    DeclareOpInterfaceMethods<MemoryEffectsOpInterface>,
-  ]> {
-  let builders = [
-    OpBuilder<(ins "Value":$value), [{
-      return build($_builder, $_state, value.getType(), value);
-    }]>];
-
-  let description = [{
-    Clones the data in the input view into an implicitly defined output view.
-
-    Usage:
-
-    ```mlir
-    %arg1 = memref.clone %arg0 : memref<?xf32> to memref<?xf32>
-    ```
-
-    Valid implementations of this operation may alias the input and output
-    views or create an actual copy. Mutating the source or result
-    of the clone operation after the clone operation thus leads to undefined
-    behavior.
-  }];
-
-  let arguments = (ins Arg<AnyRankedOrUnrankedMemRef, "", []>:$input);
-  let results = (outs Arg<AnyRankedOrUnrankedMemRef, "", []>:$output);
-
-  let extraClassDeclaration = [{
-    Value getSource() { return input(); }
-    Value getTarget() { return output(); }
-  }];
-
-  let assemblyFormat = "$input attr-dict `:` type($input) `to` type($output)";
-
-  let hasFolder = 1;
-  let verifier = ?;
-  let hasCanonicalizer = 1;
-}
-
-//===----------------------------------------------------------------------===//
 // CopyOp
 //===----------------------------------------------------------------------===//
 
 def CopyOp : MemRef_Op<"copy",
     [CopyOpInterface, SameOperandsElementType, SameOperandsShape]> {
 
   let description = [{
     Copies the data from the source to the destination memref.
let extraClassDeclaration = [{
     void setMemRef(Value value) { setOperand(0, value); }
     MemRefType getMemRefType() {
       return getMemRef().getType().cast<MemRefType>();
     }
 
     operand_range getIndices() { return {operand_begin() + 1, operand_end()}; }
   }];
 
-  let hasCanonicalizer = 1;
   let hasFolder = 1;
 
   let assemblyFormat = "$memref `[` $indices `]` attr-dict `:` type($memref)";
 }
 
 //===----------------------------------------------------------------------===//
 // PrefetchOp
 //===----------------------------------------------------------------------===//
let extraClassDeclaration = extraBaseClassDeclaration # [{
     llvm::SmallDenseSet<unsigned> getDroppedDims();
   }];
 
   let hasCanonicalizer = 1;
   let hasFolder = 1;
 }
 
 //===----------------------------------------------------------------------===//
-// TensorLoadOp
-//===----------------------------------------------------------------------===//
-
-def TensorLoadOp : MemRef_Op<"tensor_load",
-    [SameOperandsAndResultShape, SameOperandsAndResultElementType,
-     TypesMatchWith<"result type matches tensor equivalent of 'memref'",
-                    "memref", "result",
-                    "getTensorTypeFromMemRefType($_self)">]> {
-  let summary = "tensor load operation";
-  let description = [{
-    Create a tensor from a memref, making an independent copy of the element
-    data. The result value is a tensor whose shape and element type match the
-    memref operand.
-
-    The opposite of this op is buffer_cast. Together, these two ops are
-    useful for source/target materializations when doing type conversions
-    involving tensors and memrefs.
-
-    Example:
-
-    ```mlir
-    // Produces a value of tensor<4x?xf32> type.
-    %12 = memref.tensor_load %10 : memref<4x?xf32, #layout, memspace0>
-    ```
-
-    If tensor load is used in the bufferization steps, mutating the source
-    buffer after loading leads to undefined behavior.
-  }];
-
-  let arguments = (ins Arg<AnyRankedOrUnrankedMemRef,
-                       "the reference to load from", [MemRead]>:$memref);
-  let results = (outs AnyTensor:$result);
-  // TensorLoadOp is fully verified by traits.
-  let verifier = ?;
-
-  let builders = [
-    OpBuilder<(ins "Value":$memref), [{
-      $_state.addOperands(memref);
-      $_state.addTypes(getTensorTypeFromMemRefType(memref.getType()));
-    }]>];
-
-  let extraClassDeclaration = [{
-    /// The result of a tensor_load is always a tensor.
-    TensorType getType() {
-      Type resultType = getResult().getType();
-      if (resultType.isa<TensorType>())
-        return resultType.cast<TensorType>();
-      return {};
-    }
-  }];
-
-  let assemblyFormat = "$memref attr-dict `:` type($memref)";
-
-  let hasCanonicalizer = 1;
-  let hasFolder = 1;
-}
-
-//===----------------------------------------------------------------------===//
 // TensorStoreOp
 //===----------------------------------------------------------------------===//
 
 def TensorStoreOp : MemRef_Op<"tensor_store",
     [SameOperandsShape, SameOperandsElementType,
      TypesMatchWith<"type of 'value' matches tensor equivalent of 'memref'",
                     "memref", "tensor",
                     "getTensorTypeFromMemRefType($_self)">]> {

mlir/include/mlir/Dialect/SCF/Passes.td

 //===-- Passes.td - SCF pass definition file ---------------*- tablegen -*-===//
 //
 // 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_PASSES
 #define MLIR_DIALECT_SCF_PASSES
 
 include "mlir/Pass/PassBase.td"
 
 def SCFBufferize : FunctionPass<"scf-bufferize"> {
   let summary = "Bufferize the scf dialect.";
   let constructor = "mlir::createSCFBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect"];
 }
 
 // Note: Making these canonicalization patterns would require a dependency
 // of the SCF dialect on the Affine/Tensor/MemRef dialects or vice versa.
 def SCFForLoopCanonicalization
     : FunctionPass<"for-loop-canonicalization"> {
   let summary = "Canonicalize operations within scf.for loop bodies";
   let constructor = "mlir::createSCFForLoopCanonicalizationPass()";

mlir/include/mlir/Dialect/Shape/Transforms/Passes.td

def ShapeToShapeLowering : FunctionPass<"shape-to-shape-lowering"> {
   let summary = "Legalize Shape dialect to be convertible to Standard";
   let constructor = "mlir::createShapeToShapeLowering()";
 }
 
 // TODO: Generalize this to allow any type conversions desired.
 def ShapeBufferize : FunctionPass<"shape-bufferize"> {
   let summary = "Bufferize the shape dialect.";
   let constructor = "mlir::createShapeBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect"];
 }
 #endif // MLIR_DIALECT_SHAPE_TRANSFORMS_PASSES

mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.td

func @kernel_matvec(%arga: tensor<?x?xf64, #SparseMatrix>,
       return %0 : tensor<?xf64>
     }
     ```
   }];
   let constructor = "mlir::createSparsificationPass()";
   let dependentDialects = [
     "AffineDialect",
     "arith::ArithmeticDialect",
+    "bufferization::BufferizationDialect",
     "LLVM::LLVMDialect",
     "memref::MemRefDialect",
     "scf::SCFDialect",
     "sparse_tensor::SparseTensorDialect",
     "vector::VectorDialect",
   ];
   let options = [
     Option<"parallelization", "parallelization-strategy", "int32_t", "0",
```mlir
       After:
         %c1 = arith.constant 1 : index
         %0 = call @sparsePointers(%arg0, %c1) : (!llvm.ptr<i8>, index) -> memref<?xindex>
     ```
   }];
   let constructor = "mlir::createSparseTensorConversionPass()";
   let dependentDialects = [
     "arith::ArithmeticDialect",
+    "bufferization::BufferizationDialect",
     "LLVM::LLVMDialect",
     "linalg::LinalgDialect",
     "memref::MemRefDialect",
     "scf::SCFDialect",
     "sparse_tensor::SparseTensorDialect",
     "vector::VectorDialect",
   ];
 }
 
 #endif // MLIR_DIALECT_SPARSETENSOR_TRANSFORMS_PASSES

mlir/include/mlir/Dialect/StandardOps/Transforms/Passes.td

 //===-- Passes.td - StandardOps pass definition file -------*- tablegen -*-===//
 //
 // 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_STANDARD_TRANSFORMS_PASSES
 #define MLIR_DIALECT_STANDARD_TRANSFORMS_PASSES
 
 include "mlir/Pass/PassBase.td"
 
 def StdBufferize : FunctionPass<"std-bufferize"> {
   let summary = "Bufferize the std dialect";
   let constructor = "mlir::createStdBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect", "scf::SCFDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect", "scf::SCFDialect"];
 }
 
 def StdExpandOps : FunctionPass<"std-expand"> {
   let summary = "Legalize std operations to be convertible to LLVM.";
   let constructor = "mlir::createStdExpandOpsPass()";
 }
 
 def FuncBufferize : Pass<"func-bufferize", "ModuleOp"> {
let description = [{
     As this pass rewrites function operations, it also rewrites the
     corresponding return operations. Other return-like operations that
     implement the `ReturnLike` trait are not rewritten in general, as they
     require that the corresponding parent operation is also rewritten.
     Finally, this pass fails for unknown terminators, as we cannot decide
     whether they need rewriting.
   }];
   let constructor = "mlir::createFuncBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect"];
 }
 
 def TensorConstantBufferize : Pass<"tensor-constant-bufferize", "ModuleOp"> {
   let summary = "Bufferize tensor constants.";
   let description = [{
     This pass bufferizes tensor constants.
 
     This pass needs to be a module pass because it inserts memref.global
     ops into the module, which cannot be done safely from a function pass due to
     multi-threading. Most other bufferization passes can run in parallel at
     function granularity.
   }];
   let constructor = "mlir::createTensorConstantBufferizePass()";
-  let dependentDialects = ["memref::MemRefDialect"];
+  let dependentDialects = ["bufferization::BufferizationDialect",
+                           "memref::MemRefDialect"];
   let options = [
     Option<"alignment", "alignment", "unsigned", /*default=*/"0",
            "Create global memrefs with a specified alignment">,
   ];
 }
 
 #endif // MLIR_DIALECT_STANDARD_TRANSFORMS_PASSES

mlir/include/mlir/Dialect/Tensor/Transforms/Passes.td

 //===-- Passes.td - pass definition file -------------------*- tablegen -*-===//
 //
 // 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_TENSOR_TRANSFORMS_PASSES
 #define MLIR_DIALECT_TENSOR_TRANSFORMS_PASSES
 
 include "mlir/Pass/PassBase.td"
 
 def TensorBufferize : FunctionPass<"tensor-bufferize"> {
   let summary = "Bufferize the `tensor` dialect";
   let constructor = "mlir::createTensorBufferizePass()";
-  let dependentDialects = ["scf::SCFDialect", "memref::MemRefDialect"];
+  let dependentDialects = [
+    "bufferization::BufferizationDialect",
+    "memref::MemRefDialect",
+    "scf::SCFDialect"
+  ];
 }
 
 #endif // MLIR_DIALECT_TENSOR_TRANSFORMS_PASSES

mlir/include/mlir/InitAllDialects.h

 #define MLIR_INITALLDIALECTS_H_
 
 #include "mlir/Dialect/AMX/AMXDialect.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
 #include "mlir/Dialect/ArmNeon/ArmNeonDialect.h"
 #include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
 #include "mlir/Dialect/Async/IR/Async.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Complex/IR/Complex.h"
 #include "mlir/Dialect/DLTI/DLTI.h"
 #include "mlir/Dialect/EmitC/IR/EmitC.h"
 #include "mlir/Dialect/GPU/GPUDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/LLVMIR/NVVMDialect.h"
 #include "mlir/Dialect/LLVMIR/ROCDLDialect.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"

 inline void registerAllDialects(DialectRegistry &registry) {
   // clang-format off
   registry.insert<acc::OpenACCDialect,
                   AffineDialect,
                   arith::ArithmeticDialect,
                   amx::AMXDialect,
                   arm_neon::ArmNeonDialect,
                   async::AsyncDialect,
+                  bufferization::BufferizationDialect,
                   complex::ComplexDialect,
                   DLTIDialect,
                   emitc::EmitCDialect,
                   gpu::GPUDialect,
                   LLVM::LLVMDialect,
                   linalg::LinalgDialect,
                   math::MathDialect,
                   memref::MemRefDialect,

mlir/lib/Dialect/Affine/IR/AffineOps.cpp

 //===- AffineOps.cpp - MLIR Affine Operations -----------------------------===//
 //
 // 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/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Affine/IR/AffineValueMap.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
+#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/IR/BuiltinOps.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/IR/Matchers.h"
 #include "mlir/IR/OpImplementation.h"
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/SetVector.h"

mlir/lib/Dialect/Arithmetic/Transforms/Bufferize.cpp

 //===- Bufferize.cpp - Bufferization for Arithmetic ops ---------*- 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
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Transforms/Bufferize.h"
 #include "PassDetail.h"
 #include "mlir/Dialect/Arithmetic/Transforms/Passes.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 
 using namespace mlir;
 
 namespace {
 
 /// Bufferize arith.index_cast.
 struct BufferizeIndexCastOp : public OpConversionPattern<arith::IndexCastOp> {

mlir/lib/Dialect/Arithmetic/Transforms/PassDetail.h

 #define DIALECT_ARITHMETIC_TRANSFORMS_PASSDETAIL_H_
 
 #include "mlir/Pass/Pass.h"
 
 namespace mlir {
 
 class StandardOpsDialect;
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace memref {
 class MemRefDialect;
 } // end namespace memref
 
 #define GEN_PASS_CLASSES
 #include "mlir/Dialect/Arithmetic/Transforms/Passes.h.inc"
 
 } // end namespace mlir
 
 #endif // DIALECT_ARITHMETIC_TRANSFORMS_PASSDETAIL_H_

mlir/lib/Dialect/Bufferization/IR/BufferizationDialect.cpp

+//===----------------------------------------------------------------------===//
+//
+// 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/Bufferization/IR/Bufferization.h"
+
+using namespace mlir;
+using namespace mlir::bufferization;
+
+#include "mlir/Dialect/Bufferization/IR/BufferizationOpsDialect.cpp.inc"
+
+//===----------------------------------------------------------------------===//
+// Bufferization Dialect
+//===----------------------------------------------------------------------===//
+
+void mlir::bufferization::BufferizationDialect::initialize() {
+  addOperations<
+#define GET_OP_LIST
+#include "mlir/Dialect/Bufferization/IR/BufferizationOps.cpp.inc"
+      >();
+}

mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp

+
+//===----------------------------------------------------------------------===//
+//
+// 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/Bufferization/IR/Bufferization.h"
+#include "mlir/Dialect/MemRef/Utils/MemRefUtils.h"
+
+using namespace mlir;
+using namespace mlir::bufferization;
+
+//===----------------------------------------------------------------------===//
+// CloneOp
+//===----------------------------------------------------------------------===//
+
+void CloneOp::getEffects(
+    SmallVectorImpl<SideEffects::EffectInstance<MemoryEffects::Effect>>
+        &effects) {
+  effects.emplace_back(MemoryEffects::Read::get(), input(),
+                       SideEffects::DefaultResource::get());
+  effects.emplace_back(MemoryEffects::Write::get(), output(),
+                       SideEffects::DefaultResource::get());
+  effects.emplace_back(MemoryEffects::Allocate::get(), output(),
+                       SideEffects::DefaultResource::get());
+}
+
+OpFoldResult CloneOp::fold(ArrayRef<Attribute> operands) {
+  return succeeded(memref::foldMemRefCast(*this)) ? getResult() : Value();
+}
+
+namespace {
+
+/// Merge the clone and its source (by converting the clone to a cast) when
+/// possible.
+struct SimplifyClones : public OpRewritePattern<CloneOp> {
+  using OpRewritePattern<CloneOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(CloneOp cloneOp,
+                                PatternRewriter &rewriter) const override {
+    if (cloneOp.use_empty()) {
+      rewriter.eraseOp(cloneOp);
+      return success();
+    }
+
+    Value source = cloneOp.input();
+
+    // This only finds dealloc operations for the immediate value. It should
+    // also consider aliases. That would also make the safety check below
+    // redundant.
+    llvm::Optional<Operation *> maybeCloneDeallocOp =
+        findDealloc(cloneOp.output());
+    // Skip if either of them has > 1 deallocate operations.
+    if (!maybeCloneDeallocOp.hasValue())
+      return failure();
+    llvm::Optional<Operation *> maybeSourceDeallocOp = findDealloc(source);
+    if (!maybeSourceDeallocOp.hasValue())
+      return failure();
+    Operation *cloneDeallocOp = *maybeCloneDeallocOp;
+    Operation *sourceDeallocOp = *maybeSourceDeallocOp;
+
+    // If both are deallocated in the same block, their in-block lifetimes
+    // might not fully overlap, so we cannot decide which one to drop.
+    if (cloneDeallocOp && sourceDeallocOp &&
+        cloneDeallocOp->getBlock() == sourceDeallocOp->getBlock())
+      return failure();
+
+    Block *currentBlock = cloneOp->getBlock();
+    Operation *redundantDealloc = nullptr;
+    if (cloneDeallocOp && cloneDeallocOp->getBlock() == currentBlock) {
+      redundantDealloc = cloneDeallocOp;
+    } else if (sourceDeallocOp && sourceDeallocOp->getBlock() == currentBlock) {
+      redundantDealloc = sourceDeallocOp;
+    }
+
+    if (!redundantDealloc)
+      return failure();
+
+    // Safety check that there are no other deallocations inbetween
+    // cloneOp and redundantDealloc, as otherwise we might deallocate an alias
+    // of source before the uses of the clone. With alias information, we could
+    // restrict this to only fail of the dealloc's operand is an alias
+    // of the source.
+    for (Operation *pos = cloneOp->getNextNode(); pos != redundantDealloc;
+         pos = pos->getNextNode()) {
+      auto effectInterface = dyn_cast<MemoryEffectOpInterface>(pos);
+      if (!effectInterface)
+        continue;
+      if (effectInterface.hasEffect<MemoryEffects::Free>())
+        return failure();
+    }
+
+    rewriter.replaceOpWithNewOp<memref::CastOp>(cloneOp, cloneOp.getType(),
+                                                source);
+    rewriter.eraseOp(redundantDealloc);
+    return success();
+  }
+};
+
+} // namespace.
+
+void CloneOp::getCanonicalizationPatterns(OwningRewritePatternList &results,
+                                          MLIRContext *context) {
+  results.insert<SimplifyClones>(context);
+}
+
+//===----------------------------------------------------------------------===//
+// ToTensorOp
+//===----------------------------------------------------------------------===//
+
+OpFoldResult ToTensorOp::fold(ArrayRef<Attribute>) {
+  if (auto toMemref = memref().getDefiningOp<ToMemrefOp>())
+    // Approximate alias analysis by conservatively folding only when no there
+    // is no interleaved operation.
+    if (toMemref->getBlock() == this->getOperation()->getBlock() &&
+        toMemref->getNextNode() == this->getOperation())
+      return toMemref.tensor();
+  return {};
+}
+
+namespace {
+
+struct DimOfToTensorFolder : public OpRewritePattern<tensor::DimOp> {
+  using OpRewritePattern<tensor::DimOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(tensor::DimOp dimOp,
+                                PatternRewriter &rewriter) const override {
+    auto memrefToTensorOp = dimOp.source().getDefiningOp<ToTensorOp>();
+    if (!memrefToTensorOp)
+      return failure();
+
+    rewriter.replaceOpWithNewOp<memref::DimOp>(dimOp, memrefToTensorOp.memref(),
+                                               dimOp.index());
+    return success();
+  }
+};
+
+} // namespace
+
+void ToTensorOp::getCanonicalizationPatterns(RewritePatternSet &results,
+                                             MLIRContext *context) {
+  results.add<DimOfToTensorFolder>(context);
+}
+
+//===----------------------------------------------------------------------===//
+// ToMemrefOp
+//===----------------------------------------------------------------------===//
+
+OpFoldResult ToMemrefOp::fold(ArrayRef<Attribute>) {
+  if (auto memrefToTensor = tensor().getDefiningOp<ToTensorOp>())
+    if (memrefToTensor.memref().getType() == getType())
+      return memrefToTensor.memref();
+  return {};
+}
+
+namespace {
+
+/// Replace tensor.cast + to_memref by to_memref + memref.cast.
+struct ToMemrefOfCast : public OpRewritePattern<ToMemrefOp> {
+  using OpRewritePattern<ToMemrefOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(ToMemrefOp toMemref,
+                                PatternRewriter &rewriter) const final {
+    auto tensorCastOperand =
+        toMemref.getOperand().getDefiningOp<tensor::CastOp>();
+    if (!tensorCastOperand)
+      return failure();
+    auto srcTensorType =
+        tensorCastOperand.getOperand().getType().dyn_cast<RankedTensorType>();
+    if (!srcTensorType)
+      return failure();
+    auto memrefType = MemRefType::get(srcTensorType.getShape(),
+                                      srcTensorType.getElementType());
+    Value memref = rewriter.create<ToMemrefOp>(toMemref.getLoc(), memrefType,
+                                               tensorCastOperand.getOperand());
+    rewriter.replaceOpWithNewOp<memref::CastOp>(toMemref, toMemref.getType(),
+                                                memref);
+    return success();
+  }
+};
+
+/// Canonicalize bufferization.to_tensor + bufferization.to_memref to
+/// memref.cast when type mismatches prevent `ToMemrefOp::fold` to kick in.
+struct TensorLoadToMemref : public OpRewritePattern<ToMemrefOp> {
+  using OpRewritePattern<ToMemrefOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(ToMemrefOp toMemref,
+                                PatternRewriter &rewriter) const final {
+    auto memrefToTensor = toMemref.tensor().getDefiningOp<ToTensorOp>();
+    // Bail unless we have a memref_to_tensor + tensor_to_memref with different
+    // types. `ToMemrefOp::fold` handles the same type case.
+    if (!memrefToTensor ||
+        memrefToTensor.memref().getType() == toMemref.getType())
+      return failure();
+    // If types are definitely not cast-compatible, bail.
+    if (!memref::CastOp::areCastCompatible(memrefToTensor.memref().getType(),
+                                           toMemref.getType()))
+      return failure();
+
+    // We already know that the types are potentially cast-compatible. However
+    // in case the affine maps are different, we may need to use a copy if we go
+    // from dynamic to static offset or stride (the canonicalization cannot know
+    // at this point that it is really cast compatible).
+    auto isGuaranteedCastCompatible = [](MemRefType source, MemRefType target) {
+      int64_t sourceOffset, targetOffset;
+      SmallVector<int64_t, 4> sourceStrides, targetStrides;
+      if (failed(getStridesAndOffset(source, sourceStrides, sourceOffset)) ||
+          failed(getStridesAndOffset(target, targetStrides, targetOffset)))
+        return false;
+      auto dynamicToStatic = [](int64_t a, int64_t b) {
+        return a == MemRefType::getDynamicStrideOrOffset() &&
+               b != MemRefType::getDynamicStrideOrOffset();
+      };
+      if (dynamicToStatic(sourceOffset, targetOffset))
+        return false;
+      for (auto it : zip(sourceStrides, targetStrides))
+        if (dynamicToStatic(std::get<0>(it), std::get<1>(it)))
+          return false;
+      return true;
+    };
+
+    auto memrefToTensorType =
+        memrefToTensor.memref().getType().dyn_cast<MemRefType>();
+    auto toMemrefType = toMemref.getType().dyn_cast<MemRefType>();
+    if (memrefToTensorType && toMemrefType &&
+        !isGuaranteedCastCompatible(memrefToTensorType, toMemrefType)) {
+      MemRefType resultType = toMemrefType;
+      auto loc = toMemref.getLoc();
+      SmallVector<Value, 4> dynamicOperands;
+      for (int i = 0; i < resultType.getRank(); ++i) {
+        if (resultType.getShape()[i] != ShapedType::kDynamicSize)
+          continue;
+        auto index = rewriter.createOrFold<arith::ConstantIndexOp>(loc, i);
+        Value size = rewriter.create<tensor::DimOp>(loc, memrefToTensor, index);
+        dynamicOperands.push_back(size);
+      }
+      auto copy =
+          rewriter.create<memref::AllocOp>(loc, resultType, dynamicOperands);
+      rewriter.create<memref::CopyOp>(loc, memrefToTensor.memref(), copy);
+      rewriter.replaceOp(toMemref, {copy});
+    } else
+      rewriter.replaceOpWithNewOp<memref::CastOp>(toMemref, toMemref.getType(),
+                                                  memrefToTensor.memref());
+    return success();
+  }
+};
+
+/// Fold a load on a to_memref operation into an tensor.extract on the
+/// corresponding tensor.
+struct LoadOfToMemref : public OpRewritePattern<memref::LoadOp> {
+  using OpRewritePattern<memref::LoadOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(memref::LoadOp load,
+                                PatternRewriter &rewriter) const override {
+    auto toMemref = load.memref().getDefiningOp<ToMemrefOp>();
+    if (!toMemref)
+      return failure();
+
+    rewriter.replaceOpWithNewOp<tensor::ExtractOp>(load, toMemref.tensor(),
+                                                   load.indices());
+    return success();
+  }
+};
+
+/// Fold dim of a to_memref into the dim of the tensor.
+struct DimOfCastOp : public OpRewritePattern<memref::DimOp> {
+  using OpRewritePattern<memref::DimOp>::OpRewritePattern;
+
+  LogicalResult matchAndRewrite(memref::DimOp dimOp,
+                                PatternRewriter &rewriter) const override {
+    auto castOp = dimOp.source().getDefiningOp<ToMemrefOp>();
+    if (!castOp)
+      return failure();
+    Value newSource = castOp.getOperand();
+    rewriter.replaceOpWithNewOp<tensor::DimOp>(dimOp, newSource, dimOp.index());
+    return success();
+  }
+};
+
+} // namespace
+
+void ToMemrefOp::getCanonicalizationPatterns(RewritePatternSet &results,
+                                             MLIRContext *context) {
+  results.add<DimOfCastOp, LoadOfToMemref, ToMemrefOfCast, TensorLoadToMemref>(
+      context);
+}
+
+Optional<Operation *> CloneOp::buildDealloc(OpBuilder &builder, Value alloc) {
+  return builder.create<memref::DeallocOp>(alloc.getLoc(), alloc)
+      .getOperation();
+}
+
+Optional<Value> CloneOp::buildClone(OpBuilder &builder, Value alloc) {
+  return builder.create<CloneOp>(alloc.getLoc(), alloc).getResult();
+}
+
+//===----------------------------------------------------------------------===//
+// TableGen'd op method definitions
+//===----------------------------------------------------------------------===//
+
+#define GET_OP_CLASSES
+#include "mlir/Dialect/Bufferization/IR/BufferizationOps.cpp.inc"

mlir/lib/Dialect/Bufferization/IR/CMakeLists.txt

-add_mlir_library(MLIRAllocationOpInterface
+add_mlir_dialect_library(MLIRBufferization
   AllocationOpInterface.cpp
+  BufferizationOps.cpp
+  BufferizationDialect.cpp
 
   ADDITIONAL_HEADER_DIRS
-  ${MLIR_MAIN_INCLUDE_DIR}/mlir/Dialect/Bufferization/IR
+  ${MLIR_MAIN_INCLUDE_DIR}/mlir/Dialect/Bufferization
 
   DEPENDS
   MLIRAllocationOpInterfaceIncGen
+  MLIRBufferizationOpsIncGen
 
   LINK_LIBS PUBLIC
+  MLIRDialect
   MLIRIR
+  MLIRTensor
+  MLIRMemRef
   )

mlir/lib/Dialect/Linalg/Analysis/DependenceAnalysis.cpp

 //===- DependenceAnalysis.cpp - Dependence analysis on SSA views ----------===//
 //
 // 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 view-based alias and dependence analyses.
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/Linalg/Analysis/DependenceAnalysis.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
-#include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/IR/BuiltinOps.h"
 
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
 
 #define DEBUG_TYPE "linalg-dependence-analysis"
 
while (true) {
     Operation *defOp = v.getDefiningOp();
     if (!defOp)
       return v;
 
     // Treat RegionBranchOpInterfaces like an allocate and don't try to follow
     // the aliasing further.
     if (isa<RegionBranchOpInterface>(defOp))
       return v;
-    if (isa<memref::BufferCastOp>(defOp))
+    if (isa<bufferization::ToMemrefOp>(defOp))
       return v;
 
     if (auto memEffect = dyn_cast<MemoryEffectOpInterface>(defOp)) {
       // Collect all memory effects on `v`.
       SmallVector<MemoryEffects::EffectInstance, 1> effects;
       memEffect.getEffectsOnValue(v, effects);
 
       // If we have the 'Allocate' memory effect on `v`, then `v` should be the

mlir/lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp

 //===- BufferizableOpInterface.cpp - Comprehensive Bufferize --------------===//
 //
 // 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/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h"
 
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/IR/AsmState.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/IR/BuiltinOps.h"
 #include "mlir/IR/Operation.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
 #include "llvm/Support/Debug.h"
mlir::linalg::comprehensive_bufferize::bufferize(Block *block,
   return success();
 }
 
 LogicalResult
 mlir::linalg::comprehensive_bufferize::bufferize(Operation *op,
                                                  BufferizationState &state) {
   OpBuilder b(op->getContext());
 
-  // Skip BufferCast and TensorLoad ops.
-  if (isa<memref::BufferCastOp, memref::TensorLoadOp>(op))
+  // Skip ToMemrefOp and ToTensorOp.
+  if (isa<bufferization::ToMemrefOp, bufferization::ToTensorOp>(op))
     return success();
 
   // Check if op has tensor results or operands.
   auto isaTensor = [](Type t) { return t.isa<TensorType>(); };
   bool hasTensorResult = any_of(op->getResultTypes(), isaTensor);
   bool hasTensorOperand = any_of(op->getOperandTypes(), isaTensor);
 
   // No tensor results or operands: Simply bufferize all nested ops.

mlir/lib/Dialect/Linalg/ComprehensiveBufferize/CMakeLists.txt

 set(LLVM_OPTIONAL_SOURCES
   BufferizableOpInterface.cpp
   ComprehensiveBufferize.cpp
   LinalgInterfaceImpl.cpp
   TensorInterfaceImpl.cpp
   VectorInterfaceImpl.cpp
 )
 
 add_mlir_dialect_library(MLIRBufferizableOpInterface
   BufferizableOpInterface.cpp
 
   DEPENDS
   MLIRBufferizableOpInterfaceIncGen
 
   LINK_LIBS PUBLIC
   MLIRIR
+  MLIRBufferization
   MLIRMemRef
 )
 
 add_mlir_dialect_library(MLIRLinalgBufferizableOpInterfaceImpl
   LinalgInterfaceImpl.cpp
 
   LINK_LIBS PUBLIC
   MLIRBufferizableOpInterface

mlir/lib/Dialect/Linalg/ComprehensiveBufferize/ComprehensiveBufferize.cpp

 //  c. at the end of this bufferization, 3 cases may occur:
 //     i. inplaceable function arguments may be reused in place after the
 //        function itself has been bufferized. This is encoded by IR resembling:
 //
 //        ```
 //          #map = affine_map<(d0)[s0, s1] -> (d0 * s1 + s0)>
 //           func @foo(%A: tensor<?xf32> {linalg.inplaceable = true})
 //              -> tensor<?xf32> {
-//            %0 = memref.buffer_cast %A : memref<?xf32, #map>
+//            %0 = bufferization.to_memref %A : memref<?xf32, #map>
 //            // ... uses of %0
-//            %res = memref.tensor_load %0 : memref<?xf32, #map>
+//            %res = bufferization.to_tensor %0 : memref<?xf32, #map>
 //            return %res : tensor<?xf32>
 //          }
 //        ```
 //
 //        this is the cue for the bufferization of the function foo (and calls
 //        to it) may bufferize to `func @foo(%A: memref<?xf32, some_layout>)`.
 //        To fully achieve bufferization, an additional analysis is needed to
 //        determine whether function argument/operand pairs bufferize to a
 //        single inplace buffer argument (i.e. functions may return tensors in
 //        arbitrary order that may not match argument numbers).
 //
 //    ii. results that don't map to an inplaceable function argument are
 //        generally allocated. Since memref semantics wrt ownership of the
 //        underlying memory region are not well-defined, comprehensive
 //        bufferization chooses to perform allocations in a scoped fashion:
 //        returning memrefs is always considered illegal.
 //        Such scenarios are encoded by IR resembling:
 //
 //        ```
 //          #map = affine_map<(d0)[s0, s1] -> (d0 * s1 + s0)>
 //          func @foo(%A: tensor<?xf32> {linalg.inplaceable = true})
 //              -> tensor<?xf32> {
-//            %0 = memref.buffer_cast %A : memref<?xf32, #map>
+//            %0 = bufferization.to_memref %A : memref<?xf32, #map>
 //            %1 = memref.dim %0, %c0 : memref<?xf32, #map>
 //            %2 = memref.alloc(%1) : memref<?xf32>
 //            %3 = memref.cast %2 : memref<?xf32> to memref<?xf32, #map>
 //            // ... uses of %3
 //            memref.dealloc %2 : memref<?xf32, #map>
-//            %res = memref.tensor_load %3 : memref<?xf32, #map>
+//            %res = bufferization.to_tensor %3 : memref<?xf32, #map>
 //            return %res : tensor<?xf32>
 //          }
 //       ```
 //
 //        this is the cue for the bufferization of the function foo (and calls
 //        to it) that it must bufferize to `func @foo(%A: memref<?xf32,
 //        some_layout>,
 //                   %B: memref<?xf32, some_layout>)` (i.e. make a cloned
 //        allocation of the result tensor)
 //        To fully achieve bufferization, the alloc/dealloc pair must be lifted
 //        out of the function at each call site.
 //
 //   iii. as an optimization over ii., it may be possible to reuse an argument
 //        and only want to return a slice.
 //        This may forego allocation by letting *all* callers decide whether to
 //        pass a new *aliasing* memref function argument (i.e. a subview).
 //        Without loss of generality, callers may agree to allocate a new buffer
 //        to avoid this aliasing. Such scenarios are encoded by IR resembling:
 //
 //        ```
 //          #map = affine_map<(d0)[s0, s1] -> (d0 * s1 + s0)>
 //          func @foo(%arg0: tensor<?xf32> {linalg.inplaceable = true})
 //              -> tensor<4xf32> {
-//            %0 = memref.buffer_cast %arg0 : memref<?xf32, #map>
+//            %0 = bufferization.to_memref %arg0 : memref<?xf32, #map>
 //            %1 = memref.subview %0[0] [4] [1] : memref<?xf32, #map> to
 //                                                memref<4xf32, #map>
 //            // ... inplace computes into %1
-//            %3 = memref.tensor_load %1 : memref<4xf32, #map>
+//            %3 = bufferization.to_tensor %1 : memref<4xf32, #map>
 //            return %3 : tensor<4xf32>
 //          }
 //        ```
 //
 //  Note: In the future, it may be worthwhile to design special bufferization
 //  ops to encode the desired semantics at function boundaries for i., ii. and
 //  iii.
 //
 //  Lastly, note that layout map chosen to bufferize is the most dynamic
 //  canonical strided layout of the proper rank. This ensures compatibility with
 //  expected layouts after transformations. Combinations of memref.cast +
 //  canonicalization are responsible for clean ups.
 
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/ComprehensiveBufferize.h"
 
 #include <random>
 
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/Utils/StaticValueUtils.h"
 #include "mlir/IR/AsmState.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/IR/Operation.h"
 #include "mlir/IR/TypeUtilities.h"
if (!tensorType)
       continue;
     auto rankedTensorType = tensorType.dyn_cast<RankedTensorType>();
     // Cast the tensor to the most dynamic buffer possible. Further
     // canonicalizations will clean up.
     Type memRefType = rankedTensorType
                           ? getDynamicMemRefType(rankedTensorType)
                           : getContiguousOrUnrankedMemRefType(tensorType);
     Value bufferCast =
-        b.create<memref::BufferCastOp>(funcOp.getLoc(), memRefType, bbArg);
+        b.create<bufferization::ToMemrefOp>(funcOp.getLoc(), memRefType, bbArg);
     state.aliasInfo.insertNewBufferEquivalence(bufferCast, bbArg);
     state.mapBuffer(bbArg, bufferCast);
   }
 
   // Bufferize function body.
   return bufferize(&funcOp.body(), state);
 }
 
if (!tensorType) {
       continue;
     }
 
     // Get the buffer type from the bufferized function type.
     Type memrefType = bufferizedFuncType.getInput(idx);
     Value memref = frontBlock.addArgument(memrefType);
     OpBuilder b(funcOp->getContext());
     b.setInsertionPointToStart(&frontBlock);
-    // Replace all uses of bbArg through a BufferCastOp by a memref::CastOp.
+    // Replace all uses of bbArg through a ToMemrefOp by a memref::CastOp.
     for (auto &use : llvm::make_early_inc_range(bbArg.getUses())) {
-      if (auto bufferCastOp = dyn_cast<memref::BufferCastOp>(use.getOwner())) {
+      if (auto toMemrefOp =
+              dyn_cast<bufferization::ToMemrefOp>(use.getOwner())) {
         auto castOp = b.create<memref::CastOp>(
-            funcOp.getLoc(), bufferCastOp.memref().getType(), memref);
-        bufferCastOp.memref().replaceAllUsesWith(castOp);
+            funcOp.getLoc(), toMemrefOp.memref().getType(), memref);
+        toMemrefOp.memref().replaceAllUsesWith(castOp);
         aliasInfo.insertNewBufferEquivalence(castOp.dest(),
-                                             bufferCastOp.memref());
+                                             toMemrefOp.memref());
       }
     }
     // Replace all remaining uses by a tensor_load.
     if (!bbArg.use_empty()) {
-      auto tensorLoadOp =
-          b.create<memref::TensorLoadOp>(funcOp.getLoc(), memref);
-      aliasInfo.insertNewBufferEquivalence(tensorLoadOp, bbArg);
-      bbArg.replaceAllUsesWith(tensorLoadOp);
+      auto toTensorOp =
+          b.create<bufferization::ToTensorOp>(funcOp.getLoc(), memref);
+      aliasInfo.insertNewBufferEquivalence(toTensorOp, bbArg);
+      bbArg.replaceAllUsesWith(toTensorOp);
     }
     frontBlock.eraseArgument(0);
     // TODO: add support to erase aliasInfo entries if deemed necessary.
   }
 
   // 4. Rewrite the FuncOp type to buffer form.
   funcOp.setType(bufferizedFuncType);
 
if (funcOp.body().empty()) {
                 getEquivalentEnclosingFuncBBArg(returnVal, state.aliasInfo)) {
           Value oldRes = callOp->getResult(returnOperand.getOperandNumber());
           int64_t idx = bbArg.getArgNumber();
           Value buffer = state.lookupBuffer(callOp->getOperand(idx));
           // Add CallOp operand/result equivalence: this is interprocedural
           // info.
           state.aliasInfo.insertNewBufferEquivalence(oldRes, buffer);
           state.mapBuffer(oldRes, buffer);
-          // Add a TensorLoadOp to kill all uses of the CallOp return.
+          // Add a ToTensorOp to kill all uses of the CallOp return.
           // Replace all uses of the CallOp results so we can erase the CallOp.
-          // This TensorLoadOp must fold/DCE away or bufferization should be
+          // This ToTensorOp must fold/DCE away or bufferization should be
           // considered failed.
-          Value tensorLoad =
-              b.create<memref::TensorLoadOp>(callOp.getLoc(), buffer);
-          oldRes.replaceAllUsesWith(tensorLoad);
+          Value toTensor =
+              b.create<bufferization::ToTensorOp>(callOp.getLoc(), buffer);
+          oldRes.replaceAllUsesWith(toTensor);
           // Add new op equivalence info.
-          state.aliasInfo.insertNewBufferEquivalence(tensorLoad, buffer);
-          state.mapBuffer(tensorLoad, buffer);
+          state.aliasInfo.insertNewBufferEquivalence(toTensor, buffer);
+          state.mapBuffer(toTensor, buffer);
           continue;
         }
 
         // TODO: Need to hoist above function boundary.
         if (Operation *allocOp =
                 getEquivalentAlloc(returnVal, state.aliasInfo)) {
           hoistedArguments.push_back(allocOp->getResult(0));
           continue;
LogicalResult bufferize(Operation *op, OpBuilder &b,
 
     assert(isa<FuncOp>(returnOp->getParentOp()) &&
            "only support FuncOp parent for ReturnOp");
     for (OpOperand &operand : returnOp->getOpOperands()) {
       auto tensorType = operand.get().getType().dyn_cast<TensorType>();
       if (!tensorType)
         continue;
       Value v = state.lookupBuffer(operand.get());
-      Value returnTensor = b.create<memref::TensorLoadOp>(returnOp.getLoc(), v);
+      Value returnTensor =
+          b.create<bufferization::ToTensorOp>(returnOp.getLoc(), v);
       operand.set(returnTensor);
       state.aliasInfo.insertNewBufferEquivalence(returnTensor, v);
       state.mapBuffer(returnTensor, v);
     }
     return success();
   }
 };
 

mlir/lib/Dialect/Linalg/IR/LinalgInterfaces.cpp

 //===- LinalgInterfaces.cpp - Linalg interfaces implementation ------------===//
 //
 // 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/Linalg/IR/LinalgInterfaces.h"
 
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/IR/AffineExprVisitor.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "llvm/ADT/SmallSet.h"
 
 using namespace mlir;
 using namespace mlir::linalg;
 

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

 //===- Bufferize.cpp - Bufferization of linalg ops ------------------===//
 //
 // 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/Transforms/Bufferize.h"
 #include "PassDetail.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
 #include "mlir/Dialect/Linalg/Passes.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 #include "mlir/Dialect/Math/IR/Math.h"
 #include "mlir/Dialect/StandardOps/Transforms/Passes.h"
 #include "mlir/Dialect/StandardOps/Utils/Utils.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"

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

 //===- ComprehensiveBufferize.cpp - Single pass bufferization -------------===//
 //
 // 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/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/ComprehensiveBufferize.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/LinalgInterfaceImpl.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/TensorInterfaceImpl.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/VectorInterfaceImpl.h"
 #include "mlir/Dialect/Linalg/Passes.h"
 #include "mlir/Pass/Pass.h"
 #include "mlir/Pass/PassManager.h"
struct LinalgComprehensiveModuleBufferize
 
   LinalgComprehensiveModuleBufferize(
       const LinalgComprehensiveModuleBufferize &p) {}
 
   void runOnOperation() override;
 
   void getDependentDialects(DialectRegistry &registry) const override {
     registry
-        .insert<linalg::LinalgDialect, memref::MemRefDialect,
-                tensor::TensorDialect, vector::VectorDialect, scf::SCFDialect,
+        .insert<bufferization::BufferizationDialect, linalg::LinalgDialect,
+                memref::MemRefDialect, tensor::TensorDialect,
+                vector::VectorDialect, scf::SCFDialect,
                 arith::ArithmeticDialect, StandardOpsDialect, AffineDialect>();
     registerBufferizableOpInterfaceExternalModels(registry);
     linalg_ext::registerBufferizableOpInterfaceExternalModels(registry);
     tensor_ext::registerBufferizableOpInterfaceExternalModels(registry);
     vector_ext::registerBufferizableOpInterfaceExternalModels(registry);
   }
 };
 } // end namespace

mlir/lib/Dialect/Linalg/Transforms/PassDetail.h

 // Forward declaration from Dialect.h
 template <typename ConcreteDialect>
 void registerDialect(DialectRegistry &registry);
 
 namespace arith {
 class ArithmeticDialect;
 } // end namespace arith
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace linalg {
 class LinalgDialect;
 } // end namespace linalg
 
 namespace scf {
 class SCFDialect;
 } // end namespace scf
 

mlir/lib/Dialect/MemRef/IR/MemRefDialect.cpp

 //===----------------------------------------------------------------------===//
 //
 // 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/MemRef/IR/MemRef.h"
-#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Transforms/InliningUtils.h"
 
 using namespace mlir;
 using namespace mlir::memref;
 
 #include "mlir/Dialect/MemRef/IR/MemRefOpsDialect.cpp.inc"
 
 //===----------------------------------------------------------------------===//

mlir/lib/Dialect/MemRef/IR/MemRefOps.cpp

 //===----------------------------------------------------------------------===//
 //
 // 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/Arithmetic/IR/Arithmetic.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/MemRef/Utils/MemRefUtils.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/StandardOps/Utils/Utils.h"
-#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Utils/StaticValueUtils.h"
 #include "mlir/IR/AffineMap.h"
 #include "mlir/IR/Builders.h"
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/Matchers.h"
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/Interfaces/InferTypeOpInterface.h"

 
 //===----------------------------------------------------------------------===//
 // Common canonicalization pattern support logic
 //===----------------------------------------------------------------------===//
 
 /// This is a common class used for patterns of the form
 /// "someop(memrefcast) -> someop".  It folds the source of any memref.cast
 /// into the root operation directly.
-static LogicalResult foldMemRefCast(Operation *op, Value inner = nullptr) {
+LogicalResult mlir::memref::foldMemRefCast(Operation *op, Value inner) {
   bool folded = false;
   for (OpOperand &operand : op->getOpOperands()) {
     auto cast = operand.get().getDefiningOp<CastOp>();
     if (cast && operand.get() != inner &&
         !cast.getOperand().getType().isa<UnrankedMemRefType>()) {
       operand.set(cast.getOperand());
       folded = true;
     }
   }
   return success(folded);
 }
 
-//===----------------------------------------------------------------------===//
-// Helpers for GlobalOp
-//===----------------------------------------------------------------------===//
-
-static Type getTensorTypeFromMemRefType(Type type) {
+/// Return an unranked/ranked tensor type for the given unranked/ranked memref
+/// type.
+Type mlir::memref::getTensorTypeFromMemRefType(Type type) {
   if (auto memref = type.dyn_cast<MemRefType>())
     return RankedTensorType::get(memref.getShape(), memref.getElementType());
   if (auto memref = type.dyn_cast<UnrankedMemRefType>())
     return UnrankedTensorType::get(memref.getElementType());
   return NoneType::get(type.getContext());
 }
 
 //===----------------------------------------------------------------------===//

 static LogicalResult verify(AssumeAlignmentOp op) {
   unsigned alignment = op.alignment();
   if (!llvm::isPowerOf2_32(alignment))
     return op.emitOpError("alignment must be power of 2");
   return success();
 }
 
 //===----------------------------------------------------------------------===//
-// BufferCastOp
-//===----------------------------------------------------------------------===//
-
-OpFoldResult BufferCastOp::fold(ArrayRef<Attribute>) {
-  if (auto tensorLoad = tensor().getDefiningOp<TensorLoadOp>())
-    if (tensorLoad.memref().getType() == getType())
-      return tensorLoad.memref();
-  return {};
-}
-
-namespace {
-/// Replace tensor_cast + buffer_cast by buffer_cast + memref_cast.
-struct BufferCast : public OpRewritePattern<BufferCastOp> {
-  using OpRewritePattern<BufferCastOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(BufferCastOp bufferCast,
-                                PatternRewriter &rewriter) const final {
-    auto tensorCastOperand =
-        bufferCast.getOperand().getDefiningOp<tensor::CastOp>();
-    if (!tensorCastOperand)
-      return failure();
-    auto srcTensorType =
-        tensorCastOperand.getOperand().getType().dyn_cast<RankedTensorType>();
-    if (!srcTensorType)
-      return failure();
-    auto memrefType = MemRefType::get(srcTensorType.getShape(),
-                                      srcTensorType.getElementType());
-    Value memref = rewriter.create<BufferCastOp>(
-        bufferCast.getLoc(), memrefType, tensorCastOperand.getOperand());
-    rewriter.replaceOpWithNewOp<CastOp>(bufferCast, bufferCast.getType(),
-                                        memref);
-    return success();
-  }
-};
-
-/// Canonicalize memref.tensor_load + memref.buffer_cast to memref.cast when
-/// type mismatches prevent `BufferCastOp::fold` to kick in.
-struct TensorLoadToMemRef : public OpRewritePattern<BufferCastOp> {
-  using OpRewritePattern<BufferCastOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(BufferCastOp bufferCast,
-                                PatternRewriter &rewriter) const final {
-    auto tensorLoad = bufferCast.tensor().getDefiningOp<TensorLoadOp>();
-    // Bail unless we have a tensor_load + memref.buffer_cast with different
-    // types. `BufferCastOp::fold` handles the same type case.
-    if (!tensorLoad || tensorLoad.memref().getType() == bufferCast.getType())
-      return failure();
-    // If types are definitely not cast-compatible, bail.
-    if (!CastOp::areCastCompatible(tensorLoad.memref().getType(),
-                                   bufferCast.getType()))
-      return failure();
-
-    // We already know that the types are potentially cast-compatible. However
-    // in case the affine maps are different, we may need to use a copy if we go
-    // from dynamic to static offset or stride (the canonicalization cannot know
-    // at this point that it is really cast compatible).
-    auto isGuaranteedCastCompatible = [](MemRefType source, MemRefType target) {
-      int64_t sourceOffset, targetOffset;
-      SmallVector<int64_t, 4> sourceStrides, targetStrides;
-      if (failed(getStridesAndOffset(source, sourceStrides, sourceOffset)) ||
-          failed(getStridesAndOffset(target, targetStrides, targetOffset)))
-        return false;
-      auto dynamicToStatic = [](int64_t a, int64_t b) {
-        return a == MemRefType::getDynamicStrideOrOffset() &&
-               b != MemRefType::getDynamicStrideOrOffset();
-      };
-      if (dynamicToStatic(sourceOffset, targetOffset))
-        return false;
-      for (auto it : zip(sourceStrides, targetStrides))
-        if (dynamicToStatic(std::get<0>(it), std::get<1>(it)))
-          return false;
-      return true;
-    };
-
-    auto tensorLoadType = tensorLoad.memref().getType().dyn_cast<MemRefType>();
-    auto bufferCastType = bufferCast.getType().dyn_cast<MemRefType>();
-    if (tensorLoadType && bufferCastType &&
-        !isGuaranteedCastCompatible(tensorLoadType, bufferCastType)) {
-      MemRefType resultType = bufferCastType;
-      auto loc = bufferCast.getLoc();
-      SmallVector<Value, 4> dynamicOperands;
-      for (int i = 0; i < resultType.getRank(); ++i) {
-        if (resultType.getShape()[i] != ShapedType::kDynamicSize)
-          continue;
-        auto index = rewriter.createOrFold<arith::ConstantIndexOp>(loc, i);
-        Value size = rewriter.create<tensor::DimOp>(loc, tensorLoad, index);
-        dynamicOperands.push_back(size);
-      }
-      auto copy =
-          rewriter.create<memref::AllocOp>(loc, resultType, dynamicOperands);
-      rewriter.create<CopyOp>(loc, tensorLoad.memref(), copy);
-      rewriter.replaceOp(bufferCast, {copy});
-    } else
-      rewriter.replaceOpWithNewOp<CastOp>(bufferCast, bufferCast.getType(),
-                                          tensorLoad.memref());
-    return success();
-  }
-};
-
-} // namespace
-
-void BufferCastOp::getCanonicalizationPatterns(RewritePatternSet &results,
-                                               MLIRContext *context) {
-  results.add<BufferCast, TensorLoadToMemRef>(context);
-}
-
-//===----------------------------------------------------------------------===//
 // CastOp
 //===----------------------------------------------------------------------===//
 
 /// Determines whether MemRef_CastOp casts to a more dynamic version of the
 /// source memref. This is useful to to fold a memref.cast into a consuming op
 /// and implement canonicalization patterns for ops in different dialects that
 /// may consume the results of memref.cast operations. Such foldable memref.cast
 /// operations are typically inserted as `view` and `subview` ops are
bool CastOp::areCastCompatible(TypeRange inputs, TypeRange outputs) {
   return false;
 }
 
 OpFoldResult CastOp::fold(ArrayRef<Attribute> operands) {
   return succeeded(foldMemRefCast(*this)) ? getResult() : Value();
 }
 
 //===----------------------------------------------------------------------===//
-// CloneOp
-//===----------------------------------------------------------------------===//
-
-void CloneOp::getEffects(
-    SmallVectorImpl<SideEffects::EffectInstance<MemoryEffects::Effect>>
-        &effects) {
-  effects.emplace_back(MemoryEffects::Read::get(), input(),
-                       SideEffects::DefaultResource::get());
-  effects.emplace_back(MemoryEffects::Write::get(), output(),
-                       SideEffects::DefaultResource::get());
-  effects.emplace_back(MemoryEffects::Allocate::get(), output(),
-                       SideEffects::DefaultResource::get());
-}
-
-namespace {
-/// Merge the clone and its source (by converting the clone to a cast) when
-/// possible.
-struct SimplifyClones : public OpRewritePattern<CloneOp> {
-  using OpRewritePattern<CloneOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(CloneOp cloneOp,
-                                PatternRewriter &rewriter) const override {
-    if (cloneOp.use_empty()) {
-      rewriter.eraseOp(cloneOp);
-      return success();
-    }
-
-    Value source = cloneOp.input();
-
-    // This only finds dealloc operations for the immediate value. It should
-    // also consider aliases. That would also make the safety check below
-    // redundant.
-    llvm::Optional<Operation *> maybeCloneDeallocOp =
-        findDealloc(cloneOp.output());
-    // Skip if either of them has > 1 deallocate operations.
-    if (!maybeCloneDeallocOp.hasValue())
-      return failure();
-    llvm::Optional<Operation *> maybeSourceDeallocOp = findDealloc(source);
-    if (!maybeSourceDeallocOp.hasValue())
-      return failure();
-    Operation *cloneDeallocOp = *maybeCloneDeallocOp;
-    Operation *sourceDeallocOp = *maybeSourceDeallocOp;
-
-    // If both are deallocated in the same block, their in-block lifetimes
-    // might not fully overlap, so we cannot decide which one to drop.
-    if (cloneDeallocOp && sourceDeallocOp &&
-        cloneDeallocOp->getBlock() == sourceDeallocOp->getBlock())
-      return failure();
-
-    Block *currentBlock = cloneOp->getBlock();
-    Operation *redundantDealloc = nullptr;
-    if (cloneDeallocOp && cloneDeallocOp->getBlock() == currentBlock) {
-      redundantDealloc = cloneDeallocOp;
-    } else if (sourceDeallocOp && sourceDeallocOp->getBlock() == currentBlock) {
-      redundantDealloc = sourceDeallocOp;
-    }
-
-    if (!redundantDealloc)
-      return failure();
-
-    // Safety check that there are no other deallocations inbetween
-    // cloneOp and redundantDealloc, as otherwise we might deallocate an alias
-    // of source before the uses of the clone. With alias information, we could
-    // restrict this to only fail of the dealloc's operand is an alias
-    // of the source.
-    for (Operation *pos = cloneOp->getNextNode(); pos != redundantDealloc;
-         pos = pos->getNextNode()) {
-      auto effectInterface = dyn_cast<MemoryEffectOpInterface>(pos);
-      if (!effectInterface)
-        continue;
-      if (effectInterface.hasEffect<MemoryEffects::Free>())
-        return failure();
-    }
-
-    rewriter.replaceOpWithNewOp<memref::CastOp>(cloneOp, cloneOp.getType(),
-                                                source);
-    rewriter.eraseOp(redundantDealloc);
-    return success();
-  }
-};
-
-} // end anonymous namespace.
-
-void CloneOp::getCanonicalizationPatterns(OwningRewritePatternList &results,
-                                          MLIRContext *context) {
-  results.insert<SimplifyClones>(context);
-}
-
-OpFoldResult CloneOp::fold(ArrayRef<Attribute> operands) {
-  return succeeded(foldMemRefCast(*this)) ? getResult() : Value();
-}
-
-//===----------------------------------------------------------------------===//
 // DeallocOp
 //===----------------------------------------------------------------------===//
 
 LogicalResult DeallocOp::fold(ArrayRef<Attribute> cstOperands,
                               SmallVectorImpl<OpFoldResult> &results) {
   /// dealloc(memrefcast) -> dealloc
   return foldMemRefCast(*this);
 }
LogicalResult matchAndRewrite(DimOp dim,
     Value load = rewriter.create<LoadOp>(loc, reshape.shape(), dim.index());
     if (load.getType() != dim.getType())
       load = rewriter.create<arith::IndexCastOp>(loc, dim.getType(), load);
     rewriter.replaceOp(dim, load);
     return success();
   }
 };
 
-/// Fold dim of a cast into the dim of the source of the memref cast.
-struct DimOfCastOp : public OpRewritePattern<DimOp> {
-  using OpRewritePattern<DimOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(DimOp dimOp,
-                                PatternRewriter &rewriter) const override {
-    auto castOp = dimOp.source().getDefiningOp<BufferCastOp>();
-    if (!castOp)
-      return failure();
-    Value newSource = castOp.getOperand();
-    rewriter.replaceOpWithNewOp<tensor::DimOp>(dimOp, newSource, dimOp.index());
-    return success();
-  }
-};
 } // end anonymous namespace.
 
 void DimOp::getCanonicalizationPatterns(RewritePatternSet &results,
                                         MLIRContext *context) {
-  results.add<DimOfMemRefReshape, DimOfCastOp>(context);
+  results.add<DimOfMemRefReshape>(context);
 }
 
 // ---------------------------------------------------------------------------
 // DmaStartOp
 // ---------------------------------------------------------------------------
 
 void DmaStartOp::build(OpBuilder &builder, OperationState &result,
                        Value srcMemRef, ValueRange srcIndices, Value destMemRef,

 
 OpFoldResult LoadOp::fold(ArrayRef<Attribute> cstOperands) {
   /// load(memrefcast) -> load
   if (succeeded(foldMemRefCast(*this)))
     return getResult();
   return OpFoldResult();
 }
 
-namespace {
-/// Fold a load on a buffer_cast operation into an tensor.extract on the
-/// corresponding tensor.
-struct LoadOfBufferCast : public OpRewritePattern<LoadOp> {
-  using OpRewritePattern<LoadOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(LoadOp load,
-                                PatternRewriter &rewriter) const override {
-    auto buffercast = load.memref().getDefiningOp<BufferCastOp>();
-    if (!buffercast)
-      return failure();
-
-    rewriter.replaceOpWithNewOp<tensor::ExtractOp>(load, buffercast.tensor(),
-                                                   load.indices());
-    return success();
-  }
-};
-} // end anonymous namespace.
-
-void LoadOp::getCanonicalizationPatterns(RewritePatternSet &results,
-                                         MLIRContext *context) {
-  results.add<LoadOfBufferCast>(context);
-}
-
 //===----------------------------------------------------------------------===//
 // PrefetchOp
 //===----------------------------------------------------------------------===//
 
 static void print(OpAsmPrinter &p, PrefetchOp op) {
   p << " " << op.memref() << '[';
   p.printOperands(op.indices());
   p << ']' << ", " << (op.isWrite() ? "write" : "read");
if (resultShapedType.hasStaticShape() &&
       resultShapedType == sourceShapedType) {
     return getViewSource();
   }
 
   return {};
 }
 
 //===----------------------------------------------------------------------===//
-// TensorLoadOp
-//===----------------------------------------------------------------------===//
-
-OpFoldResult TensorLoadOp::fold(ArrayRef<Attribute>) {
-  if (auto bufferCast = memref().getDefiningOp<BufferCastOp>())
-    // Approximate alias analysis by conservatively folding only when no there
-    // is no interleaved operation.
-    if (bufferCast->getBlock() == this->getOperation()->getBlock() &&
-        bufferCast->getNextNode() == this->getOperation())
-      return bufferCast.tensor();
-  return {};
-}
-
-namespace {
-struct DimOfTensorLoadFolder : public OpRewritePattern<tensor::DimOp> {
-  using OpRewritePattern<tensor::DimOp>::OpRewritePattern;
-
-  LogicalResult matchAndRewrite(tensor::DimOp dimOp,
-                                PatternRewriter &rewriter) const override {
-    auto tensorLoadOp = dimOp.source().getDefiningOp<TensorLoadOp>();
-    if (!tensorLoadOp)
-      return failure();
-
-    rewriter.replaceOpWithNewOp<DimOp>(dimOp, tensorLoadOp.memref(),
-                                       dimOp.index());
-    return success();
-  }
-};
-} // namespace
-
-void TensorLoadOp::getCanonicalizationPatterns(RewritePatternSet &results,
-                                               MLIRContext *context) {
-  results.add<DimOfTensorLoadFolder>(context);
-}
-
-//===----------------------------------------------------------------------===//
 // TransposeOp
 //===----------------------------------------------------------------------===//
 
 /// Build a strided memref type by applying `permutationMap` tp `memRefType`.
 static MemRefType inferTransposeResultType(MemRefType memRefType,
                                            AffineMap permutationMap) {
   auto rank = memRefType.getRank();
   auto originalSizes = memRefType.getShape();

mlir/lib/Dialect/SCF/CMakeLists.txt

 add_mlir_dialect_library(MLIRSCF
   SCF.cpp
 
   ADDITIONAL_HEADER_DIRS
   ${MLIR_MAIN_INCLUDE_DIR}/mlir/LoopOps
 
   DEPENDS
   MLIRSCFOpsIncGen
 
   LINK_LIBS PUBLIC
   MLIRArithmetic
+  MLIRBufferization
   MLIRIR
   MLIRLoopLikeInterface
   MLIRMemRef
   MLIRSideEffectInterfaces
   MLIRStandard
   )
 
 add_subdirectory(Transforms)

mlir/lib/Dialect/SCF/SCF.cpp

 //===- SCF.cpp - Structured Control Flow Operations -----------------------===//
 //
 // 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/SCF.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 
for (auto it : llvm::zip(op.getIterOpOperands(), op.getResults())) {
           op.getLoc(), incomingCast.dest().getType(), replacements[returnIdx]);
       rewriter.replaceOp(op, replacements);
       return success();
     }
     return failure();
   }
 };
 
-/// Canonicalize the iter_args of an scf::ForOp that involve a tensor_load and
-/// for which only the last loop iteration is actually visible outside of the
-/// loop. The canonicalization looks for a pattern such as:
+/// Canonicalize the iter_args of an scf::ForOp that involve a
+/// `bufferization.to_tensor` and for which only the last loop iteration is
+/// actually visible outside of the loop. The canonicalization looks for a
+/// pattern such as:
 /// ```
 ///    %t0 = ... : tensor_type
 ///    %0 = scf.for ... iter_args(%bb0 : %t0) -> (tensor_type) {
 ///      ...
 ///      // %m is either buffer_cast(%bb00) or defined above the loop
 ///      %m... : memref_type
 ///      ... // uses of %m with potential inplace updates
-///      %new_tensor = tensor_load %m : memref_type
+///      %new_tensor = bufferization.to_tensor %m : memref_type
 ///      ...
 ///      scf.yield %new_tensor : tensor_type
 ///    }
 /// ```
 ///
 /// `%bb0` may have either 0 or 1 use. If it has 1 use it must be exactly a
-/// `%m = buffer_cast %bb0` op that feeds into the yielded `tensor_load`
-/// op.
+/// `%m = buffer_cast %bb0` op that feeds into the yielded
+/// `bufferization.to_tensor` op.
 ///
 /// If no aliasing write to the memref `%m`, from which `%new_tensor`is loaded,
-/// occurs between tensor_load and yield then the value %0 visible outside of
-/// the loop is the last `tensor_load` produced in the loop.
+/// occurs between `bufferization.to_tensor and yield then the value %0
+/// visible outside of the loop is the last `bufferization.to_tensor`
+/// produced in the loop.
 ///
 /// For now, we approximate the absence of aliasing by only supporting the case
-/// when the tensor_load is the operation immediately preceding the yield.
-///
+/// when the bufferization.to_tensor is the operation immediately preceding
+/// the yield.
+//
 /// The canonicalization rewrites the pattern as:
 /// ```
 ///    // %m is either a buffer_cast or defined above
 ///    %m... : memref_type
 ///    scf.for ... iter_args(%bb0 : %t0) -> (tensor_type) {
 ///      ... // uses of %m with potential inplace updates
 ///      scf.yield %bb0: tensor_type
 ///    }
-///    %0 = tensor_load %m : memref_type
+///    %0 = bufferization.to_tensor %m : memref_type
 /// ```
 ///
 /// A later bbArg canonicalization will further rewrite as:
 /// ```
 ///    // %m is either a buffer_cast or defined above
 ///    %m... : memref_type
 ///    scf.for ... { // no iter_args
 ///      ... // uses of %m with potential inplace updates
 ///    }
-///    %0 = tensor_load %m : memref_type
+///    %0 = bufferization.to_tensor %m : memref_type
 /// ```
 struct LastTensorLoadCanonicalization : public OpRewritePattern<ForOp> {
   using OpRewritePattern<ForOp>::OpRewritePattern;
 
   LogicalResult matchAndRewrite(ForOp forOp,
                                 PatternRewriter &rewriter) const override {
     assert(std::next(forOp.region().begin()) == forOp.region().end() &&
            "unexpected multiple blocks");
 
     Location loc = forOp.getLoc();
     DenseMap<Value, Value> replacements;
     for (BlockArgument bbArg : forOp.getRegionIterArgs()) {
       unsigned idx = bbArg.getArgNumber() - /*numIv=*/1;
       auto yieldOp = cast<scf::YieldOp>(forOp.region().front().getTerminator());
       Value yieldVal = yieldOp->getOperand(idx);
-      auto tensorLoadOp = yieldVal.getDefiningOp<memref::TensorLoadOp>();
+      auto tensorLoadOp = yieldVal.getDefiningOp<bufferization::ToTensorOp>();
       bool isTensor = bbArg.getType().isa<TensorType>();
 
-      memref::BufferCastOp bufferCastOp;
+      bufferization::ToMemrefOp tensorToMemref;
       // Either bbArg has no use or it has a single buffer_cast use.
       if (bbArg.hasOneUse())
-        bufferCastOp =
-            dyn_cast<memref::BufferCastOp>(*bbArg.getUsers().begin());
-      if (!isTensor || !tensorLoadOp || (!bbArg.use_empty() && !bufferCastOp))
+        tensorToMemref =
+            dyn_cast<bufferization::ToMemrefOp>(*bbArg.getUsers().begin());
+      if (!isTensor || !tensorLoadOp || (!bbArg.use_empty() && !tensorToMemref))
         continue;
-      // If bufferCastOp is present, it must feed into the `tensorLoadOp`.
-      if (bufferCastOp && tensorLoadOp.memref() != bufferCastOp)
+      // If tensorToMemref is present, it must feed into the `ToTensorOp`.
+      if (tensorToMemref && tensorLoadOp.memref() != tensorToMemref)
         continue;
       // TODO: Any aliasing write of tensorLoadOp.memref() nested under `forOp`
-      // must be before `tensorLoadOp` in the block so that the lastWrite
+      // must be before `ToTensorOp` in the block so that the lastWrite
       // property is not subject to additional side-effects.
-      // For now, we only support the case when tensorLoadOp appears immediately
-      // before the terminator.
+      // For now, we only support the case when ToTensorOp appears
+      // immediately before the terminator.
       if (tensorLoadOp->getNextNode() != yieldOp)
         continue;
 
-      // Clone the optional bufferCastOp before forOp.
-      if (bufferCastOp) {
+      // Clone the optional tensorToMemref before forOp.
+      if (tensorToMemref) {
         rewriter.setInsertionPoint(forOp);
-        rewriter.replaceOpWithNewOp<memref::BufferCastOp>(
-            bufferCastOp, bufferCastOp.memref().getType(),
-            bufferCastOp.tensor());
+        rewriter.replaceOpWithNewOp<bufferization::ToMemrefOp>(
+            tensorToMemref, tensorToMemref.memref().getType(),
+            tensorToMemref.tensor());
       }
 
       // Clone the tensorLoad after forOp.
       rewriter.setInsertionPointAfter(forOp);
-      Value newTensorLoad =
-          rewriter.create<memref::TensorLoadOp>(loc, tensorLoadOp.memref());
+      Value newTensorLoad = rewriter.create<bufferization::ToTensorOp>(
+          loc, tensorLoadOp.memref());
       Value forOpResult = forOp.getResult(bbArg.getArgNumber() - /*iv=*/1);
       replacements.insert(std::make_pair(forOpResult, newTensorLoad));
 
       // Make the terminator just yield the bbArg, the old tensorLoadOp + the
       // old bbArg (that is now directly yielded) will canonicalize away.
       rewriter.startRootUpdate(yieldOp);
       yieldOp.setOperand(idx, bbArg);
       rewriter.finalizeRootUpdate(yieldOp);

mlir/lib/Dialect/SCF/Transforms/Bufferize.cpp

 //===- Bufferize.cpp - scf bufferize pass ---------------------------------===//
 //
 // 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/Transforms/Bufferize.h"
 #include "PassDetail.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/Passes.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/SCF/Transforms.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Transforms/DialectConversion.h"
 
 using namespace mlir;

mlir/lib/Dialect/SCF/Transforms/PassDetail.h

 void registerDialect(DialectRegistry &registry);
 
 class AffineDialect;
 
 namespace arith {
 class ArithmeticDialect;
 } // end namespace arith
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace memref {
 class MemRefDialect;
 } // end namespace memref
 
 namespace tensor {
 class TensorDialect;
 } // end namespace tensor
 
 #define GEN_PASS_CLASSES
 #include "mlir/Dialect/SCF/Passes.h.inc"
 
 } // end namespace mlir
 
 #endif // DIALECT_LOOPOPS_TRANSFORMS_PASSDETAIL_H_

mlir/lib/Dialect/Shape/Transforms/Bufferize.cpp

 //====----- Bufferize.cpp - Bufferization of shape ops  ---------*- 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
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Transforms/Bufferize.h"
 #include "PassDetail.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/Shape/Transforms/Passes.h"
 #include "mlir/Pass/Pass.h"
 
 using namespace mlir;
 
 namespace {
 struct ShapeBufferizePass : public ShapeBufferizeBase<ShapeBufferizePass> {

mlir/lib/Dialect/Shape/Transforms/PassDetail.h

 //===- PassDetail.h - Shape Pass class details ------------------*- 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 DIALECT_SHAPE_TRANSFORMS_PASSDETAIL_H_
 #define DIALECT_SHAPE_TRANSFORMS_PASSDETAIL_H_
 
 #include "mlir/Pass/Pass.h"
 
 namespace mlir {
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace memref {
 class MemRefDialect;
 } // end namespace memref
 
 #define GEN_PASS_CLASSES
 #include "mlir/Dialect/Shape/Transforms/Passes.h.inc"
 
 } // end namespace mlir
 
 #endif // DIALECT_SHAPE_TRANSFORMS_PASSDETAIL_H_

mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorConversion.cpp

 //===- SparseTensorConversion.cpp - Sparse tensor primitives conversion ---===//
 //
 // 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
 //
 //===----------------------------------------------------------------------===//
 //
 // Convert sparse tensor primitives to calls into a runtime support library.
 // Note that this is a current implementation choice to keep the conversion
 // simple. In principle, these primitives could also be converted to actual
 // elaborate IR code that implements the primitives on the selected sparse
 // tensor storage schemes.
 //
 //===----------------------------------------------------------------------===//
 
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
if (!encDst && encSrc) {
       rewriter.setInsertionPointToEnd(before);
       Value cond = genGetNextCall(rewriter, op, iter, ind, elemPtr);
       rewriter.create<scf::ConditionOp>(loc, cond, before->getArguments());
       Block *after = rewriter.createBlock(&whileOp.after(), {}, noTypes);
       rewriter.setInsertionPointToStart(after);
       insertScalarIntoDenseTensor(rewriter, loc, elemPtr, dst, rank, ind);
       rewriter.create<scf::YieldOp>(loc);
       rewriter.setInsertionPointAfter(whileOp);
-      rewriter.replaceOpWithNewOp<memref::TensorLoadOp>(op, resType, dst);
+      rewriter.replaceOpWithNewOp<bufferization::ToTensorOp>(op, resType, dst);
       return success();
     }
     if (!encDst && !encSrc) {
       // dense => dense
       return failure();
     }
     // This is a dense => sparse conversion or a sparse constant in COO =>
     // sparse conversion, which is handled as follows:

mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorPasses.cpp

 //===- SparseTensorPasses.cpp - Pass for autogen sparse tensor code -------===//
 //
 // 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/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
 #include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
 #include "mlir/Dialect/StandardOps/Transforms/FuncConversions.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 
 using namespace mlir;
void runOnOperation() override {
     target.addDynamicallyLegalOp<tensor::CastOp>([&](tensor::CastOp op) {
       return converter.isLegal(op.getOperand().getType());
     });
     // The following operations and dialects may be introduced by the
     // rewriting rules, and are therefore marked as legal.
     target.addLegalOp<arith::CmpFOp, arith::CmpIOp, arith::ConstantOp,
                       arith::IndexCastOp, linalg::FillOp, linalg::YieldOp,
                       tensor::ExtractOp>();
-    target.addLegalDialect<LLVM::LLVMDialect, memref::MemRefDialect,
-                           scf::SCFDialect>();
+    target
+        .addLegalDialect<bufferization::BufferizationDialect, LLVM::LLVMDialect,
+                         memref::MemRefDialect, scf::SCFDialect>();
     // Populate with rules and apply rewriting rules.
     populateFuncOpTypeConversionPattern(patterns, converter);
     populateCallOpTypeConversionPattern(patterns, converter);
     populateSparseTensorConversionPatterns(converter, patterns);
     if (failed(applyPartialConversion(getOperation(), target,
                                       std::move(patterns))))
       signalPassFailure();
   }

mlir/lib/Dialect/SparseTensor/Transforms/Sparsification.cpp

 //===- Sparsification.cpp - Implementation of sparsification --------------===//
 //
 // 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 converting sparse tensor types to actual sparse code.
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h"
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/SCF/Transforms.h"
 #include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
static Value genOutputBuffer(CodeGen &codegen, PatternRewriter &rewriter,
                              ArrayRef<Value> args) {
   Location loc = op.getLoc();
   Value tensor = op.getOutputOperand(0)->get();
   // The output tensor simply could materialize from the buffer that will
   // be generated for the tensor present in the outs() clause. This has
   // the major advantage that the sparse kernel only updates the nonzero
   // positions for the output tensor.
   if (isInPlace(tensor))
-    return rewriter.create<memref::BufferCastOp>(loc, denseTp, tensor);
+    return rewriter.create<bufferization::ToMemrefOp>(loc, denseTp, tensor);
   // By default, a new buffer is allocated which is initialized to the
   // tensor defined in the outs() clause. This is always correct but
   // introduces a dense initialization component that may negatively
   // impact the running complexity of the sparse kernel. If the tensor
   // materializes into the computation, we need to preserve the zero
   // initialization assumption of all sparse output buffers.
   if (isMaterializing(tensor)) {
     Type tp = denseTp.getElementType();
     Value alloc = rewriter.create<memref::AllocOp>(loc, denseTp, args);
     Value zero =
         rewriter.create<arith::ConstantOp>(loc, tp, rewriter.getZeroAttr(tp));
     rewriter.create<linalg::FillOp>(loc, zero, alloc);
     return alloc;
   }
-  Value init = rewriter.create<memref::BufferCastOp>(loc, denseTp, tensor);
+  Value init = rewriter.create<bufferization::ToMemrefOp>(loc, denseTp, tensor);
   Value alloc = rewriter.create<memref::AllocOp>(loc, denseTp, args);
   rewriter.create<memref::CopyOp>(loc, init, alloc);
   return alloc;
 }
 
 /// Local bufferization of all dense and sparse data structures.
 /// This code enables testing the first prototype sparse compiler.
 // TODO: replace this with a proliferated bufferization strategy
for (OpOperand *t : op.getInputAndOutputOperands()) {
     // Sparse inputs use sparse primitives to obtain the values.
     // We also accept in-place all-dense annotated "sparse" outputs.
     Type elementType = getElementTypeOrSelf(t->get().getType());
     if (!enc) {
       // Non-annotated dense tensors.
       auto denseTp = MemRefType::get(shape, elementType);
       if (tensor < op.getNumInputs())
         codegen.buffers[tensor] =
-            rewriter.create<memref::BufferCastOp>(loc, denseTp, t->get());
+            rewriter.create<bufferization::ToMemrefOp>(loc, denseTp, t->get());
       else
         codegen.buffers[tensor] =
             genOutputBuffer(codegen, rewriter, op, denseTp, args);
     } else if (t == codegen.sparseOut) {
       // True sparse output needs a lexIdx array.
       Value rank = rewriter.create<arith::ConstantIndexOp>(loc, op.getRank(t));
       auto dynShape = {ShapedType::kDynamicSize};
       auto memTp = MemRefType::get(dynShape, rewriter.getIndexType());
if (getSparseTensorEncoding(resType)) {
     // The sparse tensor rematerializes from the original sparse tensor's
     // underlying sparse storage format.
     rewriter.replaceOpWithNewOp<LoadOp>(op, resType, lhs->get(),
                                         codegen.sparseOut == lhs);
   } else {
     // To rematerialize an non-annotated tensor, simply load it
     // from the bufferized value.
     Value val = codegen.buffers.back(); // value array
-    rewriter.replaceOpWithNewOp<memref::TensorLoadOp>(op, resType, val);
+    rewriter.replaceOpWithNewOp<bufferization::ToTensorOp>(op, resType, val);
   }
 }
 
 //===----------------------------------------------------------------------===//
 // Sparse compiler rewriting methods.
 //===----------------------------------------------------------------------===//
 
 namespace {

mlir/lib/Dialect/StandardOps/Transforms/Bufferize.cpp

 //===- Bufferize.cpp - Bufferization for std ops --------------------------===//
 //
 // 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 bufferization of std ops.
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Transforms/Bufferize.h"
 #include "PassDetail.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/StandardOps/Transforms/Passes.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/Transforms/DialectConversion.h"
 

mlir/lib/Dialect/StandardOps/Transforms/FuncBufferize.cpp

 //===- Bufferize.cpp - Bufferization for std ops --------------------------===//
 //
 // 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 bufferization of std.func's and std.call's.
 //
 //===----------------------------------------------------------------------===//
 
 #include "PassDetail.h"
-#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/StandardOps/Transforms/FuncConversions.h"
 #include "mlir/Dialect/StandardOps/Transforms/Passes.h"
 #include "mlir/Transforms/Bufferize.h"
 #include "mlir/Transforms/DialectConversion.h"
 
 using namespace mlir;
 
target.addDynamicallyLegalOp<FuncOp>([&](FuncOp op) {
              typeConverter.isLegal(&op.getBody());
     });
     populateCallOpTypeConversionPattern(patterns, typeConverter);
     target.addDynamicallyLegalOp<CallOp>(
         [&](CallOp op) { return typeConverter.isLegal(op); });
 
     populateBranchOpInterfaceTypeConversionPattern(patterns, typeConverter);
     populateReturnOpTypeConversionPattern(patterns, typeConverter);
-    target.addLegalOp<ModuleOp, memref::TensorLoadOp, memref::BufferCastOp>();
+    target.addLegalOp<ModuleOp, bufferization::ToTensorOp,
+                      bufferization::ToMemrefOp>();
 
     target.markUnknownOpDynamicallyLegal([&](Operation *op) {
       return isNotBranchOpInterfaceOrReturnLikeOp(op) ||
              isLegalForBranchOpInterfaceTypeConversionPattern(op,
                                                               typeConverter) ||
              isLegalForReturnOpTypeConversionPattern(op, typeConverter);
     });
 

mlir/lib/Dialect/StandardOps/Transforms/PassDetail.h

 
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Pass/Pass.h"
 
 namespace mlir {
 
 class AtomicRMWOp;
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace memref {
 class MemRefDialect;
 } // end namespace memref
 
 #define GEN_PASS_CLASSES
 #include "mlir/Dialect/StandardOps/Transforms/Passes.h.inc"
 
 } // end namespace mlir
 
 #endif // DIALECT_STANDARD_TRANSFORMS_PASSDETAIL_H_

mlir/lib/Dialect/StandardOps/Transforms/TensorConstantBufferize.cpp

 //===- Bufferize.cpp - Bufferization for std ops --------------------------===//
 //
 // 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 bufferization of tensor-valued arith.constant ops.
 //
 //===----------------------------------------------------------------------===//
 
 #include "PassDetail.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/StandardOps/Transforms/Passes.h"
 #include "mlir/IR/BlockAndValueMapping.h"
 #include "mlir/Transforms/BufferUtils.h"
 #include "mlir/Transforms/Bufferize.h"
 #include "mlir/Transforms/DialectConversion.h"
 

mlir/lib/Dialect/Tensor/Transforms/Bufferize.cpp

 //===- Bufferize.cpp - Bufferization for `tensor` dialect ops -------------===//
 //
 // 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 bufferization of `tensor` dialect ops
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Transforms/Bufferize.h"
 #include "PassDetail.h"
 #include "mlir/Dialect/Arithmetic/IR/Arithmetic.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Tensor/Transforms/Passes.h"
 #include "mlir/Transforms/DialectConversion.h"
 
 using namespace mlir;

mlir/lib/Dialect/Tensor/Transforms/PassDetail.h

 //===- PassDetail.h - GPU Pass class details --------------------*- 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 DIALECT_TENSOR_TRANSFORMS_PASSDETAIL_H_
 #define DIALECT_TENSOR_TRANSFORMS_PASSDETAIL_H_
 
 #include "mlir/Pass/Pass.h"
 
 namespace mlir {
 
+namespace bufferization {
+class BufferizationDialect;
+} // end namespace bufferization
+
 namespace memref {
 class MemRefDialect;
 } // end namespace memref
 
 namespace scf {
 class SCFDialect;
 } // end namespace scf
 
 #define GEN_PASS_CLASSES
 #include "mlir/Dialect/Tensor/Transforms/Passes.h.inc"
 
 } // end namespace mlir
 
 #endif // DIALECT_TENSOR_TRANSFORMS_PASSDETAIL_H_

mlir/lib/Dialect/Vector/VectorTransferPermutationMapRewritePatterns.cpp

 //===- VectorTransferPermutationMapRewritePatterns.cpp - Xfer map rewrite -===//
 //
 // 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 rewrite patterns for the permutation_map attribute of
 // vector.transfer operations.
 //
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Vector/VectorTransforms.h"
 #include "mlir/Interfaces/VectorInterfaces.h"
 
 using namespace mlir;
 using namespace mlir::vector;
 
 /// Transpose a vector transfer op's `in_bounds` attribute according to given
 /// indices.

mlir/lib/Transforms/BufferDeallocation.cpp

 // However, structured control-flow loops are fully supported. Furthermore, it
 // doesn't accept functions which return buffers already.
 //
 //===----------------------------------------------------------------------===//
 
 #include "PassDetail.h"
 
 #include "mlir/Dialect/Bufferization/IR/AllocationOpInterface.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/IR/Operation.h"
 #include "mlir/Interfaces/ControlFlowInterfaces.h"
 #include "mlir/Interfaces/LoopLikeInterface.h"
 #include "mlir/Pass/Pass.h"
 #include "mlir/Transforms/BufferUtils.h"
 #include "mlir/Transforms/Passes.h"
if (it != aliasToAllocations.end()) {
       builder.create<memref::DeallocOp>(alloc.getLoc(), alloc);
     }
     return success();
   }
 
   /// Builds a clone operation compatible with the given allocation value. If
   /// there is no registered AllocationOpInterface implementation for the given
   /// value (e.g. in the case of a function parameter), this method builds a
-  /// memref::CloneOp.
+  /// bufferization::CloneOp.
   FailureOr<Value> buildClone(Operation *op, Value alloc) {
     OpBuilder builder(op);
     auto it = aliasToAllocations.find(alloc);
     if (it != aliasToAllocations.end()) {
       // Call the allocation op interface to build a supported and
       // compatible clone operation.
       auto clone = it->second.buildClone(builder, alloc);
       if (clone)
         return *clone;
       return (LogicalResult)(op->emitError()
                              << "allocations without compatible clone ops "
                                 "are not supported");
     }
     // Build a "default" CloneOp for unknown allocation sources.
-    return builder.create<memref::CloneOp>(alloc.getLoc(), alloc).getResult();
+    return builder.create<bufferization::CloneOp>(alloc.getLoc(), alloc)
+        .getResult();
   }
 
   /// The dominator info to find the appropriate start operation to move the
   /// allocs.
   DominanceInfo dominators;
 
   /// The post dominator info to move the dependent allocs in the right
   /// position.
   PostDominanceInfo postDominators;
 
   /// Stores already cloned buffers to avoid additional clones of clones.
   ValueSetT clonedValues;
 
   /// Maps aliases to their source allocation interfaces (inverse mapping).
   AliasAllocationMapT aliasToAllocations;
 };
 
 //===----------------------------------------------------------------------===//
 // BufferDeallocationPass
 //===----------------------------------------------------------------------===//
 
-template <typename T>
 struct DefaultAllocationInterface
     : public bufferization::AllocationOpInterface::ExternalModel<
-          DefaultAllocationInterface<T>, T> {
+          DefaultAllocationInterface, memref::AllocOp> {
   static Optional<Operation *> buildDealloc(OpBuilder &builder, Value alloc) {
     return builder.create<memref::DeallocOp>(alloc.getLoc(), alloc)
         .getOperation();
   }
+  static Optional<Value> buildClone(OpBuilder &builder, Value alloc) {
+    return builder.create<bufferization::CloneOp>(alloc.getLoc(), alloc)
+        .getResult();
+  }
 };
 
 /// The actual buffer deallocation pass that inserts and moves dealloc nodes
 /// into the right positions. Furthermore, it inserts additional clones if
 /// necessary. It uses the algorithm described at the top of the file.
 struct BufferDeallocationPass : BufferDeallocationBase<BufferDeallocationPass> {
   void getDependentDialects(DialectRegistry &registry) const override {
+    registry.insert<bufferization::BufferizationDialect>();
     registry.insert<memref::MemRefDialect>();
-    registry.addOpInterface<memref::AllocOp,
-                            DefaultAllocationInterface<memref::AllocOp>>();
-    registry.addOpInterface<memref::CloneOp,
-                            DefaultAllocationInterface<memref::CloneOp>>();
+    registry.addOpInterface<memref::AllocOp, DefaultAllocationInterface>();
   }
 
   void runOnFunction() override {
     // Ensure that there are supported loops only.
     FuncOp func = getFunction();
     Backedges backedges(func);
     if (backedges.size()) {
       func.emitError("Only structured control-flow loops are supported.");

mlir/lib/Transforms/Bufferize.cpp

 //===- Bufferize.cpp - Bufferization utilities ----------------------------===//
 //
 // 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/Transforms/Bufferize.h"
 #include "PassDetail.h"
-#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/IR/Operation.h"
 #include "mlir/Transforms/Passes.h"
 
 using namespace mlir;
 
 //===----------------------------------------------------------------------===//
 // BufferizeTypeConverter
 //===----------------------------------------------------------------------===//
 
 static Value materializeTensorLoad(OpBuilder &builder, TensorType type,
                                    ValueRange inputs, Location loc) {
   assert(inputs.size() == 1);
   assert(inputs[0].getType().isa<BaseMemRefType>());
-  return builder.create<memref::TensorLoadOp>(loc, type, inputs[0]);
+  return builder.create<bufferization::ToTensorOp>(loc, type, inputs[0]);
 }
 
 /// Registers conversions into BufferizeTypeConverter
 BufferizeTypeConverter::BufferizeTypeConverter() {
   // Keep all types unchanged.
   addConversion([](Type type) { return type; });
   // Convert RankedTensorType to MemRefType.
   addConversion([](RankedTensorType type) -> Type {
     return MemRefType::get(type.getShape(), type.getElementType());
   });
   // Convert UnrankedTensorType to UnrankedMemRefType.
   addConversion([](UnrankedTensorType type) -> Type {
     return UnrankedMemRefType::get(type.getElementType(), 0);
   });
   addArgumentMaterialization(materializeTensorLoad);
   addSourceMaterialization(materializeTensorLoad);
   addTargetMaterialization([](OpBuilder &builder, BaseMemRefType type,
                               ValueRange inputs, Location loc) -> Value {
     assert(inputs.size() == 1);
     assert(inputs[0].getType().isa<TensorType>());
-    return builder.create<memref::BufferCastOp>(loc, type, inputs[0]);
+    return builder.create<bufferization::ToMemrefOp>(loc, type, inputs[0]);
   });
 }
 
 void mlir::populateBufferizeMaterializationLegality(ConversionTarget &target) {
-  target.addLegalOp<memref::TensorLoadOp, memref::BufferCastOp>();
+  target.addLegalOp<bufferization::ToTensorOp, bufferization::ToMemrefOp>();
 }
 
 namespace {
 // In a finalizing bufferize conversion, we know that all tensors have been
 // converted to memrefs, thus, this op becomes an identity.
-class BufferizeTensorLoadOp : public OpConversionPattern<memref::TensorLoadOp> {
+class BufferizeTensorLoadOp
+    : public OpConversionPattern<bufferization::ToTensorOp> {
 public:
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(memref::TensorLoadOp op, OpAdaptor adaptor,
+  matchAndRewrite(bufferization::ToTensorOp op, OpAdaptor adaptor,
                   ConversionPatternRewriter &rewriter) const override {
     rewriter.replaceOp(op, adaptor.memref());
     return success();
   }
 };
 } // namespace
 
 namespace {
 // In a finalizing bufferize conversion, we know that all tensors have been
 // converted to memrefs, thus, this op becomes an identity.
-class BufferizeCastOp : public OpConversionPattern<memref::BufferCastOp> {
+class BufferizeCastOp : public OpConversionPattern<bufferization::ToMemrefOp> {
 public:
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(memref::BufferCastOp op, OpAdaptor adaptor,
+  matchAndRewrite(bufferization::ToMemrefOp op, OpAdaptor adaptor,
                   ConversionPatternRewriter &rewriter) const override {
     rewriter.replaceOp(op, adaptor.tensor());
     return success();
   }
 };
 } // namespace
 
 void mlir::populateEliminateBufferizeMaterializationsPatterns(

mlir/lib/Transforms/CMakeLists.txt

add_mlir_library(MLIRTransforms
 
   DEPENDS
   MLIRStandardOpsIncGen
   MLIRTransformsPassIncGen
 
   LINK_LIBS PUBLIC
   MLIRAffine
   MLIRAnalysis
-  MLIRAllocationOpInterface
+  MLIRBufferization
   MLIRCopyOpInterface
   MLIRLoopLikeInterface
   MLIRMemRef
   MLIRSCF
   MLIRPass
   MLIRSupport
   MLIRTransformUtils
   MLIRVector
   )

mlir/test/Dialect/Arithmetic/bufferize.mlir

 // RUN: mlir-opt %s -arith-bufferize | FileCheck %s
 
 // CHECK-LABEL:   func @index_cast(
 // CHECK-SAME:  %[[TENSOR:.*]]: tensor<i32>, %[[SCALAR:.*]]: i32
 func @index_cast(%tensor: tensor<i32>, %scalar: i32) -> (tensor<index>, index) {
   %index_tensor = arith.index_cast %tensor : tensor<i32> to tensor<index>
   %index_scalar = arith.index_cast %scalar : i32 to index
   return %index_tensor, %index_scalar : tensor<index>, index
 }
-// CHECK:  %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<i32>
+// CHECK:  %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<i32>
 // CHECK-NEXT: %[[INDEX_MEMREF:.*]] = arith.index_cast %[[MEMREF]]
 // CHECK-SAME:   memref<i32> to memref<index>
-// CHECK-NEXT: %[[INDEX_TENSOR:.*]] = memref.tensor_load %[[INDEX_MEMREF]]
+// CHECK-NEXT: %[[INDEX_TENSOR:.*]] = bufferization.to_tensor %[[INDEX_MEMREF]]
 // CHECK: return %[[INDEX_TENSOR]]

mlir/test/Dialect/Bufferization/canonicalize.mlir

+// RUN: mlir-opt %s -canonicalize --split-input-file \
+// RUN:   -allow-unregistered-dialect |\
+// RUN: FileCheck %s
+
+// Basic folding of to_tensor(to_memref(t)) -> t
+// CHECK-LABEL: func @tensor_load_of_buffer_cast(
+func @tensor_load_of_buffer_cast(%arg0: tensor<?xf32>) -> tensor<?xf32> {
+  %0 = bufferization.to_memref %arg0 : memref<?xf32>
+  %1 = bufferization.to_tensor %0 : memref<?xf32>
+  return %1 : tensor<?xf32>
+}
+// CHECK-SAME:   %[[TENSOR:.*]]: tensor<?xf32>) -> tensor<?xf32> {
+// CHECK: return %[[TENSOR]]
+
+// -----
+
+// Basic folding of to_memref(to_tensor(m)) -> m
+// CHECK-LABEL: func @buffer_cast_of_tensor_load(
+func @buffer_cast_of_tensor_load(%arg0: memref<?xf32>) -> memref<?xf32> {
+  %0 = bufferization.to_tensor %arg0 : memref<?xf32>
+  %1 = bufferization.to_memref %0 : memref<?xf32>
+  return %1 : memref<?xf32>
+}
+// CHECK-SAME:   %[[MEMREF:.*]]: memref<?xf32>) -> memref<?xf32> {
+// CHECK: return %[[MEMREF]]
+
+// -----
+
+// If the memrefs are not the same type, don't fold them.
+// If the memrefs are not cast-compatible (e.g. different address space), don't
+// canonicalize them either.
+// CHECK-LABEL: func @no_fold_buffer_cast_of_tensor_load(
+//  CHECK-SAME:   %[[MEMREF_ADDRSPACE2:.*]]: memref<?xf32, 2>)
+//  CHECK-SAME:     -> memref<?xf32, 7> {
+//       CHECK: %[[TENSOR:.*]] = bufferization.to_tensor
+//  CHECK-SAME:   %[[MEMREF_ADDRSPACE2]] : memref<?xf32, 2>
+//       CHECK: %[[MEMREF_ADDRSPACE7:.*]] = bufferization.to_memref
+//  CHECK-SAME:   %[[TENSOR]] : memref<?xf32, 7>
+//       CHECK: return %[[MEMREF_ADDRSPACE7]]
+func @no_fold_buffer_cast_of_tensor_load(%arg0: memref<?xf32, 2>)
+    -> memref<?xf32, 7> {
+  %0 = bufferization.to_tensor %arg0 : memref<?xf32, 2>
+  %1 = bufferization.to_memref %0 : memref<?xf32, 7>
+  return %1 : memref<?xf32, 7>
+}
+
+// -----
+
+// CHECK-DAG: #[[$OFF_3:[a-z0-9]+]] = affine_map<(d0) -> (d0 + 3)>
+// CHECK-DAG: #[[$OFF_UNK:[a-z0-9]+]] = affine_map<(d0)[s0] -> (d0 + s0)>
+
+// If the memrefs are definitely cast-compatible, canonicalize to
+//            cast.
+// CHECK-LABEL: func @canonicalize_buffer_cast_of_tensor_load(
+//  CHECK-SAME:   %[[M:.*]]: memref<?xf32, #[[$OFF_3]]>)
+//  CHECK-SAME:     -> memref<?xf32, #[[$OFF_UNK]]> {
+//   CHECK-NOT: bufferization.to_tensor
+//   CHECK-NOT: bufferization.to_memref
+//       CHECK: %[[R:.*]] = memref.cast %[[M]]
+//  CHECK-SAME:   memref<?xf32, #[[$OFF_3]]> to memref<?xf32, #[[$OFF_UNK]]>
+//       CHECK: return %[[R]]
+func @canonicalize_buffer_cast_of_tensor_load(
+  %arg0: memref<?xf32, offset: 3, strides: [1]>)
+  -> memref<?xf32, offset: ?, strides: [1]>
+{
+  %0 = bufferization.to_tensor %arg0 : memref<?xf32, offset: 3, strides: [1]>
+  %1 = bufferization.to_memref %0 : memref<?xf32, offset: ?, strides: [1]>
+  return %1 : memref<?xf32, offset: ?, strides: [1]>
+}
+
+// -----
+
+// CHECK-DAG: #[[$OFF_UNK:[a-z0-9]+]] = affine_map<(d0)[s0] -> (d0 + s0)>
+// CHECK-DAG: #[[$OFF_3:[a-z0-9]+]] = affine_map<(d0) -> (d0 + 3)>
+
+// If the memrefs are potentially cast-compatible, canonicalize to
+//            copy.
+// CHECK-LABEL: func @canonicalize_buffer_cast_of_tensor_load_to_copy(
+func @canonicalize_buffer_cast_of_tensor_load_to_copy(
+  %arg0: memref<?xf32, offset: ?, strides: [1]>)
+  -> memref<?xf32, offset: 3, strides: [1]> {
+  %0 = bufferization.to_tensor %arg0 : memref<?xf32, offset: ?, strides: [1]>
+  %1 = bufferization.to_memref %0 : memref<?xf32, offset: 3, strides: [1]>
+  return %1 : memref<?xf32, offset: 3, strides: [1]>
+}
+// CHECK-SAME:   %[[M:.*]]: memref<?xf32, #[[$OFF_UNK]]>)
+// CHECK-SAME:     -> memref<?xf32, #[[$OFF_3]]> {
+//  CHECK-NOT: bufferization.to_tensor
+//  CHECK-NOT: bufferization.to_memref
+//      CHECK: %[[C0:.*]] = arith.constant 0 : index
+//      CHECK: %[[DIM:.*]] = memref.dim %[[M]], %[[C0]] : memref<?xf32, #[[$OFF_UNK]]>
+//      CHECK: %[[ALLOC:.*]] = memref.alloc(%[[DIM]]) : memref<?xf32, #[[$OFF_3]]>
+//      CHECK: memref.copy %[[M]], %[[ALLOC]]
+// CHECK-SAME:   memref<?xf32, #[[$OFF_UNK]]> to memref<?xf32, #[[$OFF_3]]>
+//      CHECK: return %[[ALLOC]]
+
+// -----
+
+
+// Basic folding of tensor.dim(to_tensor(m)) -> memref.dim(m).
+// CHECK-LABEL: func @dim_of_tensor_load(
+//  CHECK-SAME:     %[[MEMREF:[0-9a-z]*]]: memref<?xf32>
+//       CHECK:   %[[C0:.*]] = arith.constant 0
+//       CHECK:   %[[D:.*]] = memref.dim %[[MEMREF]], %[[C0]]
+//       CHECK:   return %[[D]] : index
+func @dim_of_tensor_load(%arg0: memref<?xf32>) -> index {
+  %c0 = arith.constant 0 : index
+  %0 = bufferization.to_tensor %arg0 : memref<?xf32>
+  %1 = tensor.dim %0, %c0 : tensor<?xf32>
+  return %1 : index
+}
+
+// -----
+
+// CHECK-LABEL: @clone_before_dealloc
+func @clone_before_dealloc(%arg0: memref<?xf32>) -> memref<?xf32> {
+  %0 = bufferization.clone %arg0 : memref<?xf32> to memref<?xf32>
+  memref.dealloc %arg0 : memref<?xf32>
+  return %0 : memref<?xf32>
+}
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+// CHECK-NEXT: return %[[ARG]]
+
+// -----
+
+// CHECK-LABEL: @clone_before_dealloc
+func @clone_before_dealloc(%arg0: memref<?xf32>) -> memref<?xf32> {
+  %0 = bufferization.clone %arg0 : memref<?xf32> to memref<?xf32>
+  "use"(%0) : (memref<?xf32>) -> ()
+  memref.dealloc %0 : memref<?xf32>
+  return %arg0 : memref<?xf32>
+}
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+// CHECK-NEXT: "use"(%arg0)
+// CHECK-NEXT: return %[[ARG]]
+
+// -----
+
+// CHECK-LABEL: @clone_after_cast
+func @clone_after_cast(%arg0: memref<?xf32>) -> memref<32xf32> {
+  %0 = memref.cast %arg0 : memref<?xf32> to memref<32xf32>
+  %1 = bufferization.clone %0 : memref<32xf32> to memref<32xf32>
+  return %1 : memref<32xf32>
+}
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+// CHECK-NEXT: bufferization.clone %[[ARG]] : memref<?xf32> to memref<32xf32>
+// CHECK-NOT: memref.cast
+
+// -----
+
+// CHECK-LABEL: @clone_and_cast
+func @clone_and_cast(%arg0: memref<?xf32>) -> memref<32xf32> {
+  %0 = bufferization.clone %arg0 : memref<?xf32> to memref<32xf32>
+  memref.dealloc %arg0 : memref<?xf32>
+  return %0 : memref<32xf32>
+}
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+// CHECK-NEXT: %[[RES:.*]] = memref.cast %[[ARG]]
+// CHECK-SAME:   memref<?xf32> to memref<32xf32>
+// CHECK-NEXT: return %[[RES]]
+
+// -----
+
+// CHECK-LABEL: @alias_is_freed
+func @alias_is_freed(%arg0 : memref<?xf32>) {
+  %0 = memref.cast %arg0 : memref<?xf32> to memref<32xf32>
+  %1 = bufferization.clone %0 : memref<32xf32> to memref<32xf32>
+  memref.dealloc %arg0 : memref<?xf32>
+  "use"(%1) : (memref<32xf32>) -> ()
+  memref.dealloc %1 : memref<32xf32>
+  return
+}
+// CHECK: bufferization.clone
+// CHECK: memref.dealloc
+// CHECK: memref.dealloc
+
+// -----
+
+// Verify SimplifyClones skips clones with multiple deallocations.
+// CHECK-LABEL: @clone_multiple_dealloc_of_source
+func @clone_multiple_dealloc_of_source(%arg0: memref<?xf32>) -> memref<?xf32> {
+  %0 = bufferization.clone %arg0 : memref<?xf32> to memref<?xf32>
+  "if_else"() ({
+    memref.dealloc %arg0 : memref<?xf32>
+    }, {
+    memref.dealloc %arg0 : memref<?xf32>
+    }) : () -> ()
+  return %0 : memref<?xf32>
+}
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+// CHECK-NEXT: %[[RES:.*]] = bufferization.clone %[[ARG]]
+// CHECK: memref.dealloc %[[ARG]]
+// CHECK: memref.dealloc %[[ARG]]
+// CHECK: return %[[RES]]
+
+// -----
+
+// CHECK-LABEL: @clone_multiple_dealloc_of_clone
+// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
+func @clone_multiple_dealloc_of_clone(%arg0: memref<?xf32>) -> memref<?xf32> {
+  // CHECK-NEXT: %[[CLONE:.*]] = bufferization.clone %[[ARG]]
+  // CHECK: memref.dealloc %[[CLONE]]
+  // CHECK: memref.dealloc %[[CLONE]]
+  // CHECK: return %[[ARG]]
+  %0 = bufferization.clone %arg0 : memref<?xf32> to memref<?xf32>
+  "use"(%0) : (memref<?xf32>) -> ()
+  "if_else"() ({
+    memref.dealloc %0 : memref<?xf32>
+    }, {
+    memref.dealloc %0 : memref<?xf32>
+    }) : () -> ()
+  return %arg0 : memref<?xf32>
+}
+
+// -----
+
+
+// CHECK-LABEL: func @tensor_cast_to_memref
+//  CHECK-SAME:   %[[ARG0:.+]]: tensor<4x6x16x32xi8>
+func @tensor_cast_to_memref(%arg0 : tensor<4x6x16x32xi8>) ->
+  memref<?x?x16x32xi8> {
+  %0 = tensor.cast %arg0 : tensor<4x6x16x32xi8> to tensor<?x?x16x32xi8>
+  %1 = bufferization.to_memref %0 : memref<?x?x16x32xi8>
+  return %1 : memref<?x?x16x32xi8>
+}
+// CHECK:   %[[M:.+]] = bufferization.to_memref %[[ARG0]] : memref<4x6x16x32xi8>
+// CHECK:   %[[M1:.+]] = memref.cast %[[M]] 
+// CHECK-SAME: memref<4x6x16x32xi8> to memref<?x?x16x32xi8>
+// CHECK:   return %[[M1]] : memref<?x?x16x32xi8>
+
+// -----
+
+// Folding of memref.load(to_memref(%v, %idxs)) -> tensor.extract(%v, %idx)
+// CHECK-LABEL: func @load_from_buffer_cast(
+func @load_from_buffer_cast(%arg0: index, %arg1: index,
+                            %arg2: tensor<?x?xf32>) -> f32 {
+  %0 = bufferization.to_memref %arg2 : memref<?x?xf32>
+  %1 = memref.load %0[%arg0, %arg1] : memref<?x?xf32>
+  return %1 : f32
+}
+//  CHECK-SAME: %[[IDX0:[0-9a-z]+]]: index, %[[IDX1:[0-9a-z]+]]: index
+//  CHECK-SAME: %[[TENSOR:[0-9a-z]+]]: tensor<?x?xf32>
+//       CHECK: %[[RES:.*]] = tensor.extract %[[TENSOR]][%[[IDX0]], %[[IDX1]]]
+//   CHECK-NOT: memref.load
+//       CHECK: return %[[RES]] : f32

mlir/test/Dialect/Bufferization/ops.mlir

+// RUN: mlir-opt %s | mlir-opt | FileCheck %s
+// RUN: mlir-opt %s --mlir-print-op-generic | mlir-opt | FileCheck %s
+
+// CHECK-LABEL: func @test_clone
+func @test_clone(%buf : memref<*xf32>) -> memref<*xf32> {
+  %clone = bufferization.clone %buf : memref<*xf32> to memref<*xf32>
+  return %clone : memref<*xf32>
+}
+
+// CHECK-LABEL: test_to_memref
+func @test_to_memref(%arg0: tensor<?xi64>, %arg1: tensor<*xi64>)
+    -> (memref<?xi64, affine_map<(d0) -> (d0 + 7)>>, memref<*xi64, 1>) {
+  %0 = bufferization.to_memref %arg0
+    : memref<?xi64, affine_map<(d0) -> (d0 + 7)>>
+  %1 = bufferization.to_memref %arg1
+    : memref<*xi64, 1>
+  return %0, %1 : memref<?xi64, affine_map<(d0) -> (d0 + 7)>>, memref<*xi64, 1>
+}
+
+// CHECK-LABEL: func @test_to_tensor
+func @test_to_tensor(%buf : memref<2xf32>) -> tensor<2xf32> {
+  %tensor = bufferization.to_tensor %buf : memref<2xf32>
+  return %tensor : tensor<2xf32>
+}

mlir/test/Dialect/Linalg/bufferize.mlir

 // RUN: mlir-opt -linalg-bufferize  -canonicalize -cse -split-input-file %s | FileCheck %s
 
 #map0 = affine_map<(d0) -> (d0)>
 
 // In-depth checking of a basic case, this is testing
-// - memref.buffer_cast / memref.tensor_load materializations are properly inserted
+// - bufferization.to_memref / bufferization.to_tensor materializations are
+//   properly inserted
 // - payload is correctly carried over
 // - affine maps are correctly carried over
 // Later tests will not check all these details.
 
 // CHECK: #map = affine_map<(d0) -> (d0)>
 // CHECK-LABEL:   func @basic(
 // CHECK-SAME:                %[[TENSOR:.*]]: tensor<4xf32>) -> tensor<4xf32> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<4xf32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<4xf32>
 // CHECK:           %[[RESULT_MEMREF:.*]] = memref.alloc() : memref<4xf32>
 // CHECK:           linalg.generic {indexing_maps = [#map, #map], iterator_types = ["parallel"]}
 // CHECK-SAME:      ins(%[[MEMREF]] : memref<4xf32>)
 // CHECK-SAME:      outs(%[[RESULT_MEMREF]] : memref<4xf32>) {
 // CHECK:           ^bb0(%[[RESULT1:.*]]: f32, %[[UNUSED:.*]]: f32):
 // CHECK:             %[[DIM1:.*]] = math.exp %[[RESULT1]] : f32
 // CHECK:             linalg.yield %[[DIM1]] : f32
 // CHECK:           }
-// CHECK:           %[[RESULT:.*]] = memref.tensor_load %[[RESULT_MEMREF]] : memref<4xf32>
+// CHECK:           %[[RESULT:.*]] = bufferization.to_tensor %[[RESULT_MEMREF]] : memref<4xf32>
 // CHECK:           return %[[RESULT]] : tensor<4xf32>
 func @basic(%arg0: tensor<4xf32>) -> tensor<4xf32> {
     %0 = linalg.generic {
       indexing_maps = [#map0, #map0],
       iterator_types = ["parallel"]
     } ins(%arg0 : tensor<4xf32>)
       outs(%arg0 : tensor<4xf32>) {
       ^bb0(%gen_arg1: f32, %out: f32):
         %tmp1 = math.exp %gen_arg1 : f32
         linalg.yield %tmp1 : f32
     } -> tensor<4xf32>
     return %0 : tensor<4xf32>
 }
 
 
 // -----
 
 #map0 = affine_map<(d0) -> (d0)>
 
 // Same as above but with linalg.init_tensor op.
 
 // CHECK: #map = affine_map<(d0) -> (d0)>
 // CHECK-LABEL: func @init_tensor(
 // CHECK-SAME:      %[[IN:.*]]: tensor<?xf32>, %[[SIZE:.*]]: index)
-// CHECK:         %[[MEMREF:.*]] = memref.buffer_cast %[[IN]] : memref<?xf32>
+// CHECK:         %[[MEMREF:.*]] = bufferization.to_memref %[[IN]] : memref<?xf32>
 // CHECK:         %[[OUT_BUF:.*]] = memref.alloc(%[[SIZE]]) : memref<?xf32>
 // CHECK:         linalg.generic
 // CHECK-SAME:    ins(%[[MEMREF]] : memref<?xf32>)
 // CHECK-SAME:    outs(%[[OUT_BUF]] : memref<?xf32>) {
 func @init_tensor(%in : tensor<?xf32>, %size: index) -> tensor<?xf32> {
   %init = linalg.init_tensor [%size] : tensor<?xf32>
   %0 = linalg.generic {
     indexing_maps = [#map0, #map0],

 
 // Check that the allocs properly consider the different shapes of the output
 // operands. The permuted indexing maps translate to different output shapes.
 
 // CHECK-LABEL:   func @dynamic_results(
 // CHECK-SAME:                          %[[ARG:.*]]: tensor<?x?xf32>
 // CHECK-DAG:       %[[C0:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[C1:.*]] = arith.constant 1 : index
-// CHECK:           %[[MEMREF_ARG:.*]] = memref.buffer_cast %[[ARG]] : memref<?x?xf32>
+// CHECK:           %[[MEMREF_ARG:.*]] = bufferization.to_memref %[[ARG]] : memref<?x?xf32>
 // CHECK:           %[[DIM0:.*]] = tensor.dim %[[ARG]], %[[C0]] : tensor<?x?xf32>
 // CHECK:           %[[DIM1:.*]] = tensor.dim %[[ARG]], %[[C1]] : tensor<?x?xf32>
 // CHECK:           %[[RESULT0:.*]] = memref.alloc(%[[DIM0]], %[[DIM1]]) : memref<?x?xf32>
 // CHECK:           %[[RESULT1:.*]] = memref.alloc(%[[DIM0]], %[[DIM1]]) : memref<?x?xf32>
 // CHECK:           linalg.generic
 // CHECK-SAME:      ins(%[[MEMREF_ARG]] : memref<?x?xf32>)
 // CHECK-SAME:      outs(%[[RESULT0]], %[[RESULT1]] : memref<?x?xf32>, memref<?x?xf32>)
 func @dynamic_results(%arg0: tensor<?x?xf32>)
#trait = {
   iterator_types = ["parallel", "parallel", "reduction"]
 }
 
 // Check the bufferization of init tensors.
 
 // CHECK-LABEL:   func @generic_with_init_tensor(
 // CHECK-SAME:                                   %[[ARG0_TENSOR:.*]]: tensor<2x3x4xvector<3x4xi4>>,
 // CHECK-SAME:                                   %[[ARG1_TENSOR:.*]]: tensor<3x2xf32>) -> tensor<3x2xf32> {
-// CHECK-DAG:           %[[ARG0_MEMREF:.*]] = memref.buffer_cast %[[ARG0_TENSOR]] : memref<2x3x4xvector<3x4xi4>>
-// CHECK-DAG:           %[[ARG1_MEMREF:.*]] = memref.buffer_cast %[[ARG1_TENSOR]] : memref<3x2xf32>
+// CHECK-DAG:           %[[ARG0_MEMREF:.*]] = bufferization.to_memref %[[ARG0_TENSOR]] : memref<2x3x4xvector<3x4xi4>>
+// CHECK-DAG:           %[[ARG1_MEMREF:.*]] = bufferization.to_memref %[[ARG1_TENSOR]] : memref<3x2xf32>
 // CHECK:           %[[INIT_BUFFER:.*]] = memref.alloc() : memref<3x2xf32>
 // CHECK:           linalg.copy(%[[ARG1_MEMREF]], %[[INIT_BUFFER]]) : memref<3x2xf32>, memref<3x2xf32>
 // CHECK:           linalg.generic
 // CHECK-SAME:      ins(%[[ARG0_MEMREF]] : memref<2x3x4xvector<3x4xi4>>)
 // CHECK-SAME:      outs(%[[INIT_BUFFER]] : memref<3x2xf32>) {
 func @generic_with_init_tensor(%arg0: tensor<2x3x4xvector<3x4xi4>>,
   %arg1: tensor<3x2xf32>) -> (tensor<3x2xf32>) {
 

 // CHECK-DAG: #[[$MAP0:[0-9a-z]*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
 // CHECK-DAG: #[[$MAP1:[0-9a-z]*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1 * 2)>
 
 func private @make_index() -> index
 
 // CHECK-LABEL: func @bufferize_slice(
 //  CHECK-SAME:   %[[T:[0-9a-z]*]]: tensor<?x?xf32>
 func @bufferize_slice(%t : tensor<?x?xf32>) -> (tensor<2x3xf32>, tensor<2x?xf32>) {
-  //      CHECK: %[[M:.*]] = memref.buffer_cast %[[T]] : memref<?x?xf32>
+  //      CHECK: %[[M:.*]] = bufferization.to_memref %[[T]] : memref<?x?xf32>
 
   //      CHECK: %[[IDX:.*]] = call @make_index() : () -> index
   %i0 = call @make_index() : () -> index
 
   // CHECK-NEXT: %[[A0:.*]] = memref.alloc() : memref<2x3xf32>
   // CHECK-NEXT: %[[SM0:.*]] = memref.subview %[[M]][0, 0] [2, 3] [1, 1]
   // CHECK-SAME:   memref<?x?xf32> to memref<2x3xf32, #[[$MAP0]]>
   // CHECK-NEXT: linalg.copy(%[[SM0]], %[[A0]]) : memref<2x3xf32, #[[$MAP0]]>, memref<2x3xf32>
-  // CHECK-NEXT: %[[RT0:.*]] = memref.tensor_load %[[A0]] : memref<2x3xf32>
+  // CHECK-NEXT: %[[RT0:.*]] = bufferization.to_tensor %[[A0]] : memref<2x3xf32>
   %st0 = tensor.extract_slice %t[0, 0][2, 3][1, 1] : tensor<?x?xf32> to tensor<2x3xf32>
 
   // CHECK-NEXT: %[[A1:.*]] = memref.alloc(%[[IDX]]) : memref<2x?xf32>
   // CHECK-NEXT: %[[SM1:.*]] = memref.subview %[[M]][0, %[[IDX]]] [2, %[[IDX]]] [1, 2]
   // CHECK-SAME:   memref<?x?xf32> to memref<2x?xf32, #[[$MAP1]]>
   // CHECK-NEXT: linalg.copy(%[[SM1]], %[[A1]]) : memref<2x?xf32, #[[$MAP1]]>, memref<2x?xf32>
-  // CHECK-NEXT: %[[RT1:.*]] = memref.tensor_load %[[A1]] : memref<2x?xf32>
+  // CHECK-NEXT: %[[RT1:.*]] = bufferization.to_tensor %[[A1]] : memref<2x?xf32>
   %st1 = tensor.extract_slice %t[0, %i0][2, %i0][1, 2] : tensor<?x?xf32> to tensor<2x?xf32>
 
   // CHECK-NEXT: return %[[RT0]], %[[RT1]]
   return %st0, %st1 : tensor<2x3xf32>, tensor<2x?xf32>
 }
 
 // -----
 
 // CHECK-DAG: #[[$MAP0:[0-9a-z]*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
 // CHECK-DAG: #[[$MAP1:[0-9a-z]*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1 * 2)>
 
 func private @make_index() -> index
 
 // CHECK-LABEL: func @bufferize_insert_slice(
 //  CHECK-SAME:   %[[T:[0-9a-z]*]]: tensor<?x?xf32>
 //  CHECK-SAME:   %[[ST0:[0-9a-z]*]]: tensor<2x3xf32>
 //  CHECK-SAME:   %[[ST1:[0-9a-z]*]]: tensor<2x?xf32>
 func @bufferize_insert_slice(%t : tensor<?x?xf32>, %st0 : tensor<2x3xf32>, %st1 : tensor<2x?xf32>) ->
     (tensor<?x?xf32>, tensor<?x?xf32>) {
-  // CHECK-DAG: %[[M:.*]] = memref.buffer_cast %[[T]] : memref<?x?xf32>
-  // CHECK-DAG: %[[SM0:.*]] = memref.buffer_cast %[[ST0]] : memref<2x3xf32>
-  // CHECK-DAG: %[[SM1:.*]] = memref.buffer_cast %[[ST1]] : memref<2x?xf32>
+  // CHECK-DAG: %[[M:.*]] = bufferization.to_memref %[[T]] : memref<?x?xf32>
+  // CHECK-DAG: %[[SM0:.*]] = bufferization.to_memref %[[ST0]] : memref<2x3xf32>
+  // CHECK-DAG: %[[SM1:.*]] = bufferization.to_memref %[[ST1]] : memref<2x?xf32>
 
   %c0 = arith.constant 0 : index
   %c1 = arith.constant 1 : index
   // CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
   // CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
   %i0 = call @make_index() : () -> index
   // CHECK: %[[IDX:.*]] = call @make_index() : () -> index
 
 
   // CHECK-NEXT: %[[DIM0:.*]] = tensor.dim %[[T]], %[[C0]] : tensor<?x?xf32>
   // CHECK-NEXT: %[[DIM1:.*]] = tensor.dim %[[T]], %[[C1]] : tensor<?x?xf32>
   // CHECK-NEXT: %[[M_COPY0:.*]] = memref.alloc(%[[DIM0]], %[[DIM1]]) : memref<?x?xf32>
   // CHECK-NEXT: linalg.copy(%[[M]], %[[M_COPY0]]) : memref<?x?xf32>, memref<?x?xf32>
   // CHECK-NEXT: %[[SUBVIEW0:.*]] = memref.subview %[[M_COPY0]][0, 0] [2, 3] [1, 1]
   // CHECK-SAME:   memref<?x?xf32> to memref<2x3xf32, #[[$MAP0]]>
   // CHECK-NEXT: linalg.copy(%[[SM0]], %[[SUBVIEW0]]) : memref<2x3xf32>, memref<2x3xf32, #[[$MAP0]]>
-  // CHECK-NEXT: %[[RT0:.*]] = memref.tensor_load %[[M_COPY0]] : memref<?x?xf32>
+  // CHECK-NEXT: %[[RT0:.*]] = bufferization.to_tensor %[[M_COPY0]] : memref<?x?xf32>
   %t0 = tensor.insert_slice %st0 into %t[0, 0][2, 3][1, 1] : tensor<2x3xf32> into tensor<?x?xf32>
 
   // CHECK-NEXT: %[[M_COPY1:.*]] = memref.alloc(%[[DIM0]], %[[DIM1]]) : memref<?x?xf32>
   // CHECK-NEXT: linalg.copy(%[[M]], %[[M_COPY1]]) : memref<?x?xf32>, memref<?x?xf32>
   // CHECK-NEXT: %[[SUBVIEW1:.*]] = memref.subview %[[M_COPY1]][0, %[[IDX]]] [2, %[[IDX]]] [1, 2]
   // CHECK-SAME:   memref<?x?xf32> to memref<2x?xf32, #[[$MAP1]]>
   // CHECK-NEXT: linalg.copy(%[[SM1]], %[[SUBVIEW1]]) : memref<2x?xf32>, memref<2x?xf32, #[[$MAP1]]>
-  // CHECK-NEXT: %[[RT1:.*]] = memref.tensor_load %[[M_COPY1]] : memref<?x?xf32>
+  // CHECK-NEXT: %[[RT1:.*]] = bufferization.to_tensor %[[M_COPY1]] : memref<?x?xf32>
   %t1 = tensor.insert_slice %st1 into %t[0, %i0][2, %i0][1, 2] : tensor<2x?xf32> into tensor<?x?xf32>
 
   //     CHECK: return %[[RT0]], %[[RT1]]
   return %t0, %t1: tensor<?x?xf32>, tensor<?x?xf32>
 }
 
 // -----
 
 // CHECK-LABEL: func @bufferize_fill(
 // CHECK-SAME:    %[[IN:.*]]: tensor<?xf32>
 func @bufferize_fill(%arg0: tensor<?xf32>) -> tensor<?xf32> {
   %c0 = arith.constant 0.0 : f32
-  // CHECK: %[[MEMREF:.*]] = memref.buffer_cast %[[IN]] : memref<?xf32>
+  // CHECK: %[[MEMREF:.*]] = bufferization.to_memref %[[IN]] : memref<?xf32>
   // CHECK: linalg.fill(%cst, %[[MEMREF]]) : f32, memref<?xf32>
-  // CHECK: %[[TENSOR:.*]] = memref.tensor_load %[[MEMREF]] : memref<?xf32>
+  // CHECK: %[[TENSOR:.*]] = bufferization.to_tensor %[[MEMREF]] : memref<?xf32>
   // CHECK: return %[[TENSOR]]
   %0 = linalg.fill(%c0, %arg0) : f32, tensor<?xf32> -> tensor<?xf32>
   return %0 : tensor<?xf32>
 }
 
 // -----
 
 // CHECK-LABEL: func @bufferize_tensor_collapse_shape(
 // CHECK-SAME:    %[[IN:.*]]: tensor<4x5xf32>
 func @bufferize_tensor_collapse_shape(%arg0: tensor<4x5xf32>) -> tensor<20xf32> {
   %out = linalg.tensor_collapse_shape %arg0 [[0, 1]] :
      tensor<4x5xf32> into tensor<20xf32>
   return %out : tensor<20xf32>
 }
-// CHECK: %[[MEMREF:.*]] = memref.buffer_cast %[[IN]] : memref<4x5xf32>
+// CHECK: %[[MEMREF:.*]] = bufferization.to_memref %[[IN]] : memref<4x5xf32>
 // CHECK: %[[RESHAPE:.*]] = memref.collapse_shape %[[MEMREF]] {{\[}}[0, 1]]
 // CHECK-SAME: : memref<4x5xf32> into memref<20xf32>
-// CHECK: %[[TENSOR:.*]] = memref.tensor_load %[[RESHAPE]] : memref<20xf32>
+// CHECK: %[[TENSOR:.*]] = bufferization.to_tensor %[[RESHAPE]] : memref<20xf32>
 // CHECK: return %[[TENSOR]]
 
 // -----
 
 // CHECK-LABEL:   func @pad_tensor_dynamic_shape(
 // CHECK-SAME:                                   %[[IN:.*]]: tensor<4x?x2x?xf32>,
 // CHECK-SAME:                                   %[[OFFSET:.*]]: index) -> tensor<4x?x?x?xf32> {
 func @pad_tensor_dynamic_shape(%arg0: tensor<4x?x2x?xf32>, %arg1: index) -> tensor<4x?x?x?xf32> {
   %c0 = arith.constant 0 : index
   %cst = arith.constant 0.0 : f32
   %out = linalg.pad_tensor %arg0 low[%c0, %c0, %arg1, %c0] high[%c0, %c0, %c0, %arg1]  {
   ^bb0(%gen_arg1: index, %gen_arg2: index, %gen_arg3: index, %gen_arg4: index):  // no predecessors
     linalg.yield %cst : f32
   } : tensor<4x?x2x?xf32> to tensor<4x?x?x?xf32>
   return %out : tensor<4x?x?x?xf32>
 }
 
 // CHECK-DAG:       %[[C3:.*]] = arith.constant 3 : index
 // CHECK-DAG:       %[[C2:.*]] = arith.constant 2 : index
 // CHECK-DAG:       %[[C1:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[CST:.*]] = arith.constant 0.000000e+00 : f32
-// CHECK:           %[[IN_MEMREF:.*]] = memref.buffer_cast %[[IN]] : memref<4x?x2x?xf32>
+// CHECK:           %[[IN_MEMREF:.*]] = bufferization.to_memref %[[IN]] : memref<4x?x2x?xf32>
 // CHECK:           %[[DIM1:.*]] = tensor.dim %[[IN]], %[[C1]] : tensor<4x?x2x?xf32>
 // CHECK:           %[[OUT_DIM2:.*]] = arith.addi %[[OFFSET]], %[[C2]] : index
 // CHECK:           %[[DIM3:.*]] = tensor.dim %[[IN]], %[[C3]] : tensor<4x?x2x?xf32>
 // CHECK:           %[[OUT_DIM3:.*]] = arith.addi %[[DIM3]], %[[OFFSET]] : index
 // CHECK:           %[[FILLED:.*]] = memref.alloc(%[[DIM1]], %[[OUT_DIM2]], %[[OUT_DIM3]]) : memref<4x?x?x?xf32>
 // CHECK:           linalg.fill(%[[CST]], %[[FILLED]]) : f32, memref<4x?x?x?xf32>
 // CHECK:           %[[OUT:.*]] = memref.alloc(%[[DIM1]], %[[OUT_DIM2]], %[[OUT_DIM3]]) : memref<4x?x?x?xf32>
 // CHECK:           linalg.copy(%[[FILLED]], %[[OUT]]) : memref<4x?x?x?xf32>, memref<4x?x?x?xf32>
 // CHECK:           %[[INTERIOR:.*]] = memref.subview %[[OUT]][0, 0, %[[OFFSET]], 0] [4, %[[DIM1]], 2, %[[DIM3]]] [1, 1, 1, 1] : memref<4x?x?x?xf32> to memref<4x?x2x?xf32, #map>
 // CHECK:           linalg.copy(%[[IN_MEMREF]], %[[INTERIOR]]) : memref<4x?x2x?xf32>, memref<4x?x2x?xf32, #map>
-// CHECK:           %[[OUT_TENSOR:.*]] = memref.tensor_load %[[OUT]] : memref<4x?x?x?xf32>
+// CHECK:           %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[OUT]] : memref<4x?x?x?xf32>
 // CHECK:           return %[[OUT_TENSOR]] : tensor<4x?x?x?xf32>
 // CHECK:         }
 
 
 // -----
 
 // CHECK-LABEL:   func @vector_transfer
 func @vector_transfer(%in: tensor<4xf32>, %out: tensor<4xf32>) {

 
 // CHECK-LABEL:   func @bufferize_dot
 func @bufferize_dot(%in: tensor<4xf32>, %out: tensor<f32>) -> tensor<f32> {
   %dot = linalg.dot ins(%in, %in : tensor<4xf32>, tensor<4xf32>)
                           outs(%out : tensor<f32>) -> tensor<f32>
   return %dot : tensor<f32>
   // CHECK: linalg.dot ins(%{{.*}}, %{{.*}} : memref<4xf32>, memref<4xf32>)
   // CHECK-SAME:       outs(%[[OUT:.*]] : memref<f32>)
-  // CHECK: %[[OUT_TENSOR:.*]] = memref.tensor_load %[[OUT]] : memref<f32>
+  // CHECK: %[[OUT_TENSOR:.*]] = bufferization.to_tensor %[[OUT]] : memref<f32>
   // CHECK: return %[[OUT_TENSOR]]
 }

mlir/test/Dialect/Linalg/lower-pad-tensor.mlir

 // RUN: mlir-opt -split-input-file --test-linalg-transform-patterns="test-transform-pad-tensor"  %s | FileCheck --check-prefix=CHECK %s
 
 // CHECK-DAG:   #[[$MAP0:.*]] = affine_map<(d0, d1, d2, d3) -> (d0, d1, d2, d3)>
 // CHECK-DAG:   #[[$MAP1:.*]] = affine_map<(d0, d1, d2, d3) -> (d0 + 1, d1 + 1, d2 + 1, d3 + 2)>
 // CHECK-LABEL: func @pad_tensor_with_memrefs
 func @pad_tensor_with_memrefs(%arg0: memref<1x28x28x1xf32>) -> memref<2x31x31x3xf32> {
   %cst = arith.constant 0.000000e+00 : f32
-  %0 = memref.tensor_load %arg0 : memref<1x28x28x1xf32>
+  %0 = bufferization.to_tensor %arg0 : memref<1x28x28x1xf32>
   %1 = linalg.pad_tensor %0 low[1, 1, 1, 2] high[0, 2, 2, 0]  {
   ^bb0(%arg1: index, %arg2: index, %arg3: index, %arg4: index):  // no predecessors
     linalg.yield %cst : f32
   } : tensor<1x28x28x1xf32> to tensor<2x31x31x3xf32>
-  %2 = memref.buffer_cast %1 : memref<2x31x31x3xf32>
+  %2 = bufferization.to_memref %1 : memref<2x31x31x3xf32>
   return %2 : memref<2x31x31x3xf32>
 }
 
 // CHECK:       linalg.fill
 // CHECK:       linalg.generic
 // CHECK-SAME:     indexing_maps = [#[[$MAP0]], #[[$MAP1]]]
 
 // -----

mlir/test/Dialect/MemRef/canonicalize.mlir

 // RUN: mlir-opt %s -canonicalize --split-input-file -allow-unregistered-dialect | FileCheck %s
 
-// Test case: Basic folding of memref.tensor_load(memref.buffer_cast(t)) -> t
-// CHECK-LABEL: func @tensor_load_of_buffer_cast(
-//  CHECK-SAME:   %[[TENSOR:.*]]: tensor<?xf32>) -> tensor<?xf32> {
-//       CHECK: return %[[TENSOR]]
-func @tensor_load_of_buffer_cast(%arg0: tensor<?xf32>) -> tensor<?xf32> {
-  %0 = memref.buffer_cast %arg0 : memref<?xf32>
-  %1 = memref.tensor_load %0 : memref<?xf32>
-  return %1 : tensor<?xf32>
-}
-
-// -----
-
-// Test case: Basic folding of memref.buffer_cast(memref.tensor_load(m)) -> m
-// CHECK-LABEL: func @buffer_cast_of_tensor_load(
-//  CHECK-SAME:   %[[MEMREF:.*]]: memref<?xf32>) -> memref<?xf32> {
-//       CHECK: return %[[MEMREF]]
-func @buffer_cast_of_tensor_load(%arg0: memref<?xf32>) -> memref<?xf32> {
-  %0 = memref.tensor_load %arg0 : memref<?xf32>
-  %1 = memref.buffer_cast %0 : memref<?xf32>
-  return %1 : memref<?xf32>
-}
-
-// -----
-
-// Test case: If the memrefs are not the same type, don't fold them.
-// Test case: If the memrefs are not cast-compatible (e.g. different address space),
-// don't canonicalize them either.
-// CHECK-LABEL: func @no_fold_buffer_cast_of_tensor_load(
-//  CHECK-SAME:   %[[MEMREF_ADDRSPACE2:.*]]: memref<?xf32, 2>)
-//  CHECK-SAME:     -> memref<?xf32, 7> {
-//       CHECK: %[[TENSOR:.*]] = memref.tensor_load
-//  CHECK_SAME:   %[[MEMREF_ADDRSPACE2]] : memref<?xf32, 2>
-//       CHECK: %[[MEMREF_ADDRSPACE7:.*]] = memref.buffer_cast
-//  CHECK_SAME:   %[[TENSOR]] : memref<?xf32, 7>
-//       CHECK: return %[[MEMREF_ADDRSPACE7]]
-func @no_fold_buffer_cast_of_tensor_load(%arg0: memref<?xf32, 2>) -> memref<?xf32, 7> {
-  %0 = memref.tensor_load %arg0 : memref<?xf32, 2>
-  %1 = memref.buffer_cast %0 : memref<?xf32, 7>
-  return %1 : memref<?xf32, 7>
-}
-
-// -----
-
-// CHECK-DAG: #[[$OFF_3:[a-z0-9]+]] = affine_map<(d0) -> (d0 + 3)>
-// CHECK-DAG: #[[$OFF_UNK:[a-z0-9]+]] = affine_map<(d0)[s0] -> (d0 + s0)>
-
-// Test case: If the memrefs are definitely cast-compatible, canonicalize to
-//            cast.
-// CHECK-LABEL: func @canonicalize_buffer_cast_of_tensor_load(
-//  CHECK-SAME:   %[[M:.*]]: memref<?xf32, #[[$OFF_3]]>)
-//  CHECK-SAME:     -> memref<?xf32, #[[$OFF_UNK]]> {
-//   CHECK-NOT: memref.tensor_load
-//   CHECK-NOT: memref.buffer_cast
-//       CHECK: %[[R:.*]] = memref.cast %[[M]]
-//  CHECK-SAME:   memref<?xf32, #[[$OFF_3]]> to memref<?xf32, #[[$OFF_UNK]]>
-//       CHECK: return %[[R]]
-func @canonicalize_buffer_cast_of_tensor_load(
-  %arg0: memref<?xf32, offset: 3, strides: [1]>)
-  -> memref<?xf32, offset: ?, strides: [1]>
-{
-  %0 = memref.tensor_load %arg0 : memref<?xf32, offset: 3, strides: [1]>
-  %1 = memref.buffer_cast %0 : memref<?xf32, offset: ?, strides: [1]>
-  return %1 : memref<?xf32, offset: ?, strides: [1]>
-}
-
-// -----
-
-// CHECK-DAG: #[[$OFF_UNK:[a-z0-9]+]] = affine_map<(d0)[s0] -> (d0 + s0)>
-// CHECK-DAG: #[[$OFF_3:[a-z0-9]+]] = affine_map<(d0) -> (d0 + 3)>
-
-// Test case: If the memrefs are potentially cast-compatible, canonicalize to
-//            copy.
-// CHECK-LABEL: func @canonicalize_buffer_cast_of_tensor_load_to_copy(
-//  CHECK-SAME:   %[[M:.*]]: memref<?xf32, #[[$OFF_UNK]]>)
-//  CHECK-SAME:     -> memref<?xf32, #[[$OFF_3]]> {
-//   CHECK-NOT: memref.tensor_load
-//   CHECK-NOT: memref.buffer_cast
-//       CHECK: %[[C0:.*]] = arith.constant 0 : index
-//       CHECK: %[[DIM:.*]] = memref.dim %[[M]], %[[C0]] : memref<?xf32, #[[$OFF_UNK]]>
-//       CHECK: %[[ALLOC:.*]] = memref.alloc(%[[DIM]]) : memref<?xf32, #[[$OFF_3]]>
-//       CHECK: memref.copy %[[M]], %[[ALLOC]]
-//  CHECK-SAME:   memref<?xf32, #[[$OFF_UNK]]> to memref<?xf32, #[[$OFF_3]]>
-//       CHECK: return %[[ALLOC]]
-func @canonicalize_buffer_cast_of_tensor_load_to_copy(
-  %arg0: memref<?xf32, offset: ?, strides: [1]>)
-  -> memref<?xf32, offset: 3, strides: [1]>
-{
-  %0 = memref.tensor_load %arg0 : memref<?xf32, offset: ?, strides: [1]>
-  %1 = memref.buffer_cast %0 : memref<?xf32, offset: 3, strides: [1]>
-  return %1 : memref<?xf32, offset: 3, strides: [1]>
-}
-
-// -----
-
 // CHECK-LABEL: func @subview_of_memcast
 //  CHECK-SAME:   %[[ARG0:.[a-z0-9A-Z_]+]]: memref<4x6x16x32xi8>
 //       CHECK:   %[[S:.+]] = memref.subview %arg0[0, 1, 0] [1, 1, 16] [1, 1, 1] : memref<4x6x16x32xi8> to memref<16x32xi8, #{{.*}}>
 //       CHECK:   %[[M:.+]] = memref.cast %[[S]] : memref<16x32xi8, #{{.*}}> to memref<16x32xi8, #{{.*}}>
 //       CHECK:   return %[[M]] : memref<16x32xi8, #{{.*}}>
 func @subview_of_memcast(%arg : memref<4x6x16x32xi8>) ->
   memref<16x32xi8, affine_map<(d0, d1)[s0] -> (d0 * 32 + d1 + s0)>>{
   %0 = memref.cast %arg : memref<4x6x16x32xi8> to memref<?x?x16x32xi8>

 //       CHECK:   %[[REDUCED1:.+]] = memref.subview %{{.+}}[0, %{{.+}}, %{{.+}}] [1, 1, %{{.+}}] [1, 1, 1]
 //  CHECK-SAME:       : memref<?x?x?xf32, #[[MAP2]]> to memref<1x?xf32, #[[MAP1]]>
 //       CHECK:   %[[REDUCED2:.+]] = memref.subview %[[REDUCED1]][0, 0] [1, %{{.+}}] [1, 1]
 //  CHECK-SAME:       : memref<1x?xf32, #[[MAP1]]> to memref<?xf32, #[[MAP0]]>
 
 
 // -----
 
-// CHECK-LABEL: @clone_before_dealloc
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_before_dealloc(%arg0: memref<?xf32>) -> memref<?xf32> {
-  // CHECK-NEXT: return %[[ARG]]
-  %0 = memref.clone %arg0 : memref<?xf32> to memref<?xf32>
-  memref.dealloc %arg0 : memref<?xf32>
-  return %0 : memref<?xf32>
-}
-
-// -----
-
-// CHECK-LABEL: @clone_before_dealloc
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_before_dealloc(%arg0: memref<?xf32>) -> memref<?xf32> {
-  // CHECK-NEXT: "use"(%arg0)
-  // CHECK-NEXT: return %[[ARG]]
-  %0 = memref.clone %arg0 : memref<?xf32> to memref<?xf32>
-  "use"(%0) : (memref<?xf32>) -> ()
-  memref.dealloc %0 : memref<?xf32>
-  return %arg0 : memref<?xf32>
-}
-
-// -----
-
-// CHECK-LABEL: @clone_after_cast
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_after_cast(%arg0: memref<?xf32>) -> memref<32xf32> {
-  // CHECK-NEXT: memref.clone %[[ARG]] : memref<?xf32> to memref<32xf32>
-  // CHECK-NOT: memref.cast
-  %0 = memref.cast %arg0 : memref<?xf32> to memref<32xf32>
-  %1 = memref.clone %0 : memref<32xf32> to memref<32xf32>
-  return %1 : memref<32xf32>
-}
-
-// -----
-
-// CHECK-LABEL: @clone_and_cast
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_and_cast(%arg0: memref<?xf32>) -> memref<32xf32> {
-  // CHECK-NEXT: %[[RES:.*]] = memref.cast %[[ARG]] : memref<?xf32> to memref<32xf32>
-  %0 = memref.clone %arg0 : memref<?xf32> to memref<32xf32>
-  // CHECK-NEXT: return %[[RES]]
-  memref.dealloc %arg0 : memref<?xf32>
-  return %0 : memref<32xf32>
-}
-
-// -----
-
-// CHECK-LABEL: @alias_is_freed
-func @alias_is_freed(%arg0 : memref<?xf32>) {
-  // CHECK: memref.clone
-  // CHECK: memref.dealloc
-  // CHECK: memref.dealloc
-  %0 = memref.cast %arg0 : memref<?xf32> to memref<32xf32>
-  %1 = memref.clone %0 : memref<32xf32> to memref<32xf32>
-  memref.dealloc %arg0 : memref<?xf32>
-  "use"(%1) : (memref<32xf32>) -> ()
-  memref.dealloc %1 : memref<32xf32>
-  return
-}
-
-// -----
-
-// Verify SimplifyClones skips clones with multiple deallocations.
-// CHECK-LABEL: @clone_multiple_dealloc_of_source
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_multiple_dealloc_of_source(%arg0: memref<?xf32>) -> memref<?xf32> {
-  // CHECK-NEXT: %[[RES:.*]] = memref.clone %[[ARG]]
-  // CHECK: memref.dealloc %[[ARG]]
-  // CHECK: memref.dealloc %[[ARG]]
-  // CHECK: return %[[RES]]
-  %0 = memref.clone %arg0 : memref<?xf32> to memref<?xf32>
-  "if_else"() ({
-    memref.dealloc %arg0 : memref<?xf32>
-    }, {
-    memref.dealloc %arg0 : memref<?xf32>
-    }) : () -> ()
-  return %0 : memref<?xf32>
-}
-
-// -----
-
-// CHECK-LABEL: @clone_multiple_dealloc_of_clone
-// CHECK-SAME: %[[ARG:.*]]: memref<?xf32>
-func @clone_multiple_dealloc_of_clone(%arg0: memref<?xf32>) -> memref<?xf32> {
-  // CHECK-NEXT: %[[CLONE:.*]] = memref.clone %[[ARG]]
-  // CHECK: memref.dealloc %[[CLONE]]
-  // CHECK: memref.dealloc %[[CLONE]]
-  // CHECK: return %[[ARG]]
-  %0 = memref.clone %arg0 : memref<?xf32> to memref<?xf32>
-  "use"(%0) : (memref<?xf32>) -> ()
-  "if_else"() ({
-    memref.dealloc %0 : memref<?xf32>
-    }, {
-    memref.dealloc %0 : memref<?xf32>
-    }) : () -> ()
-  return %arg0 : memref<?xf32>
-}
-
-// -----
-
 // CHECK-LABEL: func @dim_of_sized_view
 //  CHECK-SAME:   %{{[a-z0-9A-Z_]+}}: memref<?xi8>
 //  CHECK-SAME:   %[[SIZE:.[a-z0-9A-Z_]+]]: index
 //       CHECK:   return %[[SIZE]] : index
 func @dim_of_sized_view(%arg : memref<?xi8>, %size: index) -> index {
   %c0 = arith.constant 0 : index
   %0 = memref.reinterpret_cast %arg to offset: [0], sizes: [%size], strides: [0] : memref<?xi8> to memref<?xi8>
   %1 = memref.dim %0, %c0 : memref<?xi8>
   return %1 : index
 }
 
 // -----
 
 // CHECK-LABEL: func @no_fold_of_store
 //  CHECK:   %[[cst:.+]] = memref.cast %arg
 //  CHECK:   memref.store %[[cst]]
 func @no_fold_of_store(%arg : memref<32xi8>, %holder: memref<memref<?xi8>>) {
   %0 = memref.cast %arg : memref<32xi8> to memref<?xi8>
   memref.store %0, %holder[] : memref<memref<?xi8>>
   return
 }
 
 // -----
 
-// Test case: Folding of memref.load(memref.buffer_cast(%v, %idxs))
-//            -> tensor.extract(%v, %idx)
-// CHECK-LABEL: func @load_from_buffer_cast(
-//  CHECK-SAME:     %[[IDX0:[0-9a-z]+]]: index, %[[IDX1:[0-9a-z]+]]: index
-//  CHECK-SAME:     %[[TENSOR:[0-9a-z]+]]: tensor<?x?xf32>
-//       CHECK:   %[[RES:.*]] = tensor.extract %[[TENSOR]][%[[IDX0]], %[[IDX1]]]
-//   CHECK-NOT:   memref.load
-//       CHECK:   return %[[RES]] : f32
-func @load_from_buffer_cast(%arg0: index, %arg1: index, %arg2: tensor<?x?xf32>) -> f32 {
-  %0 = memref.buffer_cast %arg2 : memref<?x?xf32>
-  %1 = memref.load %0[%arg0, %arg1] : memref<?x?xf32>
-  return %1 : f32
-}
-
-// -----
-
-
-// Test case: Basic folding of tensor.dim(memref.tensor_load(m)) -> memref.dim(m).
-// CHECK-LABEL: func @dim_of_tensor_load(
-//  CHECK-SAME:     %[[MEMREF:[0-9a-z]*]]: memref<?xf32>
-//       CHECK:   %[[C0:.*]] = arith.constant 0
-//       CHECK:   %[[D:.*]] = memref.dim %[[MEMREF]], %[[C0]]
-//       CHECK:   return %[[D]] : index
-func @dim_of_tensor_load(%arg0: memref<?xf32>) -> index {
-  %c0 = arith.constant 0 : index
-  %0 = memref.tensor_load %arg0 : memref<?xf32>
-  %1 = tensor.dim %0, %c0 : tensor<?xf32>
-  return %1 : index
-}
-
-// -----
-
 // Test case: Folding of memref.dim(memref.alloca(%size), %idx) -> %size
 // CHECK-LABEL: func @dim_of_alloca(
 //  CHECK-SAME:     %[[SIZE:[0-9a-z]+]]: index
 //  CHECK-NEXT:   return %[[SIZE]] : index
 func @dim_of_alloca(%size: index) -> index {
   %0 = memref.alloca(%size) : memref<?xindex>
   %c0 = arith.constant 0 : index
   %1 = memref.dim %0, %c0 : memref<?xindex>
func @dim_of_memref_reshape_i32(%arg0: memref<*xf32>, %arg1: memref<?xi32>)
   %0 = memref.reshape %arg0(%arg1)
       : (memref<*xf32>, memref<?xi32>) -> memref<*xf32>
   %1 = memref.dim %0, %c3 : memref<*xf32>
   return %1 : index
 }
 
 // -----
 
-// CHECK-LABEL: func @tensor_cast_to_memref
-//  CHECK-SAME:   %[[ARG0:.+]]: tensor<4x6x16x32xi8>
-//       CHECK:   %[[M:.+]] = memref.buffer_cast %[[ARG0]] : memref<4x6x16x32xi8>
-//       CHECK:   %[[M1:.+]] = memref.cast %[[M]] : memref<4x6x16x32xi8> to memref<?x?x16x32xi8>
-//       CHECK:   return %[[M1]] : memref<?x?x16x32xi8>
-func @tensor_cast_to_memref(%arg0 : tensor<4x6x16x32xi8>) ->
-  memref<?x?x16x32xi8> {
-  %0 = tensor.cast %arg0 : tensor<4x6x16x32xi8> to tensor<?x?x16x32xi8>
-  %1 = memref.buffer_cast %0 : memref<?x?x16x32xi8>
-  return %1 : memref<?x?x16x32xi8>
-}
-
-// -----
-
 // CHECK-LABEL: func @alloc_const_fold
 func @alloc_const_fold() -> memref<?xf32> {
   // CHECK-NEXT: %0 = memref.alloc() : memref<4xf32>
   %c4 = arith.constant 4 : index
   %a = memref.alloc(%c4) : memref<?xf32>
 
   // CHECK-NEXT: %1 = memref.cast %0 : memref<4xf32> to memref<?xf32>
   // CHECK-NEXT: return %1 : memref<?xf32>

mlir/test/Dialect/MemRef/ops.mlir

 // RUN: mlir-opt %s | mlir-opt | FileCheck %s
 // RUN: mlir-opt %s --mlir-print-op-generic | mlir-opt | FileCheck %s
 
 // 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: #[[$strided2DOFF0:.*]] = affine_map<(d0, d1)[s0] -> (d0 * s0 + d1)>
 // CHECK-DAG: #[[$strided3DOFF0:.*]] = affine_map<(d0, d1, d2)[s0, s1] -> (d0 * s0 + d1 * s1 + d2)>
 
-// CHECK-LABEL: test_buffer_cast
-func @test_buffer_cast(%arg0: tensor<?xi64>, %arg1: tensor<*xi64>) -> (memref<?xi64, affine_map<(d0) -> (d0 + 7)>>, memref<*xi64, 1>) {
-  %0 = memref.buffer_cast %arg0 : memref<?xi64, affine_map<(d0) -> (d0 + 7)>>
-  %1 = memref.buffer_cast %arg1 : memref<*xi64, 1>
-  return %0, %1 : memref<?xi64, affine_map<(d0) -> (d0 + 7)>>, memref<*xi64, 1>
-}
-
 // CHECK-LABEL: func @memref_reinterpret_cast
 func @memref_reinterpret_cast(%in: memref<?xf32>)
     -> memref<10x?xf32, offset: ?, strides: [?, 1]> {
   %c0 = arith.constant 0 : index
   %c10 = arith.constant 10 : index
   %out = memref.reinterpret_cast %in to
            offset: [%c0], sizes: [10, %c10], strides: [%c10, 1]
            : memref<?xf32> to memref<10x?xf32, offset: ?, strides: [?, 1]>
func @write_global_memref() {
   %1 = arith.constant dense<[1.0, 2.0]> : tensor<2xf32>
   memref.tensor_store %1, %0 : memref<2xf32>
   return
 }
 
 // CHECK-LABEL: func @read_global_memref
 func @read_global_memref() {
   %0 = memref.get_global @memref0 : memref<2xf32>
-  %1 = memref.tensor_load %0 : memref<2xf32>
-  return
-}
-
-// CHECK-LABEL: func @memref_clone
-func @memref_clone() {
-  %0 = memref.alloc() : memref<2xf32>
-  %1 = memref.cast %0 : memref<2xf32> to memref<*xf32>
-  %2 = memref.clone %1 : memref<*xf32> to memref<*xf32>
   return
 }
 
 // CHECK-LABEL: func @memref_copy
 func @memref_copy() {
   %0 = memref.alloc() : memref<2xf32>
   %1 = memref.cast %0 : memref<2xf32> to memref<*xf32>
   %2 = memref.alloc() : memref<2xf32>

mlir/test/Dialect/SCF/bufferize.mlir

 // RUN: mlir-opt %s -scf-bufferize | FileCheck %s
 
 // CHECK-LABEL:   func @if(
 // CHECK-SAME:             %[[PRED:.*]]: i1,
 // CHECK-SAME:             %[[TRUE_TENSOR:.*]]: tensor<?xf32>,
 // CHECK-SAME:             %[[FALSE_TENSOR:.*]]: tensor<?xf32>) -> tensor<?xf32> {
-// CHECK:           %[[TRUE_MEMREF:.*]] = memref.buffer_cast %[[TRUE_TENSOR]] : memref<?xf32>
-// CHECK:           %[[FALSE_MEMREF:.*]] = memref.buffer_cast %[[FALSE_TENSOR]] : memref<?xf32>
+// CHECK:           %[[TRUE_MEMREF:.*]] = bufferization.to_memref %[[TRUE_TENSOR]] : memref<?xf32>
+// CHECK:           %[[FALSE_MEMREF:.*]] = bufferization.to_memref %[[FALSE_TENSOR]] : memref<?xf32>
 // CHECK:           %[[RESULT_MEMREF:.*]] = scf.if %[[PRED]] -> (memref<?xf32>) {
 // CHECK:             scf.yield %[[TRUE_MEMREF]] : memref<?xf32>
 // CHECK:           } else {
 // CHECK:             scf.yield %[[FALSE_MEMREF]] : memref<?xf32>
 // CHECK:           }
-// CHECK:           %[[RESULT_TENSOR:.*]] = memref.tensor_load %[[RESULT_MEMREF:.*]] : memref<?xf32>
+// CHECK:           %[[RESULT_TENSOR:.*]] = bufferization.to_tensor %[[RESULT_MEMREF:.*]] : memref<?xf32>
 // CHECK:           return %[[RESULT_TENSOR]] : tensor<?xf32>
 // CHECK:         }
 func @if(%pred: i1, %true_val: tensor<?xf32>, %false_val: tensor<?xf32>) -> tensor<?xf32> {
   %0 = scf.if %pred -> (tensor<?xf32>) {
     scf.yield %true_val : tensor<?xf32>
   } else {
     scf.yield %false_val : tensor<?xf32>
   }
   return %0 : tensor<?xf32>
 }
 
 // CHECK-LABEL:   func @for(
 // CHECK-SAME:              %[[TENSOR:.*]]: tensor<f32>,
 // CHECK-SAME:              %[[LB:.*]]: index, %[[UB:.*]]: index,
 // CHECK-SAME:              %[[STEP:.*]]: index) -> tensor<f32> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<f32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<f32>
 // CHECK:           %[[RESULT_MEMREF:.*]] = scf.for %[[VAL_6:.*]] = %[[LB]] to %[[UB]] step %[[STEP]] iter_args(%[[ITER:.*]] = %[[MEMREF]]) -> (memref<f32>) {
 // CHECK:             scf.yield %[[ITER]] : memref<f32>
 // CHECK:           }
-// CHECK:           %[[VAL_8:.*]] = memref.tensor_load %[[VAL_9:.*]] : memref<f32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_tensor %[[VAL_9:.*]] : memref<f32>
 // CHECK:           return %[[VAL_8]] : tensor<f32>
 // CHECK:         }
 func @for(%arg0: tensor<f32>, %lb: index, %ub: index, %step: index) -> tensor<f32> {
   %ret = scf.for %iv = %lb to %ub step %step iter_args(%iter = %arg0) -> tensor<f32> {
     scf.yield %iter : tensor<f32>
   }
   return %ret : tensor<f32>
 }
%0 = scf.if %pred -> (tensor<f32>) {
     scf.yield %1 : tensor<f32>
   }
   return %0 : tensor<f32>
 }
 
 // CHECK-LABEL:   func @for_correct_recursive_legalization_behavior(
 // CHECK-SAME:                                                      %[[TENSOR:.*]]: tensor<f32>,
 // CHECK-SAME:                                                      %[[INDEX:.*]]: index) -> tensor<f32> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<f32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<f32>
 // CHECK:           %[[RESULT:.*]] = scf.for %[[IV:.*]] = %[[INDEX]] to %[[INDEX]] step %[[INDEX]] iter_args(%[[MEMREF_ITER:.*]] = %[[MEMREF]]) -> (memref<f32>) {
-// CHECK:             %[[TENSOR_ITER:.*]] = memref.tensor_load %[[MEMREF_ITER]] : memref<f32>
+// CHECK:             %[[TENSOR_ITER:.*]] = bufferization.to_tensor %[[MEMREF_ITER]] : memref<f32>
 // CHECK:             %[[TENSOR_MUNGED:.*]] = "test.munge_tensor"(%[[TENSOR_ITER]]) : (tensor<f32>) -> tensor<f32>
-// CHECK:             %[[MEMREF_MUNGED:.*]] = memref.buffer_cast %[[TENSOR_MUNGED]] : memref<f32>
+// CHECK:             %[[MEMREF_MUNGED:.*]] = bufferization.to_memref %[[TENSOR_MUNGED]] : memref<f32>
 // CHECK:             scf.yield %[[MEMREF_MUNGED]] : memref<f32>
 // CHECK:           }
-// CHECK:           %[[TENSOR:.*]] = memref.tensor_load %[[RESULT:.*]] : memref<f32>
+// CHECK:           %[[TENSOR:.*]] = bufferization.to_tensor %[[RESULT:.*]] : memref<f32>
 // CHECK:           return %[[TENSOR]] : tensor<f32>
 // CHECK:         }
 func @for_correct_recursive_legalization_behavior(%arg0: tensor<f32>, %index: index) -> tensor<f32> {
   %ret = scf.for %iv = %index to %index step %index iter_args(%iter = %arg0) -> tensor<f32> {
     %0 = "test.munge_tensor"(%iter) : (tensor<f32>) -> (tensor<f32>)
     scf.yield %0 : tensor<f32>
   }
   return %ret : tensor<f32>
 }
 
 // CHECK-LABEL:   func @bufferize_while(
 // CHECK-SAME: %[[ARG0:.*]]: i64, %[[ARG1:.*]]: i64, %[[ARG2:.*]]: tensor<f32>
-// CHECK: %[[M:.*]] = memref.buffer_cast %[[ARG2]] : memref<f32>
+// CHECK: %[[M:.*]] = bufferization.to_memref %[[ARG2]] : memref<f32>
 // CHECK: %[[RES1:.*]]:3 = scf.while (%{{.*}} = %[[ARG0]], %{{.*}} = %[[M]]) : (i64, memref<f32>) -> (i64, i64, memref<f32>)
 // CHECK: scf.condition(%{{.*}}) %{{.*}}, %{{.*}}, %{{.*}} : i64, i64, memref<f32>
 // CHECK: ^bb0(%{{.*}}: i64, %{{.*}}: i64, %{{.*}}: memref<f32>):
 // CHECK: scf.yield %{{.*}}, %{{.*}} : i64, memref<f32>
-// CHECK:  %[[RES2:.*]] = memref.tensor_load %[[RES1]]#2 : memref<f32>
+// CHECK:  %[[RES2:.*]] = bufferization.to_tensor %[[RES1]]#2 : memref<f32>
 // CHECK:  return %[[RES1]]#1, %[[RES2]] : i64, tensor<f32>
 func @bufferize_while(%arg0: i64, %arg1: i64, %arg2: tensor<f32>) -> (i64, tensor<f32>) {
   %c2_i64 = arith.constant 2 : i64
   %0:3 = scf.while (%arg3 = %arg0, %arg4 = %arg2) : (i64, tensor<f32>) -> (i64, i64, tensor<f32>) {
     %1 = arith.cmpi slt, %arg3, %arg1 : i64
     scf.condition(%1) %arg3, %arg3, %arg4 : i64, i64, tensor<f32>
   } do {
   ^bb0(%arg5: i64, %arg6: i64, %arg7: tensor<f32>):
     %1 = arith.muli %arg6, %c2_i64 : i64
     scf.yield %1, %arg7 : i64, tensor<f32>
   }
   return %0#1, %0#2 : i64, tensor<f32>
 }

mlir/test/Dialect/SCF/canonicalize.mlir

 //  CHECK-SAME:   %[[T1:[0-9a-z]*]]: tensor<128x128xf32>
 //  CHECK-SAME:   %[[T2:[0-9a-z]*]]: tensor<128x128xf32>
 //  CHECK-SAME:   %[[M0:[0-9a-z]*]]: memref<128x128xf32>
 func @last_value(%t0: tensor<128x128xf32>, %t1: tensor<128x128xf32>,
                  %t2: tensor<128x128xf32>, %m0: memref<128x128xf32>,
                  %lb : index, %ub : index, %step : index)
   -> (tensor<128x128xf32>, tensor<128x128xf32>, tensor<128x128xf32>)
 {
-  // CHECK-NEXT: %[[M1:.*]] = memref.buffer_cast %[[T1]] : memref<128x128xf32>
+  // CHECK-NEXT: %[[M1:.*]] = bufferization.to_memref %[[T1]] : memref<128x128xf32>
   // CHECK-NEXT: %[[FOR_RES:.*]] = scf.for {{.*}} iter_args(%[[BBARG_T2:.*]] = %[[T2]]) -> (tensor<128x128xf32>) {
   %0:3 = scf.for %arg0 = %lb to %ub step %step iter_args(%arg1 = %t0, %arg2 = %t1, %arg3 = %t2)
     -> (tensor<128x128xf32>, tensor<128x128xf32>, tensor<128x128xf32>)
   {
-    %m1 = memref.buffer_cast %arg2 : memref<128x128xf32>
+    %m1 = bufferization.to_memref %arg2 : memref<128x128xf32>
 
     // CHECK-NEXT:   call @process(%[[M0]]) : (memref<128x128xf32>) -> ()
     call @process(%m0) : (memref<128x128xf32>) -> ()
 
     // CHECK-NEXT:   call @process(%[[M1]]) : (memref<128x128xf32>) -> ()
     call @process(%m1) : (memref<128x128xf32>) -> ()
 
     // This does not hoist (fails the bbArg has at most a single check).
     // CHECK-NEXT:   %[[T:.*]] = call @process_tensor(%[[BBARG_T2]]) : (tensor<128x128xf32>) -> memref<128x128xf32>
-    // CHECK-NEXT:   %[[YIELD_T:.*]] = memref.tensor_load %[[T:.*]]
+    // CHECK-NEXT:   %[[YIELD_T:.*]] = bufferization.to_tensor %[[T:.*]]
     %m2 = call @process_tensor(%arg3): (tensor<128x128xf32>) -> memref<128x128xf32>
-    %3 = memref.tensor_load %m2 : memref<128x128xf32>
+    %3 = bufferization.to_tensor %m2 : memref<128x128xf32>
 
     // All this stuff goes away, incrementally
-    %1 = memref.tensor_load %m0 : memref<128x128xf32>
-    %2 = memref.tensor_load %m1 : memref<128x128xf32>
+    %1 = bufferization.to_tensor %m0 : memref<128x128xf32>
+    %2 = bufferization.to_tensor %m1 : memref<128x128xf32>
 
     // CHECK-NEXT:   scf.yield %[[YIELD_T]] : tensor<128x128xf32>
     scf.yield %1, %2, %3 : tensor<128x128xf32>, tensor<128x128xf32>, tensor<128x128xf32>
 
   // CHECK-NEXT: }
   }
 
-  // CHECK-NEXT: %[[R0:.*]] = memref.tensor_load %[[M0]] : memref<128x128xf32>
-  // CHECK-NEXT: %[[R1:.*]] = memref.tensor_load %[[M1]] : memref<128x128xf32>
+  // CHECK-NEXT: %[[R0:.*]] = bufferization.to_tensor %[[M0]] : memref<128x128xf32>
+  // CHECK-NEXT: %[[R1:.*]] = bufferization.to_tensor %[[M1]] : memref<128x128xf32>
   // CHECK-NEXT: return %[[R0]], %[[R1]], %[[FOR_RES]] : tensor<128x128xf32>, tensor<128x128xf32>, tensor<128x128xf32>
   return %0#0, %0#1, %0#2 : tensor<128x128xf32>, tensor<128x128xf32>, tensor<128x128xf32>
 }
 
 // -----
 
 // CHECK-LABEL: fold_away_iter_with_no_use_and_yielded_input
 //  CHECK-SAME:   %[[A0:[0-9a-z]*]]: i32

mlir/test/Dialect/Shape/bufferize.mlir

 // RUN: mlir-opt -split-input-file -shape-bufferize <%s | FileCheck %s
 
 // -----
 
 // CHECK-LABEL:   func @shape_assuming() {
 // CHECK:           %[[WTRUE:.*]] = shape.const_witness true
 // CHECK:           %[[MEMREF:.*]] = shape.assuming %[[WTRUE]] -> (memref<2xf16>) {
 // CHECK:             %[[TENSOR_VAL:.*]] = "test.source"() : () -> tensor<2xf16>
-// CHECK:             %[[YIELDED_MEMREF:.*]] = memref.buffer_cast %[[TENSOR_VAL]] : memref<2xf16>
+// CHECK:             %[[YIELDED_MEMREF:.*]] = bufferization.to_memref %[[TENSOR_VAL]] : memref<2xf16>
 // CHECK:             shape.assuming_yield %[[YIELDED_MEMREF]] : memref<2xf16>
 // CHECK:           }
-// CHECK:           %[[TENSOR:.*]] = memref.tensor_load %[[MEMREF:.*]] : memref<2xf16>
+// CHECK:           %[[TENSOR:.*]] = bufferization.to_tensor %[[MEMREF:.*]] : memref<2xf16>
 // CHECK:           "test.sink"(%[[TENSOR]]) : (tensor<2xf16>) -> ()
 // CHECK:           return
 // CHECK:         }
 func @shape_assuming() {
   %0 = shape.const_witness true
   %1 = shape.assuming %0 -> (tensor<2xf16>) {
     %2 = "test.source"() : () -> (tensor<2xf16>)
     shape.assuming_yield %2 : tensor<2xf16>
   }
   "test.sink"(%1) : (tensor<2xf16>) -> ()
   return
 }

mlir/test/Dialect/SparseTensor/conversion_sparse2dense.mlir

 //       CHECK:   %[[Cond:.*]] = call @getNextI32(%[[Iter]], %[[IndD]], %[[ElemBuffer]]) : (!llvm.ptr<i8>, memref<?xindex>, memref<i32>) -> i1
 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<1xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<i32>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]]] : memref<13xi32>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<13xi32>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<13xi32>
 //       CHECK: return %[[T]] : tensor<13xi32>
 func @sparse_convert_1d(%arg0: tensor<13xi32, #SparseVector>) -> tensor<13xi32> {
   %0 = sparse_tensor.convert %arg0 : tensor<13xi32, #SparseVector> to tensor<13xi32>
   return %0 : tensor<13xi32>
 }
 
 // CHECK-LABEL: func @sparse_convert_1d_dyn(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<?xi32>

 //       CHECK:   %[[Cond:.*]] = call @getNextI32(%[[Iter]], %[[IndD]], %[[ElemBuffer]]) : (!llvm.ptr<i8>, memref<?xindex>, memref<i32>) -> i1
 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<1xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<i32>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]]] : memref<?xi32>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<?xi32>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<?xi32>
 //       CHECK: return %[[T]] : tensor<?xi32>
 func @sparse_convert_1d_dyn(%arg0: tensor<?xi32, #SparseVector>) -> tensor<?xi32> {
   %0 = sparse_tensor.convert %arg0 : tensor<?xi32, #SparseVector> to tensor<?xi32>
   return %0 : tensor<?xi32>
 }
 
 // CHECK-LABEL: func @sparse_convert_2d(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<2x4xf64>

 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<2xindex>
 //       CHECK:   %[[Iv1:.*]] = memref.load %[[IndS]][%[[I1]]] : memref<2xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<f64>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]], %[[Iv1]]] : memref<2x4xf64>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<2x4xf64>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<2x4xf64>
 //       CHECK: return %[[T]] : tensor<2x4xf64>
 func @sparse_convert_2d(%arg0: tensor<2x4xf64, #SparseMatrix>) -> tensor<2x4xf64> {
   %0 = sparse_tensor.convert %arg0 : tensor<2x4xf64, #SparseMatrix> to tensor<2x4xf64>
   return %0 : tensor<2x4xf64>
 }
 
 // CHECK-LABEL: func @sparse_convert_2d_dyn0(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<?x4xf64>

 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<2xindex>
 //       CHECK:   %[[Iv1:.*]] = memref.load %[[IndS]][%[[I1]]] : memref<2xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<f64>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]], %[[Iv1]]] : memref<?x4xf64>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<?x4xf64>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<?x4xf64>
 //       CHECK: return %[[T]] : tensor<?x4xf64>
 func @sparse_convert_2d_dyn0(%arg0: tensor<?x4xf64, #SparseMatrix>) -> tensor<?x4xf64> {
   %0 = sparse_tensor.convert %arg0 : tensor<?x4xf64, #SparseMatrix> to tensor<?x4xf64>
   return %0 : tensor<?x4xf64>
 }
 
 // CHECK-LABEL: func @sparse_convert_2d_dyn1(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<2x?xf64>

 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<2xindex>
 //       CHECK:   %[[Iv1:.*]] = memref.load %[[IndS]][%[[I1]]] : memref<2xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<f64>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]], %[[Iv1]]] : memref<2x?xf64>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<2x?xf64>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<2x?xf64>
 //       CHECK: return %[[T]] : tensor<2x?xf64>
 func @sparse_convert_2d_dyn1(%arg0: tensor<2x?xf64, #SparseMatrix>) -> tensor<2x?xf64> {
   %0 = sparse_tensor.convert %arg0 : tensor<2x?xf64, #SparseMatrix> to tensor<2x?xf64>
   return %0 : tensor<2x?xf64>
 }
 
 // CHECK-LABEL: func @sparse_convert_2d_dyn2(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<?x?xf64>

 //       CHECK:   scf.condition(%[[Cond]])
 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<2xindex>
 //       CHECK:   %[[Iv1:.*]] = memref.load %[[IndS]][%[[I1]]] : memref<2xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<f64>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]], %[[Iv1]]] : memref<?x?xf64>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<?x?xf64>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<?x?xf64>
 //       CHECK: return %[[T]] : tensor<?x?xf64>
 func @sparse_convert_2d_dyn2(%arg0: tensor<?x?xf64, #SparseMatrix>) -> tensor<?x?xf64> {
   %0 = sparse_tensor.convert %arg0 : tensor<?x?xf64, #SparseMatrix> to tensor<?x?xf64>
   return %0 : tensor<?x?xf64>
 }
 
 // CHECK-LABEL: func @sparse_convert_3d(
 //  CHECK-SAME: %[[Arg:.*]]: !llvm.ptr<i8>) -> tensor<2x3x4xf64>

 //       CHECK: } do {
 //       CHECK:   %[[Iv0:.*]] = memref.load %[[IndS]][%[[I0]]] : memref<3xindex>
 //       CHECK:   %[[Iv1:.*]] = memref.load %[[IndS]][%[[I1]]] : memref<3xindex>
 //       CHECK:   %[[Iv2:.*]] = memref.load %[[IndS]][%[[I2]]] : memref<3xindex>
 //       CHECK:   %[[ElemVal:.*]] = memref.load %[[ElemBuffer]][] : memref<f64>
 //       CHECK:   memref.store %[[ElemVal]], %[[M]][%[[Iv0]], %[[Iv1]], %[[Iv2]]] : memref<2x3x4xf64>
 //       CHECK:   scf.yield
 //       CHECK: }
-//       CHECK: %[[T:.*]] = memref.tensor_load %[[M]] : memref<2x3x4xf64>
+//       CHECK: %[[T:.*]] = bufferization.to_tensor %[[M]] : memref<2x3x4xf64>
 //       CHECK: return %[[T]] : tensor<2x3x4xf64>
 func @sparse_convert_3d(%arg0: tensor<2x3x4xf64, #SparseTensor>) -> tensor<2x3x4xf64> {
   %0 = sparse_tensor.convert %arg0 : tensor<2x3x4xf64, #SparseTensor> to tensor<2x3x4xf64>
   return %0 : tensor<2x3x4xf64>
 }

mlir/test/Dialect/SparseTensor/dense.mlir

 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32> {linalg.inplaceable = false}) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 1.000000e+00 : f32
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_11:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_12:.*]] = arith.muli %[[VAL_10]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_13:.*]] = arith.addi %[[VAL_12]], %[[VAL_11]] : index
 // CHECK:               %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xf32>
 // CHECK:               %[[VAL_15:.*]] = arith.addf %[[VAL_14]], %[[VAL_2]] : f32
 // CHECK:               memref.store %[[VAL_15]], %[[VAL_9]]{{\[}}%[[VAL_10]], %[[VAL_11]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_16:.*]] = memref.tensor_load %[[VAL_9]] : memref<32x16xf32>
+// CHECK:           %[[VAL_16:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_16]] : tensor<32x16xf32>
 // CHECK:         }
 func @dense1(%arga: tensor<32x16xf32, #DenseMatrix>,
              %argx: tensor<32x16xf32> {linalg.inplaceable = false})
        -> tensor<32x16xf32> {
   %c = arith.constant 1.0 : f32
   %0 = linalg.generic #trait_2d
      ins(%arga: tensor<32x16xf32, #DenseMatrix>)

 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32> {linalg.inplaceable = true}) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 1.000000e+00 : f32
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_10:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_11:.*]] = arith.muli %[[VAL_9]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_12:.*]] = arith.addi %[[VAL_11]], %[[VAL_10]] : index
 // CHECK:               %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xf32>
 // CHECK:               %[[VAL_14:.*]] = arith.addf %[[VAL_13]], %[[VAL_2]] : f32
 // CHECK:               memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_9]], %[[VAL_10]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32x16xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_15]] : tensor<32x16xf32>
 // CHECK:         }
 func @dense2(%arga: tensor<32x16xf32, #DenseMatrix>,
              %argx: tensor<32x16xf32> {linalg.inplaceable = true})
        -> tensor<32x16xf32> {
   %c = arith.constant 1.0 : f32
   %0 = linalg.generic #trait_2d
      ins(%arga: tensor<32x16xf32, #DenseMatrix>)

 // CHECK-LABEL:   func @dense3(
 // CHECK-SAME:      %[[VAL_0:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> {linalg.inplaceable = true}) -> tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> {
 // CHECK-DAG:       %[[VAL_2:.*]] = arith.constant 1.000000e+00 : f32
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_6:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32x16xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32x16xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xf32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_10:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_11:.*]] = arith.muli %[[VAL_9]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_12:.*]] = arith.addi %[[VAL_11]], %[[VAL_10]] : index
 // CHECK:               %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_9]], %[[VAL_10]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_14:.*]] = arith.addf %[[VAL_13]], %[[VAL_2]] : f32
 // CHECK:               memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<?xf32>

 // CHECK-LABEL:   func @dense4(
 // CHECK-SAME:      %[[VAL_0:.*]]: tensor<32x16x8xf32>,
 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> {linalg.inplaceable = true}) -> tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> {
 // CHECK-DAG:       %[[VAL_2:.*]] = arith.constant 8 : index
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_6:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}}>> to memref<?xf32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_10:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_11:.*]] = arith.muli %[[VAL_9]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_12:.*]] = arith.addi %[[VAL_11]], %[[VAL_10]] : index
 // CHECK:               %[[VAL_13:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<?xf32>
 // CHECK:               %[[VAL_14:.*]] = scf.for %[[VAL_15:.*]] = %[[VAL_5]] to %[[VAL_2]] step %[[VAL_6]] iter_args(%[[VAL_16:.*]] = %[[VAL_13]]) -> (f32) {
 // CHECK:                 %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_9]], %[[VAL_10]], %[[VAL_15]]] : memref<32x16x8xf32>

mlir/test/Dialect/SparseTensor/sparse_1d.mlir

 // CHECK-LABEL:   func @add_d(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_10:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_9]]] : memref<?xf32>
 // CHECK:             %[[VAL_11:.*]] = arith.addf %[[VAL_10]], %[[VAL_1]] : f32
 // CHECK:             memref.store %[[VAL_11]], %[[VAL_8]]{{\[}}%[[VAL_9]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_12:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
 // CHECK:           return %[[VAL_12]] : tensor<32xf32>
 // CHECK:         }
 func @add_d(%arga: tensor<32xf32, #DV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #DV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %a, %argb : f32

 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_7:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.fill(%[[VAL_3]], %[[VAL_7]]) : f32, memref<32xf32>
 // CHECK:           scf.for %[[VAL_8:.*]] = %[[VAL_4]] to %[[VAL_2]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_9:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_8]]] : memref<?xf32>
 // CHECK:             %[[VAL_10:.*]] = arith.addf %[[VAL_9]], %[[VAL_1]] : f32
 // CHECK:             memref.store %[[VAL_10]], %[[VAL_7]]{{\[}}%[[VAL_8]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_11:.*]] = memref.tensor_load %[[VAL_7]] : memref<32xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf32>
 // CHECK:           return %[[VAL_11]] : tensor<32xf32>
 // CHECK:         }
 func @add_d_init(%arga: tensor<32xf32, #DV>, %argb: f32) -> tensor<32xf32> {
   %u = linalg.init_tensor [32] : tensor<32xf32>
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #DV>)
     outs(%u: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %a, %argb : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @mul_d(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_10:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_9]]] : memref<?xf32>
 // CHECK:             %[[VAL_11:.*]] = arith.mulf %[[VAL_10]], %[[VAL_1]] : f32
 // CHECK:             memref.store %[[VAL_11]], %[[VAL_8]]{{\[}}%[[VAL_9]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_12:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
 // CHECK:           return %[[VAL_12]] : tensor<32xf32>
 // CHECK:         }
 func @mul_d(%arga: tensor<32xf32, #DV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #DV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.mulf %a, %argb : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @add_s(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {

 // CHECK:             %[[VAL_25:.*]] = arith.addi %[[VAL_18]], %[[VAL_6]] : index
 // CHECK:             %[[VAL_26:.*]] = select %[[VAL_24]], %[[VAL_25]], %[[VAL_18]] : index
 // CHECK:             %[[VAL_27:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_26]], %[[VAL_27]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_28:.*]] = %[[VAL_29:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             memref.store %[[VAL_1]], %[[VAL_11]]{{\[}}%[[VAL_28]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_30:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf32>
+// CHECK:           %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf32>
 // CHECK:           return %[[VAL_30]] : tensor<32xf32>
 // CHECK:         }
 func @add_s(%arga: tensor<32xf32, #SV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %a, %argb : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @repeated_add_s(
 // CHECK-SAME:                         %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                         %[[VAL_1:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
 // CHECK:             %[[VAL_15:.*]] = arith.addf %[[VAL_13]], %[[VAL_14]] : f32
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
 // CHECK:             %[[VAL_18:.*]] = arith.addf %[[VAL_16]], %[[VAL_17]] : f32
 // CHECK:             %[[VAL_19:.*]] = arith.addf %[[VAL_15]], %[[VAL_18]] : f32
 // CHECK:             memref.store %[[VAL_19]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_20:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xf32>
+// CHECK:           %[[VAL_20:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
 // CHECK:           return %[[VAL_20]] : tensor<32xf32>
 // CHECK:         }
 func @repeated_add_s(%arga: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %a, %a : f32  // same tensor
         %1 = arith.addf %a, %a : f32  // should yield
         %2 = arith.addf %0, %1 : f32  // one guard
         linalg.yield %2 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @mul_s(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_10]] to %[[VAL_11]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xf32>
 // CHECK:             %[[VAL_15:.*]] = arith.mulf %[[VAL_14]], %[[VAL_1]] : f32
 // CHECK:             memref.store %[[VAL_15]], %[[VAL_9]]{{\[}}%[[VAL_13]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_16:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xf32>
+// CHECK:           %[[VAL_16:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
 // CHECK:           return %[[VAL_16]] : tensor<32xf32>
 // CHECK:         }
 func @mul_s(%arga: tensor<32xf32, #SV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.mulf %a, %argb : f32

 // CHECK-LABEL:   func @add_dd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf32>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_10]]] : memref<32xf32>
 // CHECK:             %[[VAL_13:.*]] = arith.addf %[[VAL_11]], %[[VAL_12]] : f32
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_9]]{{\[}}%[[VAL_10]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
 // CHECK:           return %[[VAL_14]] : tensor<32xf32>
 // CHECK:         }
 func @add_dd(%arga: tensor<32xf32, #DV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #DV>, tensor<32xf32>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @mul_dd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf32>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_10]]] : memref<32xf32>
 // CHECK:             %[[VAL_13:.*]] = arith.mulf %[[VAL_11]], %[[VAL_12]] : f32
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_9]]{{\[}}%[[VAL_10]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
 // CHECK:           return %[[VAL_14]] : tensor<32xf32>
 // CHECK:         }
 func @mul_dd(%arga: tensor<32xf32, #DV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #DV>, tensor<32xf32>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @add_ds(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32xf32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]]:2 = scf.while (%[[VAL_16:.*]] = %[[VAL_13]], %[[VAL_17:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_18:.*]] = arith.cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
 // CHECK:             scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index
 // CHECK:           } do {

 // CHECK:             %[[VAL_29:.*]] = select %[[VAL_27]], %[[VAL_28]], %[[VAL_19]] : index
 // CHECK:             %[[VAL_30:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_29]], %[[VAL_30]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_31:.*]] = %[[VAL_32:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_33:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_31]]] : memref<32xf32>
 // CHECK:             memref.store %[[VAL_33]], %[[VAL_12]]{{\[}}%[[VAL_31]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_34:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xf32>
+// CHECK:           %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
 // CHECK:           return %[[VAL_34]] : tensor<32xf32>
 // CHECK:         }
 func @add_ds(%arga: tensor<32xf32>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32>, tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @mul_ds(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32xf32>
+// CHECK:           %[[VAL_5:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32xf32>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_13:.*]] = %[[VAL_11]] to %[[VAL_12]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_14]]] : memref<32xf32>
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_13]]] : memref<?xf32>
 // CHECK:             %[[VAL_17:.*]] = arith.mulf %[[VAL_15]], %[[VAL_16]] : f32
 // CHECK:             memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_14]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_18:.*]] = memref.tensor_load %[[VAL_10]] : memref<32xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf32>
 // CHECK:           return %[[VAL_18]] : tensor<32xf32>
 // CHECK:         }
 func @mul_ds(%arga: tensor<32xf32>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32>, tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @add_sd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]]:2 = scf.while (%[[VAL_16:.*]] = %[[VAL_13]], %[[VAL_17:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_18:.*]] = arith.cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
 // CHECK:             scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index
 // CHECK:           } do {

 // CHECK:             %[[VAL_29:.*]] = select %[[VAL_27]], %[[VAL_28]], %[[VAL_19]] : index
 // CHECK:             %[[VAL_30:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_29]], %[[VAL_30]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_31:.*]] = %[[VAL_32:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_31]]] : memref<32xf32>
 // CHECK:             memref.store %[[VAL_33]], %[[VAL_12]]{{\[}}%[[VAL_31]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_34:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xf32>
+// CHECK:           %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
 // CHECK:           return %[[VAL_34]] : tensor<32xf32>
 // CHECK:         }
 func @add_sd(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32xf32>
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @mul_sd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_13:.*]] = %[[VAL_11]] to %[[VAL_12]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xf32>
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_14]]] : memref<32xf32>
 // CHECK:             %[[VAL_17:.*]] = arith.mulf %[[VAL_15]], %[[VAL_16]] : f32
 // CHECK:             memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_14]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_18:.*]] = memref.tensor_load %[[VAL_10]] : memref<32xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf32>
 // CHECK:           return %[[VAL_18]] : tensor<32xf32>
 // CHECK:         }
 func @mul_sd(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]]:2 = scf.while (%[[VAL_18:.*]] = %[[VAL_13]], %[[VAL_19:.*]] = %[[VAL_15]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_20:.*]] = arith.cmpi ult, %[[VAL_18]], %[[VAL_14]] : index

 // CHECK:             %[[VAL_48:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_45]]] : memref<?xf32>
 // CHECK:             memref.store %[[VAL_48]], %[[VAL_12]]{{\[}}%[[VAL_47]]] : memref<32xf32>
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_49:.*]] = %[[VAL_50:.*]]#1 to %[[VAL_16]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_51:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_49]]] : memref<?xindex>
 // CHECK:             %[[VAL_52:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_49]]] : memref<?xf32>
 // CHECK:             memref.store %[[VAL_52]], %[[VAL_12]]{{\[}}%[[VAL_51]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_53:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xf32>
+// CHECK:           %[[VAL_53:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
 // CHECK:           return %[[VAL_53]] : tensor<32xf32>
 // CHECK:         }
 func @add_ss(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]]:2 = scf.while (%[[VAL_18:.*]] = %[[VAL_13]], %[[VAL_19:.*]] = %[[VAL_15]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_20:.*]] = arith.cmpi ult, %[[VAL_18]], %[[VAL_14]] : index

 // CHECK:             %[[VAL_35:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
 // CHECK:             %[[VAL_36:.*]] = arith.addi %[[VAL_23]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_37:.*]] = select %[[VAL_35]], %[[VAL_36]], %[[VAL_23]] : index
 // CHECK:             %[[VAL_38:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
 // CHECK:             %[[VAL_39:.*]] = arith.addi %[[VAL_24]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_40:.*]] = select %[[VAL_38]], %[[VAL_39]], %[[VAL_24]] : index
 // CHECK:             scf.yield %[[VAL_37]], %[[VAL_40]] : index, index
 // CHECK:           }
-// CHECK:           %[[VAL_41:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xf32>
+// CHECK:           %[[VAL_41:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
 // CHECK:           return %[[VAL_41]] : tensor<32xf32>
 // CHECK:         }
 func @mul_ss(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32, #SV>)
     outs(%argx: tensor<32xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<16xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<16xf32> to memref<16xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_15]] : index

 // CHECK:             memref.store %[[VAL_54]], %[[VAL_13]]{{\[}}%[[VAL_52]]] : memref<16xf32>
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_55:.*]] = %[[VAL_56:.*]]#1 to %[[VAL_17]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_57:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_55]]] : memref<?xindex>
 // CHECK:             %[[VAL_58:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_55]]] : memref<?xf32>
 // CHECK:             %[[VAL_59:.*]] = arith.mulf %[[VAL_58]], %[[VAL_2]] : f32
 // CHECK:             memref.store %[[VAL_59]], %[[VAL_13]]{{\[}}%[[VAL_57]]] : memref<16xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_60:.*]] = memref.tensor_load %[[VAL_13]] : memref<16xf32>
+// CHECK:           %[[VAL_60:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<16xf32>
 // CHECK:           return %[[VAL_60]] : tensor<16xf32>
 // CHECK:         }
 func @two_way_inv(%arga: tensor<16xf32, #SV>, %argb: tensor<16xf32, #SV>, %argc: f32, %argx: tensor<16xf32>) -> tensor<16xf32> {
   // Kernel "x(i) = a(i) * c + b(i) * c".
   %0 = linalg.generic #trait2
     ins(%arga, %argb: tensor<16xf32, #SV>, tensor<16xf32, #SV>)
     outs(%argx: tensor<16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):

 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<16xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<16xf32> to memref<16xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_15]] : index

 // CHECK:             memref.store %[[VAL_53]], %[[VAL_13]]{{\[}}%[[VAL_51]]] : memref<16xf32>
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_54:.*]] = %[[VAL_55:.*]]#1 to %[[VAL_17]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_56:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_54]]] : memref<?xindex>
 // CHECK:             %[[VAL_57:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_54]]] : memref<?xf32>
 // CHECK:             %[[VAL_58:.*]] = arith.mulf %[[VAL_57]], %[[VAL_2]] : f32
 // CHECK:             memref.store %[[VAL_58]], %[[VAL_13]]{{\[}}%[[VAL_56]]] : memref<16xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_59:.*]] = memref.tensor_load %[[VAL_13]] : memref<16xf32>
+// CHECK:           %[[VAL_59:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<16xf32>
 // CHECK:           return %[[VAL_59]] : tensor<16xf32>
 // CHECK:         }
 func @two_way_inv_alt(%arga: tensor<16xf32, #SV>,
                       %argb: tensor<16xf32, #SV>, %argc: f32, %argx: tensor<16xf32>) -> tensor<16xf32> {
   // Same kernel, but now expressed as "x(i) = (a(i) + b(i)) * c".
   %0 = linalg.generic #trait2
     ins(%arga, %argb: tensor<16xf32, #SV>, tensor<16xf32, #SV>)
     outs(%argx: tensor<16xf32>) {

 
 // CHECK-LABEL:   func @sum_reduction(
 // CHECK-SAME:      %[[VAL_0:.*]]: tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK-DAG:       %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_6:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK:           %[[VAL_6:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_7:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_6]], %[[VAL_7]] : memref<f32> to memref<f32>
 // CHECK-DAG:       %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK-DAG:       %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK-DAG:       %[[VAL_10:.*]] = memref.load %[[VAL_7]][] : memref<f32>
 // CHECK:           %[[VAL_11:.*]] = scf.for %[[VAL_12:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] iter_args(%[[VAL_13:.*]] = %[[VAL_10]]) -> (f32) {
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_12]]] : memref<?xf32>
 // CHECK:             %[[VAL_15:.*]] = arith.addf %[[VAL_13]], %[[VAL_14]] : f32
 // CHECK:             scf.yield %[[VAL_15]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_11]], %[[VAL_7]][] : memref<f32>
-// CHECK:           %[[VAL_17:.*]] = memref.tensor_load %[[VAL_7]] : memref<f32>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<f32>
 // CHECK:           return %[[VAL_17]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction(%arga: tensor<?xf32, #SV>, %argx: tensor<f32>) -> tensor<f32> {
   %0 = linalg.generic #trait_sum_reduction
     ins(%arga: tensor<?xf32, #SV>)
     outs(%argx: tensor<f32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %x, %a : f32

 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<f32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<f32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<f32> to memref<f32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_12]][] : memref<f32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]]:3 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]], %[[VAL_21:.*]] = %[[VAL_13]]) : (index, index, f32) -> (index, index, f32) {

 // CHECK:             scf.yield %[[VAL_62]] : f32
 // CHECK:           }
 // CHECK:           %[[VAL_63:.*]] = scf.for %[[VAL_64:.*]] = %[[VAL_65:.*]]#1 to %[[VAL_17]] step %[[VAL_4]] iter_args(%[[VAL_66:.*]] = %[[VAL_67:.*]]) -> (f32) {
 // CHECK:             %[[VAL_68:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_64]]] : memref<?xf32>
 // CHECK:             %[[VAL_69:.*]] = arith.addf %[[VAL_66]], %[[VAL_68]] : f32
 // CHECK:             scf.yield %[[VAL_69]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_70:.*]], %[[VAL_12]][] : memref<f32>
-// CHECK:           %[[VAL_71:.*]] = memref.tensor_load %[[VAL_12]] : memref<f32>
+// CHECK:           %[[VAL_71:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<f32>
 // CHECK:           return %[[VAL_71]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction_ss(%arga: tensor<16xf32, #SV>,
                        %argb: tensor<16xf32, #SV>,
                        %argx: tensor<f32>) -> tensor<f32> {
   // Just for testing. This case would be better expressed
   // as two separate reductions kernels.
   %0 = linalg.generic #trait_sum_reduction2

 // CHECK-SAME:      %[[VAL_1:.*1]]: tensor<f32>,
 // CHECK-SAME:      %[[VAL_2:.*2]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:      %[[VAL_3:.*3]]: tensor<f32>) -> tensor<f32> {
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_3]] : memref<f32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_3]] : memref<f32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<f32> to memref<f32>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_14]][] : memref<f32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_9]][] : memref<f32>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_5]]] : memref<?xindex>

 // CHECK:             scf.yield %[[VAL_68]] : f32
 // CHECK:           }
 // CHECK:           %[[VAL_69:.*]] = scf.for %[[VAL_70:.*]] = %[[VAL_71:.*]]#1 to %[[VAL_20]] step %[[VAL_5]] iter_args(%[[VAL_72:.*]] = %[[VAL_73:.*]]) -> (f32) {
 // CHECK:             %[[VAL_74:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_70]]] : memref<?xf32>
 // CHECK:             %[[VAL_75:.*]] = arith.addf %[[VAL_72]], %[[VAL_74]] : f32
 // CHECK:             scf.yield %[[VAL_75]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_76:.*]], %[[VAL_14]][] : memref<f32>
-// CHECK:           %[[VAL_77:.*]] = memref.tensor_load %[[VAL_14]] : memref<f32>
+// CHECK:           %[[VAL_77:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<f32>
 // CHECK:           return %[[VAL_77]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction_inv(%arga: tensor<16xf32, #SV>,
                         %argb: tensor<f32>,
                         %argc: tensor<16xf32, #SV>,
                         %argx: tensor<f32>) -> tensor<f32> {
   // Just for testing. This case would be better expressed
   // as two separate reductions kernels.

 // CHECK-SAME:                          %[[VAL_0:.*0]]: tensor<?xf64>,
 // CHECK-SAME:                          %[[VAL_1:.*1]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                          %[[VAL_2:.*2]]: tensor<?xf64>,
 // CHECK-SAME:                          %[[VAL_3:.*3]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                          %[[VAL_4:.*]]: tensor<?xf64>) -> tensor<?xf64> {
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_0]] : memref<?xf64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_0]] : memref<?xf64>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<?xf64>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<?xf64>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.pointers %[[VAL_3]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.indices %[[VAL_3]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = sparse_tensor.values %[[VAL_3]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
 // CHECK:           %[[VAL_16:.*]] = tensor.dim %[[VAL_4]], %[[VAL_5]] : tensor<?xf64>
-// CHECK:           %[[VAL_17:.*]] = memref.buffer_cast %[[VAL_4]] : memref<?xf64>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_memref %[[VAL_4]] : memref<?xf64>
 // CHECK:           %[[VAL_18:.*]] = memref.alloc(%[[VAL_16]]) : memref<?xf64>
 // CHECK:           memref.copy %[[VAL_17]], %[[VAL_18]] : memref<?xf64> to memref<?xf64>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_22:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_23:.*]]:3 = scf.while (%[[VAL_24:.*]] = %[[VAL_19]], %[[VAL_25:.*]] = %[[VAL_21]], %[[VAL_26:.*]] = %[[VAL_5]]) : (index, index, index) -> (index, index, index) {
 // CHECK:             %[[VAL_27:.*]] = arith.cmpi ult, %[[VAL_24]], %[[VAL_20]] : index

 // CHECK:             scf.yield %[[VAL_108]], %[[VAL_109]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_110:.*]] = %[[VAL_111:.*]]#1 to %[[VAL_16]] step %[[VAL_7]] {
 // CHECK:             %[[VAL_112:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_110]]] : memref<?xf64>
 // CHECK:             %[[VAL_113:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_110]]] : memref<?xf64>
 // CHECK:             %[[VAL_114:.*]] = arith.addf %[[VAL_112]], %[[VAL_113]] : f64
 // CHECK:             memref.store %[[VAL_114]], %[[VAL_18]]{{\[}}%[[VAL_110]]] : memref<?xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_115:.*]] = memref.tensor_load %[[VAL_18]] : memref<?xf64>
+// CHECK:           %[[VAL_115:.*]] = bufferization.to_tensor %[[VAL_18]] : memref<?xf64>
 // CHECK:           return %[[VAL_115]] : tensor<?xf64>
 // CHECK:         }
 func @four_tensors_op(%arga: tensor<?xf64>,
                       %argb: tensor<?xf64, #SV>,
                       %argc: tensor<?xf64>,
                       %argd: tensor<?xf64, #SV>,
                       %argx: tensor<?xf64>) -> tensor<?xf64> {
   %r = linalg.generic #trait_four_tensors

 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_3]] : memref<f64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_3]] : memref<f64>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<f64>
 // CHECK:           memref.copy %[[VAL_15]], %[[VAL_16]] : memref<f64> to memref<f64>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_16]][] : memref<f64>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_22:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_4]]] : memref<?xindex>

 // CHECK:             scf.yield %[[VAL_243]] : f64
 // CHECK:           }
 // CHECK:           %[[VAL_244:.*]] = scf.for %[[VAL_245:.*]] = %[[VAL_246:.*]]#0 to %[[VAL_19]] step %[[VAL_5]] iter_args(%[[VAL_247:.*]] = %[[VAL_248:.*]]) -> (f64) {
 // CHECK:             %[[VAL_249:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_245]]] : memref<?xf64>
 // CHECK:             %[[VAL_250:.*]] = arith.addf %[[VAL_247]], %[[VAL_249]] : f64
 // CHECK:             scf.yield %[[VAL_250]] : f64
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_251:.*]], %[[VAL_16]][] : memref<f64>
-// CHECK:           %[[VAL_252:.*]] = memref.tensor_load %[[VAL_16]] : memref<f64>
+// CHECK:           %[[VAL_252:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<f64>
 // CHECK:           return %[[VAL_252]] : tensor<f64>
 // CHECK:         }
 func @red3s(%arga: tensor<?xf64, #SV>,
             %argb: tensor<?xf64, #SV>,
       %argc: tensor<?xf64, #SV>, %argx: tensor<f64>) ->tensor<f64>{
  %0 = linalg.generic #trait_red3s
    ins(%arga, %argb, %argc: tensor<?xf64, #SV>, tensor<?xf64, #SV>, tensor<?xf64, #SV>)
    outs(%argx: tensor<f64>) {
      ^bb(%a: f64,%b: f64,%c: f64,%x: f64):
         %0 = arith.addf %x, %a : f64
         %1 = arith.addf %0, %b : f64
         %2 = arith.addf %1, %c : f64
       linalg.yield %2 : f64
     } -> tensor<f64>
   return %0 : tensor<f64>
 }

mlir/test/Dialect/SparseTensor/sparse_2d.mlir

 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_12:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_13:.*]] = arith.muli %[[VAL_11]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_14:.*]] = arith.addi %[[VAL_13]], %[[VAL_12]] : index
 // CHECK:               %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_14]]] : memref<?xf32>
 // CHECK:               %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_11]], %[[VAL_12]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_17:.*]] = arith.addf %[[VAL_15]], %[[VAL_16]] : f32
 // CHECK:               memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_11]], %[[VAL_12]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_18:.*]] = memref.tensor_load %[[VAL_10]] : memref<32x16xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_18]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_dd(%arga: tensor<32x16xf32, #Tdd>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tdd>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16xf32>
   return %0 : tensor<32x16xf32>
 }
 
 // CHECK-LABEL:   func @mul_dd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_12:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_13:.*]] = arith.muli %[[VAL_11]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_14:.*]] = arith.addi %[[VAL_13]], %[[VAL_12]] : index
 // CHECK:               %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_14]]] : memref<?xf32>
 // CHECK:               %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_11]], %[[VAL_12]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_17:.*]] = arith.mulf %[[VAL_15]], %[[VAL_16]] : f32
 // CHECK:               memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_11]], %[[VAL_12]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_18:.*]] = memref.tensor_load %[[VAL_10]] : memref<32x16xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_18]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_dd(%arga: tensor<32x16xf32, #Tdd>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tdd>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_7]] {
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             %[[VAL_16:.*]] = arith.addi %[[VAL_14]], %[[VAL_7]] : index
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_15]], %[[VAL_20:.*]] = %[[VAL_5]]) : (index, index) -> (index, index) {
 // CHECK:               %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_17]] : index

 // CHECK:               %[[VAL_33:.*]] = arith.addi %[[VAL_23]], %[[VAL_7]] : index
 // CHECK:               scf.yield %[[VAL_32]], %[[VAL_33]] : index, index
 // CHECK:             }
 // CHECK:             scf.for %[[VAL_34:.*]] = %[[VAL_35:.*]]#1 to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_36:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_14]], %[[VAL_34]]] : memref<32x16xf32>
 // CHECK:               memref.store %[[VAL_36]], %[[VAL_13]]{{\[}}%[[VAL_14]], %[[VAL_34]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_37:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16xf32>
+// CHECK:           %[[VAL_37:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_37]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_ds(%arga: tensor<32x16xf32, #Tds>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tds>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16xf32>
   return %0 : tensor<32x16xf32>
 }
 
 // CHECK-LABEL:   func @mul_ds(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = arith.addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_16:.*]] = %[[VAL_13]] to %[[VAL_15]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:               %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xf32>
 // CHECK:               %[[VAL_19:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_12]], %[[VAL_17]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_20:.*]] = arith.mulf %[[VAL_18]], %[[VAL_19]] : f32
 // CHECK:               memref.store %[[VAL_20]], %[[VAL_11]]{{\[}}%[[VAL_12]], %[[VAL_17]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_21:.*]] = memref.tensor_load %[[VAL_11]] : memref<32x16xf32>
+// CHECK:           %[[VAL_21:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_21]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_ds(%arga: tensor<32x16xf32, #Tds>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tds>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]]:2 = scf.while (%[[VAL_17:.*]] = %[[VAL_14]], %[[VAL_18:.*]] = %[[VAL_6]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_19:.*]] = arith.cmpi ult, %[[VAL_17]], %[[VAL_15]] : index
 // CHECK:             scf.condition(%[[VAL_19]]) %[[VAL_17]], %[[VAL_18]] : index, index
 // CHECK:           } do {

 // CHECK:             scf.yield %[[VAL_34]], %[[VAL_35]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_36:.*]] = %[[VAL_37:.*]]#1 to %[[VAL_3]] step %[[VAL_7]] {
 // CHECK:             scf.for %[[VAL_38:.*]] = %[[VAL_6]] to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_39:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_36]], %[[VAL_38]]] : memref<32x16xf32>
 // CHECK:               memref.store %[[VAL_39]], %[[VAL_13]]{{\[}}%[[VAL_36]], %[[VAL_38]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_40:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16xf32>
+// CHECK:           %[[VAL_40:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_40]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_sd(%arga: tensor<32x16xf32, #Tsd>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tsd>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16xf32>
   return %0 : tensor<32x16xf32>
 }
 
 // CHECK-LABEL:   func @mul_sd(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_12]] to %[[VAL_13]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_16:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_17:.*]] = arith.muli %[[VAL_14]], %[[VAL_3]] : index
 // CHECK:               %[[VAL_18:.*]] = arith.addi %[[VAL_17]], %[[VAL_16]] : index
 // CHECK:               %[[VAL_19:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xf32>
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_15]], %[[VAL_16]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_21:.*]] = arith.mulf %[[VAL_19]], %[[VAL_20]] : f32
 // CHECK:               memref.store %[[VAL_21]], %[[VAL_11]]{{\[}}%[[VAL_15]], %[[VAL_16]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_22:.*]] = memref.tensor_load %[[VAL_11]] : memref<32x16xf32>
+// CHECK:           %[[VAL_22:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_22]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_sd(%arga: tensor<32x16xf32, #Tsd>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tsd>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_14]], %[[VAL_15]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_16]], %[[VAL_20:.*]] = %[[VAL_6]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_17]] : index
 // CHECK:             scf.condition(%[[VAL_21]]) %[[VAL_19]], %[[VAL_20]] : index, index
 // CHECK:           } do {

 // CHECK:             scf.yield %[[VAL_52]], %[[VAL_53]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_54:.*]] = %[[VAL_55:.*]]#1 to %[[VAL_3]] step %[[VAL_7]] {
 // CHECK:             scf.for %[[VAL_56:.*]] = %[[VAL_6]] to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_57:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_54]], %[[VAL_56]]] : memref<32x16xf32>
 // CHECK:               memref.store %[[VAL_57]], %[[VAL_15]]{{\[}}%[[VAL_54]], %[[VAL_56]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_58:.*]] = memref.tensor_load %[[VAL_15]] : memref<32x16xf32>
+// CHECK:           %[[VAL_58:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_58]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_ss(%arga: tensor<32x16xf32, #Tss>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tss>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16xf32>
   return %0 : tensor<32x16xf32>
 }
 
 // CHECK-LABEL:   func @mul_ss(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_15:.*]] = %[[VAL_13]] to %[[VAL_14]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = arith.addi %[[VAL_15]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_19:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_20:.*]] = %[[VAL_17]] to %[[VAL_19]] step %[[VAL_4]] {
 // CHECK:               %[[VAL_21:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_20]]] : memref<?xf32>
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_16]], %[[VAL_21]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_24:.*]] = arith.mulf %[[VAL_22]], %[[VAL_23]] : f32
 // CHECK:               memref.store %[[VAL_24]], %[[VAL_12]]{{\[}}%[[VAL_16]], %[[VAL_21]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_25:.*]] = memref.tensor_load %[[VAL_12]] : memref<32x16xf32>
+// CHECK:           %[[VAL_25:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_25]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_ss(%arga: tensor<32x16xf32, #Tss>, %argb: tensor<32x16xf32>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tss>, tensor<32x16xf32>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_15]], %[[VAL_16]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]]:2 = scf.while (%[[VAL_22:.*]] = %[[VAL_17]], %[[VAL_23:.*]] = %[[VAL_19]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_24:.*]] = arith.cmpi ult, %[[VAL_22]], %[[VAL_18]] : index

 // CHECK:             %[[VAL_111:.*]] = arith.addi %[[VAL_107]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_112:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_111]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_113:.*]] = %[[VAL_110]] to %[[VAL_112]] step %[[VAL_4]] {
 // CHECK:               %[[VAL_114:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_113]]] : memref<?xindex>
 // CHECK:               %[[VAL_115:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_113]]] : memref<?xf32>
 // CHECK:               memref.store %[[VAL_115]], %[[VAL_16]]{{\[}}%[[VAL_109]], %[[VAL_114]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_116:.*]] = memref.tensor_load %[[VAL_16]] : memref<32x16xf32>
+// CHECK:           %[[VAL_116:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_116]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_ss_ss(%arga: tensor<32x16xf32, #Tss>, %argb: tensor<32x16xf32, #Tss>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tss>, tensor<32x16xf32, #Tss>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_15]], %[[VAL_16]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]]:2 = scf.while (%[[VAL_22:.*]] = %[[VAL_17]], %[[VAL_23:.*]] = %[[VAL_19]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_24:.*]] = arith.cmpi ult, %[[VAL_22]], %[[VAL_18]] : index

 // CHECK:             %[[VAL_66:.*]] = arith.cmpi eq, %[[VAL_29]], %[[VAL_32]] : index
 // CHECK:             %[[VAL_67:.*]] = arith.addi %[[VAL_27]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_68:.*]] = select %[[VAL_66]], %[[VAL_67]], %[[VAL_27]] : index
 // CHECK:             %[[VAL_69:.*]] = arith.cmpi eq, %[[VAL_30]], %[[VAL_32]] : index
 // CHECK:             %[[VAL_70:.*]] = arith.addi %[[VAL_28]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_71:.*]] = select %[[VAL_69]], %[[VAL_70]], %[[VAL_28]] : index
 // CHECK:             scf.yield %[[VAL_68]], %[[VAL_71]] : index, index
 // CHECK:           }
-// CHECK:           %[[VAL_72:.*]] = memref.tensor_load %[[VAL_16]] : memref<32x16xf32>
+// CHECK:           %[[VAL_72:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_72]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_ss_ss(%arga: tensor<32x16xf32, #Tss>, %argb: tensor<32x16xf32, #Tss>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tss>, tensor<32x16xf32, #Tss>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_14]], %[[VAL_15]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_16]], %[[VAL_20:.*]] = %[[VAL_5]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_17]] : index
 // CHECK:             scf.condition(%[[VAL_21]]) %[[VAL_19]], %[[VAL_20]] : index, index
 // CHECK:           } do {

 // CHECK:             %[[VAL_65:.*]] = arith.addi %[[VAL_62]], %[[VAL_7]] : index
 // CHECK:             %[[VAL_66:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_65]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_67:.*]] = %[[VAL_64]] to %[[VAL_66]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_68:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_67]]] : memref<?xindex>
 // CHECK:               %[[VAL_69:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_67]]] : memref<?xf32>
 // CHECK:               memref.store %[[VAL_69]], %[[VAL_15]]{{\[}}%[[VAL_62]], %[[VAL_68]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_70:.*]] = memref.tensor_load %[[VAL_15]] : memref<32x16xf32>
+// CHECK:           %[[VAL_70:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_70]] : tensor<32x16xf32>
 // CHECK:         }
 func @add_sd_ds(%arga: tensor<32x16xf32, #Tsd>, %argb: tensor<32x16xf32, #Tds>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tsd>, tensor<32x16xf32, #Tds>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_5]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16xf32> to memref<32x16xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_16:.*]] = %[[VAL_14]] to %[[VAL_15]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]] = arith.addi %[[VAL_17]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_21:.*]] = %[[VAL_18]] to %[[VAL_20]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:               %[[VAL_23:.*]] = arith.muli %[[VAL_16]], %[[VAL_3]] : index
 // CHECK:               %[[VAL_24:.*]] = arith.addi %[[VAL_23]], %[[VAL_22]] : index
 // CHECK:               %[[VAL_25:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK:               %[[VAL_26:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_21]]] : memref<?xf32>
 // CHECK:               %[[VAL_27:.*]] = arith.mulf %[[VAL_25]], %[[VAL_26]] : f32
 // CHECK:               memref.store %[[VAL_27]], %[[VAL_13]]{{\[}}%[[VAL_17]], %[[VAL_22]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_28:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16xf32>
+// CHECK:           %[[VAL_28:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_28]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul_sd_ds(%arga: tensor<32x16xf32, #Tsd>, %argb: tensor<32x16xf32, #Tds>, %argx: tensor<32x16xf32>) -> tensor<32x16xf32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32x16xf32, #Tsd>, tensor<32x16xf32, #Tds>)
     outs(%argx: tensor<32x16xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:      %[[VAL_2:.*]]: tensor<16xf32>) -> tensor<16xf32> {
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 16 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16x32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<16xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<16xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<16xf32> to memref<16xf32>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-DAG:         %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK-DAG:         %[[VAL_14:.*]] = arith.addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK-DAG:         %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK-DAG:         %[[VAL_16:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_12]]] : memref<16xf32>
 // CHECK:             %[[VAL_17:.*]] = scf.for %[[VAL_18:.*]] = %[[VAL_13]] to %[[VAL_15]] step %[[VAL_5]] iter_args(%[[VAL_19:.*]] = %[[VAL_16]]) -> (f32) {
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:               %[[VAL_21:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xf32>
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_20]]] : memref<32xf32>
 // CHECK:               %[[VAL_23:.*]] = arith.mulf %[[VAL_21]], %[[VAL_22]] : f32
 // CHECK:               %[[VAL_24:.*]] = arith.addf %[[VAL_23]], %[[VAL_19]] : f32
 // CHECK:               scf.yield %[[VAL_24]] : f32
 // CHECK:             }
 // CHECK:             memref.store %[[VAL_17]], %[[VAL_11]]{{\[}}%[[VAL_12]]] : memref<16xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_26:.*]] = memref.tensor_load %[[VAL_11]] : memref<16xf32>
+// CHECK:           %[[VAL_26:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<16xf32>
 // CHECK:           return %[[VAL_26]] : tensor<16xf32>
 // CHECK:         }
 func @matvec(%argA: tensor<16x32xf32, #Tds>, %argb: tensor<32xf32>, %argx: tensor<16xf32>) -> tensor<16xf32> {
   %0 = linalg.generic #trait_matvec
        ins(%argA, %argb: tensor<16x32xf32, #Tds>, tensor<32xf32>)
       outs(%argx: tensor<16xf32>) {
     ^bb(%A: f32, %b: f32, %x: f32):
       %0 = arith.mulf %A, %b : f32

 // CHECK-LABEL:   func @sum_reduction(
 // CHECK-SAME:      %[[VAL_0:.*]]: tensor<10x20xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK-DAG:       %[[VAL_2:.*]] = arith.constant 10 : index
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<10x20xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_7]], %[[VAL_8]] : memref<f32> to memref<f32>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_8]][] : memref<f32>
 // CHECK:           %[[VAL_10:.*]] = scf.for %[[VAL_11:.*]] = %[[VAL_4]] to %[[VAL_2]] step %[[VAL_3]] iter_args(%[[VAL_12:.*]] = %[[VAL_9]]) -> (f32) {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = arith.addi %[[VAL_11]], %[[VAL_3]] : index
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             %[[VAL_16:.*]] = scf.for %[[VAL_17:.*]] = %[[VAL_13]] to %[[VAL_15]] step %[[VAL_3]] iter_args(%[[VAL_18:.*]] = %[[VAL_12]]) -> (f32) {
 // CHECK:               %[[VAL_19:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_17]]] : memref<?xf32>
 // CHECK:               %[[VAL_20:.*]] = arith.addf %[[VAL_18]], %[[VAL_19]] : f32
 // CHECK:               scf.yield %[[VAL_20]] : f32
 // CHECK:             }
 // CHECK:             scf.yield %[[VAL_16]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_10]], %[[VAL_8]][] : memref<f32>
-// CHECK:           %[[VAL_23:.*]] = memref.tensor_load %[[VAL_8]] : memref<f32>
+// CHECK:           %[[VAL_23:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<f32>
 // CHECK:           return %[[VAL_23]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction(%arga: tensor<10x20xf32, #Tds>, %argx: tensor<f32>) -> tensor<f32> {
   %0 = linalg.generic #trait_sum_reduction
      ins(%arga: tensor<10x20xf32, #Tds>)
     outs(%argx: tensor<f32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %x, %a : f32

 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2.000000e+00 : f64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
 // CHECK:           %[[VAL_8:.*]] = tensor.dim %[[VAL_1]], %[[VAL_3]] : tensor<?x?xf64>
 // CHECK:           %[[VAL_9:.*]] = tensor.dim %[[VAL_1]], %[[VAL_4]] : tensor<?x?xf64>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<?x?xf64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<?x?xf64>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc(%[[VAL_8]], %[[VAL_9]]) : memref<?x?xf64>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<?x?xf64> to memref<?x?xf64>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_3]] to %[[VAL_8]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = arith.addi %[[VAL_12]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_16:.*]] = %[[VAL_13]] to %[[VAL_15]] step %[[VAL_4]] {
 // CHECK:               %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:               %[[VAL_18:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xf64>
 // CHECK:               %[[VAL_19:.*]] = arith.mulf %[[VAL_18]], %[[VAL_2]] : f64
 // CHECK:               memref.store %[[VAL_19]], %[[VAL_11]]{{\[}}%[[VAL_12]], %[[VAL_17]]] : memref<?x?xf64>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_20:.*]] = memref.tensor_load %[[VAL_11]] : memref<?x?xf64>
+// CHECK:           %[[VAL_20:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<?x?xf64>
 // CHECK:           return %[[VAL_20]] : tensor<?x?xf64>
 // CHECK:         }
 func @scale(%arga: tensor<?x?xf64, #Tds>, %argx: tensor<?x?xf64>) -> tensor<?x?xf64> {
   %0 = arith.constant 2.0 : f64
   %1 = linalg.generic #trait_scale
      ins(%arga: tensor<?x?xf64, #Tds>)
     outs(%argx: tensor<?x?xf64>) {
       ^bb(%a: f64, %x: f64):

 // CHECK-SAME:      %[[VAL_3:.*3]]: tensor<?x?xf32>) -> tensor<?x?xf32> {
 // CHECK:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<?x?xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<?x?xf32>
 // CHECK:           %[[VAL_12:.*]] = tensor.dim %[[VAL_2]], %[[VAL_4]] : tensor<?x?xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<?x?xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<?x?xf32>
 // CHECK:           %[[VAL_14:.*]] = tensor.dim %[[VAL_3]], %[[VAL_4]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_15:.*]] = tensor.dim %[[VAL_3]], %[[VAL_5]] : tensor<?x?xf32>
-// CHECK:           %[[VAL_16:.*]] = memref.buffer_cast %[[VAL_3]] : memref<?x?xf32>
+// CHECK:           %[[VAL_16:.*]] = bufferization.to_memref %[[VAL_3]] : memref<?x?xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.alloc(%[[VAL_14]], %[[VAL_15]]) : memref<?x?xf32>
 // CHECK:           memref.copy %[[VAL_16]], %[[VAL_17]] : memref<?x?xf32> to memref<?x?xf32>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_20:.*]] = %[[VAL_18]] to %[[VAL_19]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:             %[[VAL_22:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:             %[[VAL_23:.*]] = arith.addi %[[VAL_20]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_24:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_23]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_25:.*]] = %[[VAL_22]] to %[[VAL_24]] step %[[VAL_5]] {
 // CHECK-DAG:           %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_25]]] : memref<?xindex>
 // CHECK-DAG:           %[[VAL_27:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_25]]] : memref<?xf32>
 // CHECK-DAG:           %[[VAL_28:.*]] = memref.load %[[VAL_17]]{{\[}}%[[VAL_21]], %[[VAL_26]]] : memref<?x?xf32>
 // CHECK:               %[[VAL_29:.*]] = scf.for %[[VAL_30:.*]] = %[[VAL_4]] to %[[VAL_12]] step %[[VAL_5]] iter_args(%[[VAL_31:.*]] = %[[VAL_28]]) -> (f32) {
 // CHECK:                 %[[VAL_32:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_21]], %[[VAL_30]]] : memref<?x?xf32>
 // CHECK:                 %[[VAL_33:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_30]], %[[VAL_26]]] : memref<?x?xf32>
 // CHECK:                 %[[VAL_34:.*]] = arith.mulf %[[VAL_32]], %[[VAL_33]] : f32
 // CHECK:                 %[[VAL_35:.*]] = arith.mulf %[[VAL_27]], %[[VAL_34]] : f32
 // CHECK:                 %[[VAL_36:.*]] = arith.addf %[[VAL_31]], %[[VAL_35]] : f32
 // CHECK:                 scf.yield %[[VAL_36]] : f32
 // CHECK:               }
 // CHECK:               memref.store %[[VAL_29]], %[[VAL_17]]{{\[}}%[[VAL_21]], %[[VAL_26]]] : memref<?x?xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_38:.*]] = memref.tensor_load %[[VAL_17]] : memref<?x?xf32>
+// CHECK:           %[[VAL_38:.*]] = bufferization.to_tensor %[[VAL_17]] : memref<?x?xf32>
 // CHECK:           return %[[VAL_38]] : tensor<?x?xf32>
 // CHECK:         }
 func @sampled_dense_dense(%args: tensor<?x?xf32, #Tss>,
                           %arga: tensor<?x?xf32>,
                           %argb: tensor<?x?xf32>,
                           %argx: tensor<?x?xf32>) -> tensor<?x?xf32> {
   %0 = linalg.generic #trait_sampled_dense_dense
      ins(%args, %arga, %argb: tensor<?x?xf32, #Tss>, tensor<?x?xf32>, tensor<?x?xf32>)

 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_7]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_7]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_17:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_7]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_7]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_20:.*]] = memref.buffer_cast %[[VAL_3]] : memref<?xf32>
-// CHECK:           %[[VAL_21:.*]] = memref.buffer_cast %[[VAL_4]] : memref<f32>
+// CHECK:           %[[VAL_20:.*]] = bufferization.to_memref %[[VAL_3]] : memref<?xf32>
+// CHECK:           %[[VAL_21:.*]] = bufferization.to_memref %[[VAL_4]] : memref<f32>
 // CHECK:           %[[VAL_22:.*]] = tensor.dim %[[VAL_5]], %[[VAL_6]] : tensor<?xf32>
-// CHECK:           %[[VAL_23:.*]] = memref.buffer_cast %[[VAL_5]] : memref<?xf32>
+// CHECK:           %[[VAL_23:.*]] = bufferization.to_memref %[[VAL_5]] : memref<?xf32>
 // CHECK:           %[[VAL_24:.*]] = memref.alloc(%[[VAL_22]]) : memref<?xf32>
 // CHECK:           memref.copy %[[VAL_23]], %[[VAL_24]] : memref<?xf32> to memref<?xf32>
 // CHECK:           %[[VAL_25:.*]] = memref.load %[[VAL_21]][] : memref<f32>
 // CHECK:           %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_27:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_28:.*]]:2 = scf.while (%[[VAL_29:.*]] = %[[VAL_26]], %[[VAL_30:.*]] = %[[VAL_6]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_31:.*]] = arith.cmpi ult, %[[VAL_29]], %[[VAL_27]] : index
 // CHECK:             scf.condition(%[[VAL_31]]) %[[VAL_29]], %[[VAL_30]] : index, index

 // CHECK:             %[[VAL_167:.*]] = memref.load %[[VAL_17]]{{\[}}%[[VAL_166]]] : memref<?xindex>
 // CHECK:             %[[VAL_168:.*]] = scf.for %[[VAL_169:.*]] = %[[VAL_165]] to %[[VAL_167]] step %[[VAL_7]] iter_args(%[[VAL_170:.*]] = %[[VAL_164]]) -> (f32) {
 // CHECK:               %[[VAL_171:.*]] = memref.load %[[VAL_19]]{{\[}}%[[VAL_169]]] : memref<?xf32>
 // CHECK:               %[[VAL_172:.*]] = arith.addf %[[VAL_170]], %[[VAL_171]] : f32
 // CHECK:               scf.yield %[[VAL_172]] : f32
 // CHECK:             }
 // CHECK:             memref.store %[[VAL_173:.*]], %[[VAL_24]]{{\[}}%[[VAL_162]]] : memref<?xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_174:.*]] = memref.tensor_load %[[VAL_24]] : memref<?xf32>
+// CHECK:           %[[VAL_174:.*]] = bufferization.to_tensor %[[VAL_24]] : memref<?xf32>
 // CHECK:           return %[[VAL_174]] : tensor<?xf32>
 // CHECK:         }
 func @sum_kernel_with_inv(%arga: tensor<?x?xf32, #Tss>,
                           %argb: tensor<?x?xf32, #Tds>,
                           %argc: tensor<?x?xf32, #Tds>,
                           %argd: tensor<?xf32>,
                           %arge: tensor<f32>,
                           %argx: tensor<?xf32>) -> tensor<?xf32> {

mlir/test/Dialect/SparseTensor/sparse_3d.mlir

 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_6]] to %[[VAL_3]] step %[[VAL_7]] {
 // CHECK:             scf.for %[[VAL_13:.*]] = %[[VAL_6]] to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_14:.*]] = arith.muli %[[VAL_12]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_15:.*]] = arith.addi %[[VAL_14]], %[[VAL_13]] : index
 // CHECK:               scf.for %[[VAL_16:.*]] = %[[VAL_6]] to %[[VAL_5]] step %[[VAL_7]] {
 // CHECK:                 %[[VAL_17:.*]] = arith.muli %[[VAL_15]], %[[VAL_5]] : index
 // CHECK:                 %[[VAL_18:.*]] = arith.addi %[[VAL_17]], %[[VAL_16]] : index
 // CHECK:                 %[[VAL_19:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xf32>
 // CHECK:                 %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_12]], %[[VAL_13]], %[[VAL_16]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_21:.*]] = arith.addf %[[VAL_19]], %[[VAL_20]] : f32
 // CHECK:                 memref.store %[[VAL_21]], %[[VAL_11]]{{\[}}%[[VAL_12]], %[[VAL_13]], %[[VAL_16]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_22:.*]] = memref.tensor_load %[[VAL_11]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_22:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_22]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_ddd(%arga: tensor<32x16x8xf32, #Tddd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tddd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16x8xf32>
   return %0 : tensor<32x16x8xf32>
 }
 
 // CHECK-LABEL:   func @mul_ddd(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_6]] to %[[VAL_3]] step %[[VAL_7]] {
 // CHECK:             scf.for %[[VAL_13:.*]] = %[[VAL_6]] to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_14:.*]] = arith.muli %[[VAL_12]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_15:.*]] = arith.addi %[[VAL_14]], %[[VAL_13]] : index
 // CHECK:               scf.for %[[VAL_16:.*]] = %[[VAL_6]] to %[[VAL_5]] step %[[VAL_7]] {
 // CHECK:                 %[[VAL_17:.*]] = arith.muli %[[VAL_15]], %[[VAL_5]] : index
 // CHECK:                 %[[VAL_18:.*]] = arith.addi %[[VAL_17]], %[[VAL_16]] : index
 // CHECK:                 %[[VAL_19:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xf32>
 // CHECK:                 %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_12]], %[[VAL_13]], %[[VAL_16]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_21:.*]] = arith.mulf %[[VAL_19]], %[[VAL_20]] : f32
 // CHECK:                 memref.store %[[VAL_21]], %[[VAL_11]]{{\[}}%[[VAL_12]], %[[VAL_13]], %[[VAL_16]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_22:.*]] = memref.tensor_load %[[VAL_11]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_22:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_22]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_ddd(%arga: tensor<32x16x8xf32, #Tddd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tddd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_9:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_14]], %[[VAL_15]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_16:.*]] = %[[VAL_7]] to %[[VAL_4]] step %[[VAL_9]] {
 // CHECK:             scf.for %[[VAL_17:.*]] = %[[VAL_7]] to %[[VAL_5]] step %[[VAL_9]] {
 // CHECK:               %[[VAL_18:.*]] = arith.muli %[[VAL_16]], %[[VAL_5]] : index
 // CHECK:               %[[VAL_19:.*]] = arith.addi %[[VAL_18]], %[[VAL_17]] : index
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:               %[[VAL_21:.*]] = arith.addi %[[VAL_19]], %[[VAL_9]] : index

 // CHECK:                 scf.yield %[[VAL_37]], %[[VAL_38]] : index, index
 // CHECK:               }
 // CHECK:               scf.for %[[VAL_39:.*]] = %[[VAL_40:.*]]#1 to %[[VAL_6]] step %[[VAL_9]] {
 // CHECK:                 %[[VAL_41:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_16]], %[[VAL_17]], %[[VAL_39]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_41]], %[[VAL_15]]{{\[}}%[[VAL_16]], %[[VAL_17]], %[[VAL_39]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_42:.*]] = memref.tensor_load %[[VAL_15]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_42:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_42]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_dds(%arga: tensor<32x16x8xf32, #Tdds>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdds>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK:           %[[VAL_3:.*]] = arith.constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_6]] to %[[VAL_4]] step %[[VAL_7]] {
 // CHECK:             scf.for %[[VAL_15:.*]] = %[[VAL_6]] to %[[VAL_5]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_16:.*]] = arith.muli %[[VAL_14]], %[[VAL_5]] : index
 // CHECK:               %[[VAL_17:.*]] = arith.addi %[[VAL_16]], %[[VAL_15]] : index
 // CHECK:               %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:               %[[VAL_19:.*]] = arith.addi %[[VAL_17]], %[[VAL_7]] : index
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_21:.*]] = %[[VAL_18]] to %[[VAL_20]] step %[[VAL_7]] {
 // CHECK:                 %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:                 %[[VAL_23:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_21]]] : memref<?xf32>
 // CHECK:                 %[[VAL_24:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_14]], %[[VAL_15]], %[[VAL_22]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_25:.*]] = arith.mulf %[[VAL_23]], %[[VAL_24]] : f32
 // CHECK:                 memref.store %[[VAL_25]], %[[VAL_13]]{{\[}}%[[VAL_14]], %[[VAL_15]], %[[VAL_22]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_26:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_26:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_26]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_dds(%arga: tensor<32x16x8xf32, #Tdds>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdds>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_15:.*]] = %[[VAL_7]] to %[[VAL_3]] step %[[VAL_8]] {
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_17:.*]] = arith.addi %[[VAL_15]], %[[VAL_8]] : index
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]]:2 = scf.while (%[[VAL_20:.*]] = %[[VAL_16]], %[[VAL_21:.*]] = %[[VAL_7]]) : (index, index) -> (index, index) {
 // CHECK:               %[[VAL_22:.*]] = arith.cmpi ult, %[[VAL_20]], %[[VAL_18]] : index

 // CHECK:             }
 // CHECK:             scf.for %[[VAL_39:.*]] = %[[VAL_40:.*]]#1 to %[[VAL_4]] step %[[VAL_8]] {
 // CHECK:               scf.for %[[VAL_41:.*]] = %[[VAL_7]] to %[[VAL_5]] step %[[VAL_8]] {
 // CHECK:                 %[[VAL_42:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_15]], %[[VAL_39]], %[[VAL_41]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_42]], %[[VAL_14]]{{\[}}%[[VAL_15]], %[[VAL_39]], %[[VAL_41]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_43:.*]] = memref.tensor_load %[[VAL_14]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_43:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_43]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_dsd(%arga: tensor<32x16x8xf32, #Tdsd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdsd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16x8xf32>
   return %0 : tensor<32x16x8xf32>
 }
 
 // CHECK-LABEL:   func @mul_dsd(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_13:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK:             %[[VAL_15:.*]] = arith.addi %[[VAL_13]], %[[VAL_6]] : index
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_17:.*]] = %[[VAL_14]] to %[[VAL_16]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_19:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_20:.*]] = arith.muli %[[VAL_17]], %[[VAL_4]] : index
 // CHECK:                 %[[VAL_21:.*]] = arith.addi %[[VAL_20]], %[[VAL_19]] : index
 // CHECK:                 %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xf32>
 // CHECK:                 %[[VAL_23:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_13]], %[[VAL_18]], %[[VAL_19]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_24:.*]] = arith.mulf %[[VAL_22]], %[[VAL_23]] : f32
 // CHECK:                 memref.store %[[VAL_24]], %[[VAL_12]]{{\[}}%[[VAL_13]], %[[VAL_18]], %[[VAL_19]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_25:.*]] = memref.tensor_load %[[VAL_12]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_25:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_25]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_dsd(%arga: tensor<32x16x8xf32, #Tdsd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdsd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_9:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_16:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_16:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_16]], %[[VAL_17]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_18:.*]] = %[[VAL_8]] to %[[VAL_4]] step %[[VAL_9]] {
 // CHECK:             %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_20:.*]] = arith.addi %[[VAL_18]], %[[VAL_9]] : index
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:             %[[VAL_22:.*]]:2 = scf.while (%[[VAL_23:.*]] = %[[VAL_19]], %[[VAL_24:.*]] = %[[VAL_8]]) : (index, index) -> (index, index) {
 // CHECK:               %[[VAL_25:.*]] = arith.cmpi ult, %[[VAL_23]], %[[VAL_21]] : index

 // CHECK:             }
 // CHECK:             scf.for %[[VAL_58:.*]] = %[[VAL_59:.*]]#1 to %[[VAL_5]] step %[[VAL_9]] {
 // CHECK:               scf.for %[[VAL_60:.*]] = %[[VAL_8]] to %[[VAL_6]] step %[[VAL_9]] {
 // CHECK:                 %[[VAL_61:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_18]], %[[VAL_58]], %[[VAL_60]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_61]], %[[VAL_17]]{{\[}}%[[VAL_18]], %[[VAL_58]], %[[VAL_60]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_62:.*]] = memref.tensor_load %[[VAL_17]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_62:.*]] = bufferization.to_tensor %[[VAL_17]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_62]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_dss(%arga: tensor<32x16x8xf32, #Tdss>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdss>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           scf.for %[[VAL_15:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_17:.*]] = arith.addi %[[VAL_15]], %[[VAL_6]] : index
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_19:.*]] = %[[VAL_16]] to %[[VAL_18]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:               %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:               %[[VAL_22:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_22]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_24:.*]] = %[[VAL_21]] to %[[VAL_23]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_25:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_24]]] : memref<?xindex>
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK:                 %[[VAL_27:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_15]], %[[VAL_20]], %[[VAL_25]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_28:.*]] = arith.mulf %[[VAL_26]], %[[VAL_27]] : f32
 // CHECK:                 memref.store %[[VAL_28]], %[[VAL_14]]{{\[}}%[[VAL_15]], %[[VAL_20]], %[[VAL_25]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_29:.*]] = memref.tensor_load %[[VAL_14]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_29:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_29]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_dss(%arga: tensor<32x16x8xf32, #Tdss>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tdss>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_8]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]]:2 = scf.while (%[[VAL_18:.*]] = %[[VAL_15]], %[[VAL_19:.*]] = %[[VAL_7]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_20:.*]] = arith.cmpi ult, %[[VAL_18]], %[[VAL_16]] : index
 // CHECK:             scf.condition(%[[VAL_20]]) %[[VAL_18]], %[[VAL_19]] : index, index
 // CHECK:           } do {

 // CHECK:           scf.for %[[VAL_41:.*]] = %[[VAL_42:.*]]#1 to %[[VAL_3]] step %[[VAL_8]] {
 // CHECK:             scf.for %[[VAL_43:.*]] = %[[VAL_7]] to %[[VAL_4]] step %[[VAL_8]] {
 // CHECK:               scf.for %[[VAL_44:.*]] = %[[VAL_7]] to %[[VAL_5]] step %[[VAL_8]] {
 // CHECK:                 %[[VAL_45:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_41]], %[[VAL_43]], %[[VAL_44]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_45]], %[[VAL_14]]{{\[}}%[[VAL_41]], %[[VAL_43]], %[[VAL_44]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_46:.*]] = memref.tensor_load %[[VAL_14]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_46:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_46]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_sdd(%arga: tensor<32x16x8xf32, #Tsdd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsdd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<32x16x8xf32>
   return %0 : tensor<32x16x8xf32>
 }
 
 // CHECK-LABEL:   func @mul_sdd(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_15:.*]] = %[[VAL_13]] to %[[VAL_14]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_17:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_18:.*]] = arith.muli %[[VAL_15]], %[[VAL_3]] : index
 // CHECK:               %[[VAL_19:.*]] = arith.addi %[[VAL_18]], %[[VAL_17]] : index
 // CHECK:               scf.for %[[VAL_20:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_21:.*]] = arith.muli %[[VAL_19]], %[[VAL_4]] : index
 // CHECK:                 %[[VAL_22:.*]] = arith.addi %[[VAL_21]], %[[VAL_20]] : index
 // CHECK:                 %[[VAL_23:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_22]]] : memref<?xf32>
 // CHECK:                 %[[VAL_24:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_16]], %[[VAL_17]], %[[VAL_20]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_25:.*]] = arith.mulf %[[VAL_23]], %[[VAL_24]] : f32
 // CHECK:                 memref.store %[[VAL_25]], %[[VAL_12]]{{\[}}%[[VAL_16]], %[[VAL_17]], %[[VAL_20]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_26:.*]] = memref.tensor_load %[[VAL_12]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_26:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_26]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_sdd(%arga: tensor<32x16x8xf32, #Tsdd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsdd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_9:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_16:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_16:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_16]], %[[VAL_17]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_8]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_9]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]]:2 = scf.while (%[[VAL_21:.*]] = %[[VAL_18]], %[[VAL_22:.*]] = %[[VAL_8]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_23:.*]] = arith.cmpi ult, %[[VAL_21]], %[[VAL_19]] : index
 // CHECK:             scf.condition(%[[VAL_23]]) %[[VAL_21]], %[[VAL_22]] : index, index
 // CHECK:           } do {

 // CHECK:           scf.for %[[VAL_60:.*]] = %[[VAL_61:.*]]#1 to %[[VAL_4]] step %[[VAL_9]] {
 // CHECK:             scf.for %[[VAL_62:.*]] = %[[VAL_8]] to %[[VAL_5]] step %[[VAL_9]] {
 // CHECK:               scf.for %[[VAL_63:.*]] = %[[VAL_8]] to %[[VAL_6]] step %[[VAL_9]] {
 // CHECK:                 %[[VAL_64:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_60]], %[[VAL_62]], %[[VAL_63]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_64]], %[[VAL_17]]{{\[}}%[[VAL_60]], %[[VAL_62]], %[[VAL_63]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_65:.*]] = memref.tensor_load %[[VAL_17]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_65:.*]] = bufferization.to_tensor %[[VAL_17]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_65]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_sds(%arga: tensor<32x16x8xf32, #Tsds>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsds>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 16 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_17:.*]] = %[[VAL_15]] to %[[VAL_16]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_19:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_20:.*]] = arith.muli %[[VAL_17]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_21:.*]] = arith.addi %[[VAL_20]], %[[VAL_19]] : index
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:               %[[VAL_23:.*]] = arith.addi %[[VAL_21]], %[[VAL_6]] : index
 // CHECK:               %[[VAL_24:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_23]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_25:.*]] = %[[VAL_22]] to %[[VAL_24]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_25]]] : memref<?xindex>
 // CHECK:                 %[[VAL_27:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_25]]] : memref<?xf32>
 // CHECK:                 %[[VAL_28:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_18]], %[[VAL_19]], %[[VAL_26]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_29:.*]] = arith.mulf %[[VAL_27]], %[[VAL_28]] : f32
 // CHECK:                 memref.store %[[VAL_29]], %[[VAL_14]]{{\[}}%[[VAL_18]], %[[VAL_19]], %[[VAL_26]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_30:.*]] = memref.tensor_load %[[VAL_14]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_30]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_sds(%arga: tensor<32x16x8xf32, #Tsds>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsds>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_15]], %[[VAL_16]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_8]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]]:2 = scf.while (%[[VAL_20:.*]] = %[[VAL_17]], %[[VAL_21:.*]] = %[[VAL_7]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_22:.*]] = arith.cmpi ult, %[[VAL_20]], %[[VAL_18]] : index
 // CHECK:             scf.condition(%[[VAL_22]]) %[[VAL_20]], %[[VAL_21]] : index, index
 // CHECK:           } do {

 // CHECK:           scf.for %[[VAL_61:.*]] = %[[VAL_62:.*]]#1 to %[[VAL_3]] step %[[VAL_8]] {
 // CHECK:             scf.for %[[VAL_63:.*]] = %[[VAL_7]] to %[[VAL_4]] step %[[VAL_8]] {
 // CHECK:               scf.for %[[VAL_64:.*]] = %[[VAL_7]] to %[[VAL_5]] step %[[VAL_8]] {
 // CHECK:                 %[[VAL_65:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_61]], %[[VAL_63]], %[[VAL_64]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_65]], %[[VAL_16]]{{\[}}%[[VAL_61]], %[[VAL_63]], %[[VAL_64]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_66:.*]] = memref.tensor_load %[[VAL_16]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_66:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_66]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_ssd(%arga: tensor<32x16x8xf32, #Tssd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tssd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 8 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_16:.*]] = %[[VAL_14]] to %[[VAL_15]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]] = arith.addi %[[VAL_16]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_21:.*]] = %[[VAL_18]] to %[[VAL_20]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_23:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK:                 %[[VAL_24:.*]] = arith.muli %[[VAL_21]], %[[VAL_3]] : index
 // CHECK:                 %[[VAL_25:.*]] = arith.addi %[[VAL_24]], %[[VAL_23]] : index
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_25]]] : memref<?xf32>
 // CHECK:                 %[[VAL_27:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_17]], %[[VAL_22]], %[[VAL_23]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_28:.*]] = arith.mulf %[[VAL_26]], %[[VAL_27]] : f32
 // CHECK:                 memref.store %[[VAL_28]], %[[VAL_13]]{{\[}}%[[VAL_17]], %[[VAL_22]], %[[VAL_23]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_29:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_29:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_29]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_ssd(%arga: tensor<32x16x8xf32, #Tssd>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tssd>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-DAG:           %[[VAL_9:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_17:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_18:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_19:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_18]], %[[VAL_19]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_8]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_9]]] : memref<?xindex>
 // CHECK:           %[[VAL_22:.*]]:2 = scf.while (%[[VAL_23:.*]] = %[[VAL_20]], %[[VAL_24:.*]] = %[[VAL_8]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_25:.*]] = arith.cmpi ult, %[[VAL_23]], %[[VAL_21]] : index
 // CHECK:             scf.condition(%[[VAL_25]]) %[[VAL_23]], %[[VAL_24]] : index, index
 // CHECK:           } do {

 // CHECK:           scf.for %[[VAL_80:.*]] = %[[VAL_81:.*]]#1 to %[[VAL_4]] step %[[VAL_9]] {
 // CHECK:             scf.for %[[VAL_82:.*]] = %[[VAL_8]] to %[[VAL_5]] step %[[VAL_9]] {
 // CHECK:               scf.for %[[VAL_83:.*]] = %[[VAL_8]] to %[[VAL_6]] step %[[VAL_9]] {
 // CHECK:                 %[[VAL_84:.*]] = memref.load %[[VAL_17]]{{\[}}%[[VAL_80]], %[[VAL_82]], %[[VAL_83]]] : memref<32x16x8xf32>
 // CHECK:                 memref.store %[[VAL_84]], %[[VAL_19]]{{\[}}%[[VAL_80]], %[[VAL_82]], %[[VAL_83]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_85:.*]] = memref.tensor_load %[[VAL_19]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_85:.*]] = bufferization.to_tensor %[[VAL_19]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_85]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @add_sss(%arga: tensor<32x16x8xf32, #Tsss>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsss>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.addf %a, %b : f32

 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed", "compressed", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32x16x8xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16x8xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           memref.copy %[[VAL_14]], %[[VAL_15]] : memref<32x16x8xf32> to memref<32x16x8xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_18:.*]] = %[[VAL_16]] to %[[VAL_17]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_19:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_21:.*]] = arith.addi %[[VAL_18]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_22:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_23:.*]] = %[[VAL_20]] to %[[VAL_22]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_24:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_23]]] : memref<?xindex>
 // CHECK:               %[[VAL_25:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_23]]] : memref<?xindex>
 // CHECK:               %[[VAL_26:.*]] = arith.addi %[[VAL_23]], %[[VAL_5]] : index
 // CHECK:               %[[VAL_27:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_26]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_28:.*]] = %[[VAL_25]] to %[[VAL_27]] step %[[VAL_5]] {
 // CHECK:                 %[[VAL_29:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_28]]] : memref<?xindex>
 // CHECK:                 %[[VAL_30:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_28]]] : memref<?xf32>
 // CHECK:                 %[[VAL_31:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_19]], %[[VAL_24]], %[[VAL_29]]] : memref<32x16x8xf32>
 // CHECK:                 %[[VAL_32:.*]] = arith.mulf %[[VAL_30]], %[[VAL_31]] : f32
 // CHECK:                 memref.store %[[VAL_32]], %[[VAL_15]]{{\[}}%[[VAL_19]], %[[VAL_24]], %[[VAL_29]]] : memref<32x16x8xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_15]] : memref<32x16x8xf32>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<32x16x8xf32>
 // CHECK:           return %[[VAL_33]] : tensor<32x16x8xf32>
 // CHECK:         }
 func @mul_sss(%arga: tensor<32x16x8xf32, #Tsss>, %argb: tensor<32x16x8xf32>, %argx: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16x8xf32, #Tsss>, tensor<32x16x8xf32>)
     outs(%argx: tensor<32x16x8xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32

 // CHECK-SAME:                    %[[VAL_3:.*3]]: tensor<?x?xf32>) -> tensor<?x?xf32> {
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 2 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
 // CHECK:           %[[VAL_10:.*]] = tensor.dim %[[VAL_2]], %[[VAL_5]] : tensor<?x?xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<?x?xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<?x?xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<?x?xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<?x?xf32>
 // CHECK:           %[[VAL_13:.*]] = tensor.dim %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_14:.*]] = tensor.dim %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_0]] : memref<?x?xf32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_0]] : memref<?x?xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc(%[[VAL_13]], %[[VAL_14]]) : memref<?x?xf32>
 // CHECK:           memref.copy %[[VAL_15]], %[[VAL_16]] : memref<?x?xf32> to memref<?x?xf32>
 // CHECK:           scf.for %[[VAL_17:.*]] = %[[VAL_5]] to %[[VAL_13]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_18:.*]] = %[[VAL_5]] to %[[VAL_10]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_19:.*]] = arith.muli %[[VAL_10]], %[[VAL_17]] : index
 // CHECK:               %[[VAL_20:.*]] = arith.addi %[[VAL_19]], %[[VAL_18]] : index
 // CHECK:               %[[VAL_21:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:               %[[VAL_22:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_22]]] : memref<?xindex>
 // CHECK:               scf.for %[[VAL_24:.*]] = %[[VAL_21]] to %[[VAL_23]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_25:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xindex>
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK:                 scf.for %[[VAL_27:.*]] = %[[VAL_5]] to %[[VAL_14]] step %[[VAL_6]] {
 // CHECK:                   %[[VAL_28:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_18]], %[[VAL_27]]] : memref<?x?xf32>
 // CHECK:                   %[[VAL_29:.*]] = arith.mulf %[[VAL_26]], %[[VAL_28]] : f32
 // CHECK:                   %[[VAL_30:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_25]], %[[VAL_27]]] : memref<?x?xf32>
 // CHECK:                   %[[VAL_31:.*]] = arith.mulf %[[VAL_29]], %[[VAL_30]] : f32
 // CHECK:                   %[[VAL_32:.*]] = memref.load %[[VAL_16]]{{\[}}%[[VAL_17]], %[[VAL_27]]] : memref<?x?xf32>
 // CHECK:                   %[[VAL_33:.*]] = arith.addf %[[VAL_31]], %[[VAL_32]] : f32
 // CHECK:                   memref.store %[[VAL_33]], %[[VAL_16]]{{\[}}%[[VAL_17]], %[[VAL_27]]] : memref<?x?xf32>
 // CHECK:                 }
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_34:.*]] = memref.tensor_load %[[VAL_16]] : memref<?x?xf32>
+// CHECK:           %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<?x?xf32>
 // CHECK:           return %[[VAL_34]] : tensor<?x?xf32>
 // CHECK:         }
 func @kernel_3d(%arga: tensor<?x?xf32>,
                 %argb: tensor<?x?x?xf32, #Tdds>,
                 %argc: tensor<?x?xf32>,
           %argd: tensor<?x?xf32>) -> tensor<?x?xf32> {
   %0 = linalg.generic #trait_kernel_3d
        ins(%argb, %argc, %argd: tensor<?x?x?xf32, #Tdds>, tensor<?x?xf32>, tensor<?x?xf32>)

 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 2 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<f32> to memref<f32>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_10]][] : memref<f32>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = scf.for %[[VAL_15:.*]] = %[[VAL_12]] to %[[VAL_13]] step %[[VAL_3]] iter_args(%[[VAL_16:.*]] = %[[VAL_11]]) -> (f32) {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = arith.addi %[[VAL_15]], %[[VAL_3]] : index
 // CHECK:             %[[VAL_19:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_20:.*]] = scf.for %[[VAL_21:.*]] = %[[VAL_17]] to %[[VAL_19]] step %[[VAL_3]] iter_args(%[[VAL_22:.*]] = %[[VAL_16]]) -> (f32) {
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:               %[[VAL_24:.*]] = arith.addi %[[VAL_21]], %[[VAL_3]] : index
 // CHECK:               %[[VAL_25:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_24]]] : memref<?xindex>
 // CHECK:               %[[VAL_26:.*]] = scf.for %[[VAL_27:.*]] = %[[VAL_23]] to %[[VAL_25]] step %[[VAL_3]] iter_args(%[[VAL_28:.*]] = %[[VAL_22]]) -> (f32) {
 // CHECK:                 %[[VAL_29:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_27]]] : memref<?xf32>
 // CHECK:                 %[[VAL_30:.*]] = arith.addf %[[VAL_28]], %[[VAL_29]] : f32
 // CHECK:                 scf.yield %[[VAL_30]] : f32
 // CHECK:               }
 // CHECK:               scf.yield %[[VAL_26]] : f32
 // CHECK:             }
 // CHECK:             scf.yield %[[VAL_20]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_14]], %[[VAL_10]][] : memref<f32>
-// CHECK:           %[[VAL_34:.*]] = memref.tensor_load %[[VAL_10]] : memref<f32>
+// CHECK:           %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<f32>
 // CHECK:           return %[[VAL_34]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction(%arga: tensor<10x20x30xf32, #Tsss>, %argx: tensor<f32>) -> tensor<f32> {
   %0 = linalg.generic #trait_sum_reduction
      ins(%arga: tensor<10x20x30xf32, #Tsss>)
     outs(%argx: tensor<f32>) {
       ^bb(%a: f32, %x: f32):
         %0 = arith.addf %x, %a : f32

 // CHECK-SAME:      %[[VAL_0:.*]]: tensor<?x?x?xf32>,
 // CHECK-SAME:      %[[VAL_1:.*]]: tensor<?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-SAME:      %[[VAL_2:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 2 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = tensor.dim %[[VAL_0]], %[[VAL_3]] : tensor<?x?x?xf32>
 // CHECK:           %[[VAL_7:.*]] = tensor.dim %[[VAL_0]], %[[VAL_4]] : tensor<?x?x?xf32>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_0]] : memref<?x?x?xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_0]] : memref<?x?x?xf32>
 // CHECK:           %[[VAL_9:.*]] = tensor.dim %[[VAL_1]], %[[VAL_5]] : tensor<?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<f32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<f32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<f32>
 // CHECK:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<f32> to memref<f32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_12]][] : memref<f32>
 // CHECK:           %[[VAL_14:.*]] = scf.for %[[VAL_15:.*]] = %[[VAL_5]] to %[[VAL_9]] step %[[VAL_3]] iter_args(%[[VAL_16:.*]] = %[[VAL_13]]) -> (f32) {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_15]]] : memref<?xf32>
 // CHECK:             %[[VAL_18:.*]] = scf.for %[[VAL_19:.*]] = %[[VAL_5]] to %[[VAL_6]] step %[[VAL_3]] iter_args(%[[VAL_20:.*]] = %[[VAL_16]]) -> (f32) {
 // CHECK:               %[[VAL_21:.*]] = scf.for %[[VAL_22:.*]] = %[[VAL_5]] to %[[VAL_7]] step %[[VAL_3]] iter_args(%[[VAL_23:.*]] = %[[VAL_20]]) -> (f32) {
 // CHECK:                 %[[VAL_24:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_15]], %[[VAL_19]], %[[VAL_22]]] : memref<?x?x?xf32>
 // CHECK:                 %[[VAL_25:.*]] = arith.mulf %[[VAL_24]], %[[VAL_17]] : f32
 // CHECK:                 %[[VAL_26:.*]] = arith.addf %[[VAL_23]], %[[VAL_25]] : f32
 // CHECK:                 scf.yield %[[VAL_26]] : f32
 // CHECK:               }
 // CHECK:               scf.yield %[[VAL_21]] : f32
 // CHECK:             }
 // CHECK:             scf.yield %[[VAL_18]] : f32
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_14]], %[[VAL_12]][] : memref<f32>
-// CHECK:           %[[VAL_30:.*]] = memref.tensor_load %[[VAL_12]] : memref<f32>
+// CHECK:           %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<f32>
 // CHECK:           return %[[VAL_30]] : tensor<f32>
 // CHECK:         }
 func @sum_reduction_inv(%arga: tensor<?x?x?xf32>,
                         %argb: tensor<?xf32, #Td>,
             %argx: tensor<f32>) -> tensor<f32> {
   %0 = linalg.generic #trait_sum_reduction_inv
     ins(%arga, %argb: tensor<?x?x?xf32>, tensor<?xf32, #Td>)
     outs(%argx: tensor<f32>) {

 // CHECK-SAME:                     %[[VAL_2:.*]]: tensor<30xf32>,
 // CHECK-SAME:                     %[[VAL_3:.*]]: tensor<10x20x30xf32>) -> tensor<10x20x30xf32> {
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 10 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant 20 : index
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 30 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_8:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<20xf32>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<30xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<10x20x30xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<20xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<30xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<10x20x30xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<10x20x30xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<10x20x30xf32> to memref<10x20x30xf32>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_7]] to %[[VAL_4]] step %[[VAL_8]] {
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_14]]] : memref<?xf32>
 // CHECK:             scf.for %[[VAL_16:.*]] = %[[VAL_7]] to %[[VAL_5]] step %[[VAL_8]] {
 // CHECK:               %[[VAL_17:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_16]]] : memref<20xf32>
 // CHECK:               scf.for %[[VAL_18:.*]] = %[[VAL_7]] to %[[VAL_6]] step %[[VAL_8]] {
 // CHECK:                 %[[VAL_19:.*]] = arith.mulf %[[VAL_15]], %[[VAL_17]] : f32
 // CHECK:                 %[[VAL_20:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_18]]] : memref<30xf32>
 // CHECK:                 %[[VAL_21:.*]] = arith.mulf %[[VAL_19]], %[[VAL_20]] : f32
 // CHECK:                 memref.store %[[VAL_21]], %[[VAL_13]]{{\[}}%[[VAL_14]], %[[VAL_16]], %[[VAL_18]]] : memref<10x20x30xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_22:.*]] = memref.tensor_load %[[VAL_13]] : memref<10x20x30xf32>
+// CHECK:           %[[VAL_22:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<10x20x30xf32>
 // CHECK:           return %[[VAL_22]] : tensor<10x20x30xf32>
 // CHECK:         }
 func @invariants(%arga: tensor<10xf32, #Td>,
                  %argb: tensor<20xf32>,
                  %argc: tensor<30xf32>,
                  %argx: tensor<10x20x30xf32>) -> tensor<10x20x30xf32> {
   %0 = linalg.generic #trait_invariants
      ins(%arga, %argb, %argc : tensor<10xf32, #Td>, tensor<20xf32>, tensor<30xf32>)
     outs(%argx: tensor<10x20x30xf32>) {
       ^bb(%a: f32, %b: f32, %c: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32
         %1 = arith.mulf %0, %c : f32
         linalg.yield %1 : f32
   } -> tensor<10x20x30xf32>
   return %0 : tensor<10x20x30xf32>
 }

mlir/test/Dialect/SparseTensor/sparse_affine.mlir

 // CHECK-SAME:                          %[[VAL_1:.*]]: tensor<4xf32>,
 // CHECK-SAME:                          %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 3 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<4xf32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<4xf32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf32> to memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<4xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_15:.*]] = %[[VAL_13]] to %[[VAL_14]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_16]]] : memref<32xf32>
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_15]]] : memref<?xf32>
 // CHECK:             %[[VAL_19:.*]] = arith.mulf %[[VAL_18]], %[[VAL_12]] : f32
 // CHECK:             %[[VAL_20:.*]] = arith.addf %[[VAL_17]], %[[VAL_19]] : f32
 // CHECK:             memref.store %[[VAL_20]], %[[VAL_11]]{{\[}}%[[VAL_16]]] : memref<32xf32>
 // CHECK:           }
-// CHECK:           %[[VAL_21:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf32>
+// CHECK:           %[[VAL_21:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf32>
 // CHECK:           return %[[VAL_21]] : tensor<32xf32>
 // CHECK:         }
 func @mul_inv_dense1d(%arga: tensor<32xf32, #SpVec>,
                       %argb: tensor<4xf32>,
                       %argx: tensor<32xf32>) -> tensor<32xf32> {
   %0 = linalg.generic #trait1
      ins(%arga, %argb: tensor<32xf32, #SpVec>, tensor<4xf32>)
     outs(%argx: tensor<32xf32>) {

 // CHECK-SAME:                             %[[VAL_1:.*]]: tensor<34xi32>,
 // CHECK-SAME:                             %[[VAL_2:.*]]: tensor<32xi32>) -> tensor<32xi32> {
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 2 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<34xi32>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<34xi32>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32xi32>
 // CHECK:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xi32> to memref<32xi32>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_12]] to %[[VAL_13]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_14]]] : memref<?xi32>
 // CHECK:             %[[VAL_17:.*]] = arith.addi %[[VAL_15]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_17]]] : memref<34xi32>
 // CHECK:             %[[VAL_19:.*]] = arith.andi %[[VAL_16]], %[[VAL_18]] : i32
 // CHECK:             memref.store %[[VAL_19]], %[[VAL_11]]{{\[}}%[[VAL_15]]] : memref<32xi32>
 // CHECK:           }
-// CHECK:           %[[VAL_20:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xi32>
+// CHECK:           %[[VAL_20:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xi32>
 // CHECK:           return %[[VAL_20]] : tensor<32xi32>
 // CHECK:         }
 func @and_affine_dense1d(%arga: tensor<32xi32, #SpVec>,
                          %argb: tensor<34xi32>,
                          %argx: tensor<32xi32>) -> tensor<32xi32> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi32, #SpVec>, tensor<34xi32>)
     outs(%argx: tensor<32xi32>) {

 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 32 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_6:.*]] = arith.constant 2 : index
 // CHECK-DAG:       %[[VAL_7:.*]] = arith.constant 3 : index
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<34x19xf64>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<34x19xf64>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32x16xf64>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<32x16xf64>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<32x16xf64> to memref<32x16xf64>
 // CHECK:           scf.for %[[VAL_14:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK:             %[[VAL_16:.*]] = arith.addi %[[VAL_14]], %[[VAL_3]] : index
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_18:.*]] = %[[VAL_15]] to %[[VAL_17]] step %[[VAL_3]] {
 // CHECK:               %[[VAL_19:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:               %[[VAL_20:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_14]], %[[VAL_19]]] : memref<32x16xf64>
 // CHECK:               %[[VAL_21:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_18]]] : memref<?xf64>
 // CHECK:               %[[VAL_22:.*]] = arith.addi %[[VAL_14]], %[[VAL_6]] : index
 // CHECK:               %[[VAL_23:.*]] = arith.addi %[[VAL_19]], %[[VAL_7]] : index
 // CHECK:               %[[VAL_24:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_22]], %[[VAL_23]]] : memref<34x19xf64>
 // CHECK:               %[[VAL_25:.*]] = arith.mulf %[[VAL_21]], %[[VAL_24]] : f64
 // CHECK:               %[[VAL_26:.*]] = arith.addf %[[VAL_20]], %[[VAL_25]] : f64
 // CHECK:               memref.store %[[VAL_26]], %[[VAL_13]]{{\[}}%[[VAL_14]], %[[VAL_19]]] : memref<32x16xf64>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_27:.*]] = memref.tensor_load %[[VAL_13]] : memref<32x16xf64>
+// CHECK:           %[[VAL_27:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<32x16xf64>
 // CHECK:           return %[[VAL_27]] : tensor<32x16xf64>
 // CHECK:         }
 func @mul_affine_dense2d(%arga: tensor<32x16xf64, #CSR>,
                          %argb: tensor<34x19xf64>,
                          %argx: tensor<32x16xf64>) -> tensor<32x16xf64> {
   %0 = linalg.generic #trait3
      ins(%arga, %argb: tensor<32x16xf64, #CSR>, tensor<34x19xf64>)
     outs(%argx: tensor<32x16xf64>) {
       ^bb(%a: f64, %b: f64, %x: f64):
         %0 = arith.mulf %a, %b : f64
         %1 = arith.addf %x, %0 : f64
         linalg.yield %1 : f64
   } -> tensor<32x16xf64>
   return %0 : tensor<32x16xf64>
 }

mlir/test/Dialect/SparseTensor/sparse_fp_ops.mlir

 // CHECK-LABEL:   func @abs(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xf64>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
 // CHECK:           %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_10]]] : memref<?xindex>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf64>
 // CHECK:             %[[VAL_13:.*]] = math.abs %[[VAL_12]] : f64
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_7]] : memref<32xf64>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf64>
 // CHECK:           return %[[VAL_14]] : tensor<32xf64>
 func @abs(%arga: tensor<32xf64, #SV>,
           %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf64, #SV>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %x: f64):
         %0 = math.abs %a : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 
 // CHECK-LABEL:   func @ceil(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xf64>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
 // CHECK:           %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_10]]] : memref<?xindex>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf64>
 // CHECK:             %[[VAL_13:.*]] = math.ceil %[[VAL_12]] : f64
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_7]] : memref<32xf64>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf64>
 // CHECK:           return %[[VAL_14]] : tensor<32xf64>
 // CHECK:         }
 func @ceil(%arga: tensor<32xf64, #SV>,
            %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf64, #SV>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %x: f64):
         %0 = math.ceil %a : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 
 // CHECK-LABEL:   func @floor(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xf64>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
 // CHECK:           %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_10]]] : memref<?xindex>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf64>
 // CHECK:             %[[VAL_13:.*]] = math.floor %[[VAL_12]] : f64
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_7]] : memref<32xf64>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf64>
 // CHECK:           return %[[VAL_14]] : tensor<32xf64>
 // CHECK:         }
 func @floor(%arga: tensor<32xf64, #SV>,
             %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf64, #SV>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %x: f64):
         %0 = math.floor %a : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 
 // CHECK-LABEL:   func @neg(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xf64>
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
 // CHECK:           %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] {
 // CHECK:             %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_10]]] : memref<?xindex>
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf64>
 // CHECK:             %[[VAL_13:.*]] = arith.negf %[[VAL_12]] : f64
 // CHECK:             memref.store %[[VAL_13]], %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_14:.*]] = memref.tensor_load %[[VAL_7]] : memref<32xf64>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf64>
 // CHECK:           return %[[VAL_14]] : tensor<32xf64>
 // CHECK:         }
 func @neg(%arga: tensor<32xf64, #SV>,
           %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait1
      ins(%arga: tensor<32xf64, #SV>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %x: f64):

 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>

 // CHECK:             %[[VAL_28:.*]] = select %[[VAL_26]], %[[VAL_27]], %[[VAL_18]] : index
 // CHECK:             %[[VAL_29:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_28]], %[[VAL_29]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_30:.*]] = %[[VAL_31:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_32:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_30]]] : memref<32xf64>
 // CHECK:             memref.store %[[VAL_32]], %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf64>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf64>
 // CHECK:           return %[[VAL_33]] : tensor<32xf64>
 // CHECK:         }
 func @add(%arga: tensor<32xf64, #SV>,
           %argb: tensor<32xf64>,
           %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf64, #SV>, tensor<32xf64>)
     outs(%argx: tensor<32xf64>) {

 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xf64>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>

 // CHECK:             %[[VAL_30:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_29]], %[[VAL_30]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_31:.*]] = %[[VAL_32:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_31]]] : memref<32xf64>
 // CHECK:             %[[VAL_34:.*]] = arith.negf %[[VAL_33]] : f64
 // CHECK:             memref.store %[[VAL_34]], %[[VAL_11]]{{\[}}%[[VAL_31]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_35:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf64>
+// CHECK:           %[[VAL_35:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf64>
 // CHECK:           return %[[VAL_35]] : tensor<32xf64>
 // CHECK:         }
 func @sub(%arga: tensor<32xf64, #SV>,
           %argb: tensor<32xf64>,
           %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf64, #SV>, tensor<32xf64>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %b: f64, %x: f64):
         %0 = arith.subf %a, %b : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 
 // CHECK-LABEL:   func @mul(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xf64>,
 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_10]] to %[[VAL_11]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xf64>
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_13]]] : memref<32xf64>
 // CHECK:             %[[VAL_16:.*]] = arith.mulf %[[VAL_14]], %[[VAL_15]] : f64
 // CHECK:             memref.store %[[VAL_16]], %[[VAL_9]]{{\[}}%[[VAL_13]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_17:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xf64>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf64>
 // CHECK:           return %[[VAL_17]] : tensor<32xf64>
 // CHECK:         }
 func @mul(%arga: tensor<32xf64, #SV>,
           %argb: tensor<32xf64>,
           %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xf64, #SV>, tensor<32xf64>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %b: f64, %x: f64):
         %0 = arith.mulf %a, %b : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 
 // CHECK-LABEL:   func @divbyc(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2.000000e+00 : f64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xf64>
 // CHECK:             %[[VAL_14:.*]] = arith.divf %[[VAL_13]], %[[VAL_2]] : f64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xf64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xf64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf64>
 // CHECK:           return %[[VAL_15]] : tensor<32xf64>
 // CHECK:         }
 func @divbyc(%arga: tensor<32xf64, #SV>,
            %argx: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
   %c = arith.constant 2.0 : f64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xf64, #SV>)
     outs(%argx: tensor<32xf64>) {
       ^bb(%a: f64, %x: f64):
         %0 = arith.divf %a, %c : f64
         linalg.yield %0 : f64
   } -> tensor<32xf64>
   return %0 : tensor<32xf64>
 }
 

mlir/test/Dialect/SparseTensor/sparse_int_ops.mlir

 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>

 // CHECK:             %[[VAL_28:.*]] = select %[[VAL_26]], %[[VAL_27]], %[[VAL_18]] : index
 // CHECK:             %[[VAL_29:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_28]], %[[VAL_29]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_30:.*]] = %[[VAL_31:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_32:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:             memref.store %[[VAL_32]], %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xi64>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xi64>
 // CHECK:           return %[[VAL_33]] : tensor<32xi64>
 // CHECK:         }
 func @add(%arga: tensor<32xi64, #SV>,
           %argb: tensor<32xi64>,
           %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {

 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK-DAG:           %[[VAL_7:.*]] = arith.constant 0 : i64
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]]:2 = scf.while (%[[VAL_16:.*]] = %[[VAL_13]], %[[VAL_17:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_18:.*]] = arith.cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
 // CHECK:             scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_19:.*]]: index, %[[VAL_20:.*]]: index):
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<?xindex>

 // CHECK:             %[[VAL_31:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_30]], %[[VAL_31]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_32:.*]] = %[[VAL_33:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_34:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_32]]] : memref<32xi64>
 // CHECK:             %[[VAL_35:.*]] = arith.subi %[[VAL_7]], %[[VAL_34]] : i64
 // CHECK:             memref.store %[[VAL_35]], %[[VAL_12]]{{\[}}%[[VAL_32]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_36:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xi64>
+// CHECK:           %[[VAL_36:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xi64>
 // CHECK:           return %[[VAL_36]] : tensor<32xi64>
 // CHECK:         }
 func @sub(%arga: tensor<32xi64, #SV>,
           %argb: tensor<32xi64>,
           %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %b: i64, %x: i64):
         %0 = arith.subi %a, %b : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @mul(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xi64>,
 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_10]] to %[[VAL_11]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xi64>
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_13]]] : memref<32xi64>
 // CHECK:             %[[VAL_16:.*]] = arith.muli %[[VAL_14]], %[[VAL_15]] : i64
 // CHECK:             memref.store %[[VAL_16]], %[[VAL_9]]{{\[}}%[[VAL_13]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_17:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xi64>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xi64>
 // CHECK:           return %[[VAL_17]] : tensor<32xi64>
 // CHECK:         }
 func @mul(%arga: tensor<32xi64, #SV>,
           %argb: tensor<32xi64>,
           %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %b: i64, %x: i64):
         %0 = arith.muli %a, %b : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @divsbyc(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2 : i64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xi64>
 // CHECK:             %[[VAL_14:.*]] = arith.divsi %[[VAL_13]], %[[VAL_2]] : i64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xi64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xi64>
 // CHECK:           return %[[VAL_15]] : tensor<32xi64>
 // CHECK:         }
 func @divsbyc(%arga: tensor<32xi64, #SV>,
               %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %c = arith.constant 2 : i64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xi64, #SV>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %x: i64):
         %0 = arith.divsi %a, %c : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @divubyc(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2 : i64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xi64>
 // CHECK:             %[[VAL_14:.*]] = arith.divui %[[VAL_13]], %[[VAL_2]] : i64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xi64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xi64>
 // CHECK:           return %[[VAL_15]] : tensor<32xi64>
 // CHECK:         }
 func @divubyc(%arga: tensor<32xi64, #SV>,
               %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %c = arith.constant 2 : i64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xi64, #SV>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %x: i64):
         %0 = arith.divui %a, %c : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @and(
 // CHECK-SAME:              %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:              %[[VAL_1:.*]]: tensor<32xi64>,
 // CHECK-SAME:              %[[VAL_2:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_12:.*]] = %[[VAL_10]] to %[[VAL_11]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xi64>
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_13]]] : memref<32xi64>
 // CHECK:             %[[VAL_16:.*]] = arith.andi %[[VAL_14]], %[[VAL_15]] : i64
 // CHECK:             memref.store %[[VAL_16]], %[[VAL_9]]{{\[}}%[[VAL_13]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_17:.*]] = memref.tensor_load %[[VAL_9]] : memref<32xi64>
+// CHECK:           %[[VAL_17:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xi64>
 // CHECK:           return %[[VAL_17]] : tensor<32xi64>
 // CHECK:         }
 func @and(%arga: tensor<32xi64, #SV>,
           %argb: tensor<32xi64>,
           %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {

 // CHECK-SAME:             %[[VAL_2:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>

 // CHECK:             %[[VAL_28:.*]] = select %[[VAL_26]], %[[VAL_27]], %[[VAL_18]] : index
 // CHECK:             %[[VAL_29:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_28]], %[[VAL_29]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_30:.*]] = %[[VAL_31:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_32:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:             memref.store %[[VAL_32]], %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xi64>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xi64>
 // CHECK:           return %[[VAL_33]] : tensor<32xi64>
 // CHECK:         }
 func @or(%arga: tensor<32xi64, #SV>,
          %argb: tensor<32xi64>,
          %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {

 // CHECK-SAME:             %[[VAL_2:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_5:.*]] = arith.constant true
 // CHECK-DAG:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xi64>
+// CHECK:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xi64>
 // CHECK:           %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
 // CHECK:             %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
 // CHECK:           } do {
 // CHECK:           ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>

 // CHECK:             %[[VAL_28:.*]] = select %[[VAL_26]], %[[VAL_27]], %[[VAL_18]] : index
 // CHECK:             %[[VAL_29:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
 // CHECK:             scf.yield %[[VAL_28]], %[[VAL_29]] : index, index
 // CHECK:           }
 // CHECK:           scf.for %[[VAL_30:.*]] = %[[VAL_31:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             %[[VAL_32:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:             memref.store %[[VAL_32]], %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xi64>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xi64>
 // CHECK:           return %[[VAL_33]] : tensor<32xi64>
 // CHECK:         }
 func @xor(%arga: tensor<32xi64, #SV>,
           %argb: tensor<32xi64>,
           %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %0 = linalg.generic #trait2
      ins(%arga, %argb: tensor<32xi64, #SV>, tensor<32xi64>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %b: i64, %x: i64):
         %0 = arith.xori %a, %b : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @ashrbyc(
 // CHECK-SAME:                  %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                  %[[VAL_1:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2 : i64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xi64>
 // CHECK:             %[[VAL_14:.*]] = arith.shrsi %[[VAL_13]], %[[VAL_2]] : i64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xi64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xi64>
 // CHECK:           return %[[VAL_15]] : tensor<32xi64>
 // CHECK:         }
 func @ashrbyc(%arga: tensor<32xi64, #SV>,
               %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %c = arith.constant 2 : i64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xi64, #SV>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %x: i64):
         %0 = arith.shrsi %a, %c : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @lsrbyc(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2 : i64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xi64>
 // CHECK:             %[[VAL_14:.*]] = arith.shrui %[[VAL_13]], %[[VAL_2]] : i64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xi64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xi64>
 // CHECK:           return %[[VAL_15]] : tensor<32xi64>
 // CHECK:         }
 func @lsrbyc(%arga: tensor<32xi64, #SV>,
              %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %c = arith.constant 2 : i64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xi64, #SV>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %x: i64):
         %0 = arith.shrui %a, %c : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 
 // CHECK-LABEL:   func @lslbyc(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
 // CHECK-DAG:           %[[VAL_2:.*]] = arith.constant 2 : i64
 // CHECK-DAG:           %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xindex>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xi64, #sparse_tensor.encoding<{{{.*}}}>> to memref<?xi64>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xi64>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xi64>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_11]]] : memref<?xi64>
 // CHECK:             %[[VAL_14:.*]] = arith.shli %[[VAL_13]], %[[VAL_2]] : i64
 // CHECK:             memref.store %[[VAL_14]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xi64>
 // CHECK:           }
-// CHECK:           %[[VAL_15:.*]] = memref.tensor_load %[[VAL_8]] : memref<32xi64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xi64>
 // CHECK:           return %[[VAL_15]] : tensor<32xi64>
 // CHECK:         }
 func @lslbyc(%arga: tensor<32xi64, #SV>,
              %argx: tensor<32xi64> {linalg.inplaceable = true}) -> tensor<32xi64> {
   %c = arith.constant 2 : i64
   %0 = linalg.generic #traitc
      ins(%arga: tensor<32xi64, #SV>)
     outs(%argx: tensor<32xi64>) {
       ^bb(%a: i64, %x: i64):
         %0 = arith.shli %a, %c : i64
         linalg.yield %0 : i64
   } -> tensor<32xi64>
   return %0 : tensor<32xi64>
 }
 

mlir/test/Dialect/SparseTensor/sparse_kernels.mlir

 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 30 : index
 // CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<10x20xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<10x20xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<10x20xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<20x30xf32>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<10x30xf32>
+// CHECK:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_1]] : memref<20x30xf32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<10x30xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<10x30xf32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<10x30xf32> to memref<10x30xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_16:.*]] = %[[VAL_14]] to %[[VAL_15]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]] = arith.addi %[[VAL_16]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_21:.*]] = %[[VAL_18]] to %[[VAL_20]] step %[[VAL_4]] {
 // CHECK:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_21]]] : memref<?xf32>
 // CHECK:               scf.for %[[VAL_24:.*]] = %[[VAL_3]] to %[[VAL_5]] step %[[VAL_4]] {
 // CHECK:                 %[[VAL_25:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_17]], %[[VAL_24]]] : memref<10x30xf32>
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_22]], %[[VAL_24]]] : memref<20x30xf32>
 // CHECK:                 %[[VAL_27:.*]] = arith.mulf %[[VAL_23]], %[[VAL_26]] : f32
 // CHECK:                 %[[VAL_28:.*]] = arith.addf %[[VAL_25]], %[[VAL_27]] : f32
 // CHECK:                 memref.store %[[VAL_28]], %[[VAL_13]]{{\[}}%[[VAL_17]], %[[VAL_24]]] : memref<10x30xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_29:.*]] = memref.tensor_load %[[VAL_13]] : memref<10x30xf32>
+// CHECK:           %[[VAL_29:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<10x30xf32>
 // CHECK:           return %[[VAL_29]] : tensor<10x30xf32>
 // CHECK:         }
 func @matmul(%a: tensor<10x20xf32, #DCSR>,
              %b: tensor<20x30xf32>,
              %c: tensor<10x30xf32>) -> tensor<10x30xf32> {
   %0 = linalg.matmul
     ins(%a, %b: tensor<10x20xf32, #DCSR>, tensor<20x30xf32>)
     outs(%c: tensor<10x30xf32>) -> tensor<10x30xf32>
   return %0 : tensor<10x30xf32>
 }
 
 // CHECK-LABEL:   func @conv2d(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<8x8xi32>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<6x6xi32>) -> tensor<6x6xi32> {
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 6 : index
-// CHECK:           %[[VAL_6:.*]] = memref.buffer_cast %[[VAL_0]] : memref<8x8xi32>
+// CHECK:           %[[VAL_6:.*]] = bufferization.to_memref %[[VAL_0]] : memref<8x8xi32>
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<3x3xi32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<6x6xi32>
+// CHECK:           %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<6x6xi32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<6x6xi32>
 // CHECK:           memref.copy %[[VAL_12]], %[[VAL_13]] : memref<6x6xi32> to memref<6x6xi32>
 // CHECK:           %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_16:.*]] = %[[VAL_14]] to %[[VAL_15]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]] = arith.addi %[[VAL_16]], %[[VAL_4]] : index

 // CHECK:                   %[[VAL_29:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_27]], %[[VAL_28]]] : memref<8x8xi32>
 // CHECK:                   %[[VAL_30:.*]] = arith.muli %[[VAL_29]], %[[VAL_23]] : i32
 // CHECK:                   %[[VAL_31:.*]] = arith.addi %[[VAL_26]], %[[VAL_30]] : i32
 // CHECK:                   memref.store %[[VAL_31]], %[[VAL_13]]{{\[}}%[[VAL_25]], %[[VAL_24]]] : memref<6x6xi32>
 // CHECK:                 }
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_32:.*]] = memref.tensor_load %[[VAL_13]] : memref<6x6xi32>
+// CHECK:           %[[VAL_32:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<6x6xi32>
 // CHECK:           return %[[VAL_32]] : tensor<6x6xi32>
 // CHECK:         }
 func @conv2d(%input:  tensor<8x8xi32>,
              %filter: tensor<3x3xi32, #DCSR>,
              %output: tensor<6x6xi32>) -> tensor<6x6xi32> {
   %0 = linalg.conv_2d
     ins  (%input, %filter: tensor<8x8xi32>, tensor<3x3xi32, #DCSR>)
     outs (%output: tensor<6x6xi32>) -> tensor<6x6xi32>
   return %0 : tensor<6x6xi32>
 }
 
 // CHECK-LABEL:   func @quantized_matmul(
 // CHECK-SAME:                           %[[VAL_0:.*]]: tensor<5x3xi8>,
 // CHECK-SAME:                           %[[VAL_1:.*]]: tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                           %[[VAL_2:.*]]: tensor<5x6xi64>) -> tensor<5x6xi64> {
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 2 : i64
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-DAG:       %[[VAL_6:.*]] = arith.constant 5 : index
-// CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_0]] : memref<5x3xi8>
+// CHECK:           %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_0]] : memref<5x3xi8>
 // CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_5]] : tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_5]] : tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<3x6xi8, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<5x6xi64>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_2]] : memref<5x6xi64>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<5x6xi64>
 // CHECK:           memref.copy %[[VAL_13]], %[[VAL_14]] : memref<5x6xi64> to memref<5x6xi64>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_17:.*]] = %[[VAL_15]] to %[[VAL_16]] step %[[VAL_5]] {
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK:             %[[VAL_20:.*]] = arith.addi %[[VAL_17]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_22:.*]] = %[[VAL_19]] to %[[VAL_21]] step %[[VAL_5]] {
 // CHECK:               %[[VAL_23:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_22]]] : memref<?xindex>
 // CHECK:               %[[VAL_24:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_22]]] : memref<?xi8>
 // CHECK:               scf.for %[[VAL_25:.*]] = %[[VAL_4]] to %[[VAL_6]] step %[[VAL_5]] {
 // CHECK:                 %[[VAL_26:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_25]], %[[VAL_23]]] : memref<5x6xi64>
 // CHECK:                 %[[VAL_27:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_25]], %[[VAL_18]]] : memref<5x3xi8>
 // CHECK:                 %[[VAL_28:.*]] = arith.extsi %[[VAL_27]] : i8 to i64
 // CHECK:                 %[[VAL_29:.*]] = arith.subi %[[VAL_28]], %[[VAL_3]] : i64
 // CHECK:                 %[[VAL_30:.*]] = arith.extsi %[[VAL_24]] : i8 to i64
 // CHECK:                 %[[VAL_31:.*]] = arith.muli %[[VAL_29]], %[[VAL_30]] : i64
 // CHECK:                 %[[VAL_32:.*]] = arith.addi %[[VAL_26]], %[[VAL_31]] : i64
 // CHECK:                 memref.store %[[VAL_32]], %[[VAL_14]]{{\[}}%[[VAL_25]], %[[VAL_23]]] : memref<5x6xi64>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_33:.*]] = memref.tensor_load %[[VAL_14]] : memref<5x6xi64>
+// CHECK:           %[[VAL_33:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<5x6xi64>
 // CHECK:           return %[[VAL_33]] : tensor<5x6xi64>
 // CHECK:         }
 func @quantized_matmul(%input1: tensor<5x3xi8>,
                        %input2: tensor<3x6xi8, #DCSR>,
                        %output: tensor<5x6xi64>) -> tensor<5x6xi64> {
   %c0 = arith.constant 0 : i32
   %c2 = arith.constant 2 : i32
   %0 = linalg.quantized_matmul
     ins(%input1, %input2, %c2, %c0 : tensor<5x3xi8>, tensor<3x6xi8, #DCSR>, i32, i32)
     outs(%output : tensor<5x6xi64>) -> tensor<5x6xi64>
   return %0: tensor<5x6xi64>
 }

mlir/test/Dialect/SparseTensor/sparse_lower.mlir

 // CHECK-HIR-SAME:      %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-HIR-SAME:      %[[VAL_2:.*]]: tensor<32xf64>) -> tensor<32xf64> {
 // CHECK-HIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-HIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-HIR-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-HIR:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK-HIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-HIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-HIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-HIR:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf64>
 // CHECK-HIR:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf64> to memref<32xf64>
 // CHECK-HIR:           scf.for %[[VAL_12:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-HIR-DAG:         %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK-HIR-DAG:         %[[VAL_14:.*]] = arith.addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK-HIR-DAG:         %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK-HIR-DAG:         %[[VAL_16:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_12]]] : memref<32xf64>
 // CHECK-HIR:             %[[VAL_17:.*]] = scf.for %[[VAL_18:.*]] = %[[VAL_13]] to %[[VAL_15]] step %[[VAL_5]] iter_args(%[[VAL_19:.*]] = %[[VAL_16]]) -> (f64) {
 // CHECK-HIR:               %[[VAL_20:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK-HIR:               %[[VAL_21:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xf64>
 // CHECK-HIR:               %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_20]]] : memref<64xf64>
 // CHECK-HIR:               %[[VAL_23:.*]] = arith.mulf %[[VAL_21]], %[[VAL_22]] : f64
 // CHECK-HIR:               %[[VAL_24:.*]] = arith.addf %[[VAL_19]], %[[VAL_23]] : f64
 // CHECK-HIR:               scf.yield %[[VAL_24]] : f64
 // CHECK-HIR:             }
 // CHECK-HIR:             memref.store %[[VAL_17]], %[[VAL_11]]{{\[}}%[[VAL_12]]] : memref<32xf64>
 // CHECK-HIR:           }
-// CHECK-HIR:           %[[VAL_26:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_26:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf64>
 // CHECK-HIR:           return %[[VAL_26]] : tensor<32xf64>
 // CHECK-HIR:         }
 
 // CHECK-MIR-LABEL:   func @matvec(
 // CHECK-MIR-SAME:      %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-MIR-SAME:      %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-MIR-SAME:      %[[VAL_2:.*]]: tensor<32xf64>) -> tensor<32xf64> {
 // CHECK-MIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-MIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-MIR-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-MIR:           %[[VAL_6:.*]] = call @sparsePointers(%[[VAL_0]], %[[VAL_5]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_7:.*]] = call @sparseIndices(%[[VAL_0]], %[[VAL_5]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_8:.*]] = call @sparseValuesF64(%[[VAL_0]]) : (!llvm.ptr<i8>) -> memref<?xf64>
-// CHECK-MIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-MIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-MIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-MIR:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf64>
 // CHECK-MIR:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf64> to memref<32xf64>
 // CHECK-MIR:           scf.for %[[VAL_14:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-MIR-DAG:         %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>
 // CHECK-MIR-DAG:         %[[VAL_16:.*]] = arith.addi %[[VAL_14]], %[[VAL_5]] : index
 // CHECK-MIR-DAG:         %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_16]]] : memref<?xindex>
 // CHECK-MIR-DAG:         %[[VAL_18:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_14]]] : memref<32xf64>
 // CHECK-MIR:             %[[VAL_19:.*]] = scf.for %[[VAL_20:.*]] = %[[VAL_15]] to %[[VAL_17]] step %[[VAL_5]] iter_args(%[[VAL_21:.*]] = %[[VAL_18]]) -> (f64) {
 // CHECK-MIR:               %[[VAL_22:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK-MIR:               %[[VAL_23:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_20]]] : memref<?xf64>
 // CHECK-MIR:               %[[VAL_24:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_22]]] : memref<64xf64>
 // CHECK-MIR:               %[[VAL_25:.*]] = arith.mulf %[[VAL_23]], %[[VAL_24]] : f64
 // CHECK-MIR:               %[[VAL_26:.*]] = arith.addf %[[VAL_21]], %[[VAL_25]] : f64
 // CHECK-MIR:               scf.yield %[[VAL_26]] : f64
 // CHECK-MIR:             }
 // CHECK-MIR:             memref.store %[[VAL_19]], %[[VAL_11]]{{\[}}%[[VAL_14]]] : memref<32xf64>
 // CHECK-MIR:           }
-// CHECK-MIR:           %[[VAL_28:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_28:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf64>
 // CHECK-MIR:           return %[[VAL_28]] : tensor<32xf64>
 // CHECK-MIR:         }
 
 // CHECK-LIR-LABEL:   func @matvec(
 // CHECK-LIR-SAME:      %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-LIR-SAME:      %[[VAL_1:.*]]: memref<64xf64>,
 // CHECK-LIR-SAME:      %[[VAL_2:.*]]: memref<32xf64>) -> memref<32xf64> {
 // CHECK-LIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index

mlir/test/Dialect/SparseTensor/sparse_lower_col.mlir

 // CHECK-HIR-SAME:                 %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-HIR-SAME:                 %[[VAL_2:.*]]: tensor<32xf64>) -> tensor<32xf64> {
 // CHECK-HIR-DAG:       %[[VAL_3:.*]] = arith.constant 64 : index
 // CHECK-HIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-HIR-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-HIR:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], dimOrdering = affine_map<(d0, d1) -> (d1, d0)>, pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK-HIR:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], dimOrdering = affine_map<(d0, d1) -> (d1, d0)>, pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK-HIR:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x64xf64, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "compressed" ], dimOrdering = affine_map<(d0, d1) -> (d1, d0)>, pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
-// CHECK-HIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-HIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-HIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-HIR:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf64>
 // CHECK-HIR:           memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf64> to memref<32xf64>
 // CHECK-HIR:           scf.for %[[VAL_12:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-HIR:             %[[VAL_13:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_12]]] : memref<64xf64>
 // CHECK-HIR:             %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
 // CHECK-HIR:             %[[VAL_15:.*]] = arith.addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK-HIR:             %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK-HIR:             scf.for %[[VAL_17:.*]] = %[[VAL_14]] to %[[VAL_16]] step %[[VAL_5]] {
 // CHECK-HIR:               %[[VAL_18:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK-HIR:               %[[VAL_19:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_18]]] : memref<32xf64>
 // CHECK-HIR:               %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xf64>
 // CHECK-HIR:               %[[VAL_21:.*]] = arith.mulf %[[VAL_20]], %[[VAL_13]] : f64
 // CHECK-HIR:               %[[VAL_22:.*]] = arith.addf %[[VAL_19]], %[[VAL_21]] : f64
 // CHECK-HIR:               memref.store %[[VAL_22]], %[[VAL_11]]{{\[}}%[[VAL_18]]] : memref<32xf64>
 // CHECK-HIR:             }
 // CHECK-HIR:           }
-// CHECK-HIR:           %[[VAL_23:.*]] = memref.tensor_load %[[VAL_11]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_23:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf64>
 // CHECK-HIR:           return %[[VAL_23]] : tensor<32xf64>
 // CHECK-HIR:         }
 
 // CHECK-MIR-LABEL:   func @matvec(
 // CHECK-MIR-SAME:                 %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-MIR-SAME:                 %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-MIR-SAME:                 %[[VAL_2:.*]]: tensor<32xf64>) -> tensor<32xf64> {
 // CHECK-MIR-DAG:       %[[VAL_3:.*]] = arith.constant 64 : index
 // CHECK-MIR-DAG:       %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK-MIR-DAG:       %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK-MIR:           %[[VAL_7:.*]] = call @sparsePointers(%[[VAL_0]], %[[VAL_6]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_8:.*]] = call @sparseIndices(%[[VAL_0]], %[[VAL_6]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_9:.*]] = call @sparseValuesF64(%[[VAL_0]]) : (!llvm.ptr<i8>) -> memref<?xf64>
-// CHECK-MIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-MIR:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-MIR:           %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-MIR:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf64>
 // CHECK-MIR:           memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf64> to memref<32xf64>
 // CHECK-MIR:           scf.for %[[VAL_15:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK-MIR:             %[[VAL_16:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_15]]] : memref<64xf64>
 // CHECK-MIR:             %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>
 // CHECK-MIR:             %[[VAL_18:.*]] = arith.addi %[[VAL_15]], %[[VAL_6]] : index
 // CHECK-MIR:             %[[VAL_19:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK-MIR:             scf.for %[[VAL_20:.*]] = %[[VAL_17]] to %[[VAL_19]] step %[[VAL_6]] {
 // CHECK-MIR:               %[[VAL_21:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_20]]] : memref<?xindex>
 // CHECK-MIR:               %[[VAL_22:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_21]]] : memref<32xf64>
 // CHECK-MIR:               %[[VAL_23:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_20]]] : memref<?xf64>
 // CHECK-MIR:               %[[VAL_24:.*]] = arith.mulf %[[VAL_23]], %[[VAL_16]] : f64
 // CHECK-MIR:               %[[VAL_25:.*]] = arith.addf %[[VAL_22]], %[[VAL_24]] : f64
 // CHECK-MIR:               memref.store %[[VAL_25]], %[[VAL_12]]{{\[}}%[[VAL_21]]] : memref<32xf64>
 // CHECK-MIR:             }
 // CHECK-MIR:           }
-// CHECK-MIR:           %[[VAL_26:.*]] = memref.tensor_load %[[VAL_12]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_26:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf64>
 // CHECK-MIR:           return %[[VAL_26]] : tensor<32xf64>
 // CHECK-MIR:         }
 
 // CHECK-LIR-LABEL:   func @matvec(
 // CHECK-LIR-SAME:                 %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-LIR-SAME:                 %[[VAL_1:.*]]: memref<64xf64>,
 // CHECK-LIR-SAME:                 %[[VAL_2:.*]]: memref<32xf64>) -> memref<32xf64> {
 // CHECK-LIR-DAG:       %[[VAL_3:.*]] = arith.constant 64 : index

mlir/test/Dialect/SparseTensor/sparse_lower_inplace.mlir

 // CHECK-HIR-SAME:      %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-HIR-SAME:      %[[VAL_2:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-HIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-HIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-HIR-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-HIR:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x64xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK-HIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-HIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-HIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-HIR:           scf.for %[[VAL_11:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-HIR-DAG:         %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK-HIR-DAG:         %[[VAL_13:.*]] = arith.addi %[[VAL_11]], %[[VAL_5]] : index
 // CHECK-HIR-DAG:         %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK-HIR-DAG:         %[[VAL_15:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK-HIR:             %[[VAL_16:.*]] = scf.for %[[VAL_17:.*]] = %[[VAL_12]] to %[[VAL_14]] step %[[VAL_5]] iter_args(%[[VAL_18:.*]] = %[[VAL_15]]) -> (f64) {
 // CHECK-HIR:               %[[VAL_19:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK-HIR:               %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xf64>
 // CHECK-HIR:               %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<64xf64>
 // CHECK-HIR:               %[[VAL_22:.*]] = arith.mulf %[[VAL_20]], %[[VAL_21]] : f64
 // CHECK-HIR:               %[[VAL_23:.*]] = arith.addf %[[VAL_18]], %[[VAL_22]] : f64
 // CHECK-HIR:               scf.yield %[[VAL_23]] : f64
 // CHECK-HIR:             }
 // CHECK-HIR:             memref.store %[[VAL_16]], %[[VAL_10]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK-HIR:           }
-// CHECK-HIR:           %[[VAL_25:.*]] = memref.tensor_load %[[VAL_10]] : memref<32xf64>
+// CHECK-HIR:           %[[VAL_25:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf64>
 // CHECK-HIR:           return %[[VAL_25]] : tensor<32xf64>
 // CHECK-HIR:         }
 
 // CHECK-MIR-LABEL:   func @matvec(
 // CHECK-MIR-SAME:      %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-MIR-SAME:      %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-MIR-SAME:      %[[VAL_2:.*]]: tensor<32xf64> {linalg.inplaceable = true}) -> tensor<32xf64> {
 // CHECK-MIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index
 // CHECK-MIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-MIR-DAG:       %[[VAL_5:.*]] = arith.constant 1 : index
 // CHECK-MIR:           %[[VAL_6:.*]] = call @sparsePointers(%[[VAL_0]], %[[VAL_5]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_7:.*]] = call @sparseIndices(%[[VAL_0]], %[[VAL_5]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 // CHECK-MIR:           %[[VAL_8:.*]] = call @sparseValuesF64(%[[VAL_0]]) : (!llvm.ptr<i8>) -> memref<?xf64>
-// CHECK-MIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
-// CHECK-MIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<64xf64>
+// CHECK-MIR:           %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf64>
 // CHECK-MIR:           scf.for %[[VAL_11:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-MIR-DAG:         %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK-MIR-DAG:         %[[VAL_13:.*]] = arith.addi %[[VAL_11]], %[[VAL_5]] : index
 // CHECK-MIR-DAG:         %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK-MIR-DAG:         %[[VAL_15:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK-MIR:             %[[VAL_16:.*]] = scf.for %[[VAL_17:.*]] = %[[VAL_12]] to %[[VAL_14]] step %[[VAL_5]] iter_args(%[[VAL_18:.*]] = %[[VAL_15]]) -> (f64) {
 // CHECK-MIR:               %[[VAL_19:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_17]]] : memref<?xindex>
 // CHECK-MIR:               %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_17]]] : memref<?xf64>
 // CHECK-MIR:               %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<64xf64>
 // CHECK-MIR:               %[[VAL_22:.*]] = arith.mulf %[[VAL_20]], %[[VAL_21]] : f64
 // CHECK-MIR:               %[[VAL_23:.*]] = arith.addf %[[VAL_18]], %[[VAL_22]] : f64
 // CHECK-MIR:               scf.yield %[[VAL_23]] : f64
 // CHECK-MIR:             }
 // CHECK-MIR:             memref.store %[[VAL_16]], %[[VAL_10]]{{\[}}%[[VAL_11]]] : memref<32xf64>
 // CHECK-MIR:           }
-// CHECK-MIR:           %[[VAL_25:.*]] = memref.tensor_load %[[VAL_10]] : memref<32xf64>
+// CHECK-MIR:           %[[VAL_25:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf64>
 // CHECK-MIR:           return %[[VAL_25]] : tensor<32xf64>
 // CHECK-MIR:         }
 
 // CHECK-LIR-LABEL:   func @matvec(
 // CHECK-LIR-SAME:      %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-LIR-SAME:      %[[VAL_1:.*]]: memref<64xf64>,
 // CHECK-LIR-SAME:      %[[VAL_2:.*]]: memref<32xf64> {linalg.inplaceable = true}) -> memref<32xf64> {
 // CHECK-LIR-DAG:       %[[VAL_3:.*]] = arith.constant 32 : index

mlir/test/Dialect/SparseTensor/sparse_nd.mlir

 // CHECK:           %[[VAL_5:.*]] = arith.constant 10 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 20 : index
 // CHECK:           %[[VAL_7:.*]] = arith.constant 30 : index
 // CHECK:           %[[VAL_8:.*]] = arith.constant 60 : index
 // CHECK:           %[[VAL_9:.*]] = arith.constant 70 : index
 // CHECK:           %[[VAL_10:.*]] = arith.constant 80 : index
 // CHECK:           %[[VAL_11:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_12:.*]] = arith.constant 1 : index
-// CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_0]] : memref<10x20x30x40x50x60x70x80xf32>
+// CHECK:           %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_0]] : memref<10x20x30x40x50x60x70x80xf32>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<80x70x60x50x40x30x20x10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense", "compressed", "compressed", "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_15:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<80x70x60x50x40x30x20x10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense", "compressed", "compressed", "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_16:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<80x70x60x50x40x30x20x10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense", "compressed", "compressed", "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<80x70x60x50x40x30x20x10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense", "compressed", "compressed", "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<80x70x60x50x40x30x20x10xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense", "dense", "dense", "compressed", "compressed", "dense", "dense", "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
-// CHECK:           %[[VAL_19:.*]] = memref.buffer_cast %[[VAL_2]] : memref<10x20x30x40x50x60x70x80xf32>
+// CHECK:           %[[VAL_19:.*]] = bufferization.to_memref %[[VAL_2]] : memref<10x20x30x40x50x60x70x80xf32>
 // CHECK:           %[[VAL_20:.*]] = memref.alloc() : memref<10x20x30x40x50x60x70x80xf32>
 // CHECK:           memref.copy %[[VAL_19]], %[[VAL_20]] : memref<10x20x30x40x50x60x70x80xf32> to memref<10x20x30x40x50x60x70x80xf32>
 // CHECK:           scf.for %[[VAL_21:.*]] = %[[VAL_11]] to %[[VAL_10]] step %[[VAL_12]] {
 // CHECK:             scf.for %[[VAL_22:.*]] = %[[VAL_11]] to %[[VAL_9]] step %[[VAL_12]] {
 // CHECK:               %[[VAL_23:.*]] = arith.muli %[[VAL_21]], %[[VAL_9]] : index
 // CHECK:               %[[VAL_24:.*]] = arith.addi %[[VAL_23]], %[[VAL_22]] : index
 // CHECK:               scf.for %[[VAL_25:.*]] = %[[VAL_11]] to %[[VAL_8]] step %[[VAL_12]] {
 // CHECK:                 %[[VAL_26:.*]] = arith.muli %[[VAL_24]], %[[VAL_8]] : index

 // CHECK:                         }
 // CHECK:                       }
 // CHECK:                     }
 // CHECK:                   }
 // CHECK:                 }
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_50:.*]] = memref.tensor_load %[[VAL_20]] : memref<10x20x30x40x50x60x70x80xf32>
+// CHECK:           %[[VAL_50:.*]] = bufferization.to_tensor %[[VAL_20]] : memref<10x20x30x40x50x60x70x80xf32>
 // CHECK:           return %[[VAL_50]] : tensor<10x20x30x40x50x60x70x80xf32>
 // CHECK:         }
 func @mul(%arga: tensor<10x20x30x40x50x60x70x80xf32>,
           %argb: tensor<80x70x60x50x40x30x20x10xf32, #SparseTensor>,
           %argx: tensor<10x20x30x40x50x60x70x80xf32>)
         -> tensor<10x20x30x40x50x60x70x80xf32> {
   %0 = linalg.generic #trait_mul
     ins(%arga, %argb: tensor<10x20x30x40x50x60x70x80xf32>,
                       tensor<80x70x60x50x40x30x20x10xf32, #SparseTensor>)
     outs(%argx: tensor<10x20x30x40x50x60x70x80xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = arith.mulf %a, %b : f32
         linalg.yield %0 : f32
     }      -> tensor<10x20x30x40x50x60x70x80xf32>
   return %0 : tensor<10x20x30x40x50x60x70x80xf32>
 }

mlir/test/Dialect/SparseTensor/sparse_perm.mlir

 // CHECK-SAME:                          %[[VAL_0:.*]]: tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                          %[[VAL_1:.*]]: tensor<20x30x10xf32>) -> tensor<20x30x10xf32> {
 // CHECK:           %[[VAL_2:.*]] = arith.constant 20 : index
 // CHECK:           %[[VAL_3:.*]] = arith.constant 30 : index
 // CHECK:           %[[VAL_4:.*]] = arith.constant 10 : index
 // CHECK:           %[[VAL_5:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20x30xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_1]] : memref<20x30x10xf32>
+// CHECK:           %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<20x30x10xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<20x30x10xf32>
 // CHECK:           memref.copy %[[VAL_8]], %[[VAL_9]] : memref<20x30x10xf32> to memref<20x30x10xf32>
 // CHECK:           scf.for %[[VAL_10:.*]] = %[[VAL_5]] to %[[VAL_3]] step %[[VAL_6]] {
 // CHECK:             scf.for %[[VAL_11:.*]] = %[[VAL_5]] to %[[VAL_4]] step %[[VAL_6]] {
 // CHECK:               %[[VAL_12:.*]] = arith.muli %[[VAL_10]], %[[VAL_4]] : index
 // CHECK:               %[[VAL_13:.*]] = arith.addi %[[VAL_12]], %[[VAL_11]] : index
 // CHECK:               scf.for %[[VAL_14:.*]] = %[[VAL_5]] to %[[VAL_2]] step %[[VAL_6]] {
 // CHECK:                 %[[VAL_15:.*]] = arith.muli %[[VAL_13]], %[[VAL_2]] : index
 // CHECK:                 %[[VAL_16:.*]] = arith.addi %[[VAL_15]], %[[VAL_14]] : index
 // CHECK:                 %[[VAL_17:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_16]]] : memref<?xf32>
 // CHECK:                 memref.store %[[VAL_17]], %[[VAL_9]]{{\[}}%[[VAL_14]], %[[VAL_10]], %[[VAL_11]]] : memref<20x30x10xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_18:.*]] = memref.tensor_load %[[VAL_9]] : memref<20x30x10xf32>
+// CHECK:           %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<20x30x10xf32>
 // CHECK:           return %[[VAL_18]] : tensor<20x30x10xf32>
 // CHECK:         }
 func @sparse_static_dims(%arga: tensor<10x20x30xf32, #X>,
                          %argx: tensor<20x30x10xf32>) -> tensor<20x30x10xf32> {
   %0 = linalg.generic #trait
     ins(%arga: tensor<10x20x30xf32, #X>)
     outs(%argx: tensor<20x30x10xf32>) {
       ^bb(%a : f32, %x: f32):
         linalg.yield %a : f32
   } -> tensor<20x30x10xf32>
   return %0 : tensor<20x30x10xf32>
 }
 
 // CHECK-LABEL:   func @sparse_dynamic_dims(
 // CHECK-SAME:                          %[[VAL_0:.*]]: tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>,
 // CHECK-SAME:                          %[[VAL_1:.*]]: tensor<?x?x?xf32>) -> tensor<?x?x?xf32> {
 // CHECK-DAG:       %[[VAL_2:.*]] = arith.constant 2 : index
 // CHECK-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-DAG:       %[[VAL_4:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_6:.*]] = tensor.dim %[[VAL_1]], %[[VAL_3]] : tensor<?x?x?xf32>
 // CHECK:           %[[VAL_7:.*]] = tensor.dim %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32>
 // CHECK:           %[[VAL_8:.*]] = tensor.dim %[[VAL_1]], %[[VAL_2]] : tensor<?x?x?xf32>
-// CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<?x?x?xf32>
+// CHECK:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<?x?x?xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc(%[[VAL_6]], %[[VAL_7]], %[[VAL_8]]) : memref<?x?x?xf32>
 // CHECK:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<?x?x?xf32> to memref<?x?x?xf32>
 // CHECK:           scf.for %[[VAL_11:.*]] = %[[VAL_3]] to %[[VAL_7]] step %[[VAL_4]] {
 // CHECK:             scf.for %[[VAL_12:.*]] = %[[VAL_3]] to %[[VAL_8]] step %[[VAL_4]] {
 // CHECK:               %[[VAL_13:.*]] = arith.muli %[[VAL_8]], %[[VAL_11]] : index
 // CHECK:               %[[VAL_14:.*]] = arith.addi %[[VAL_13]], %[[VAL_12]] : index
 // CHECK:               scf.for %[[VAL_15:.*]] = %[[VAL_3]] to %[[VAL_6]] step %[[VAL_4]] {
 // CHECK:                 %[[VAL_16:.*]] = arith.muli %[[VAL_6]], %[[VAL_14]] : index
 // CHECK:                 %[[VAL_17:.*]] = arith.addi %[[VAL_16]], %[[VAL_15]] : index
 // CHECK:                 %[[VAL_18:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_17]]] : memref<?xf32>
 // CHECK:                 memref.store %[[VAL_18]], %[[VAL_10]]{{\[}}%[[VAL_15]], %[[VAL_11]], %[[VAL_12]]] : memref<?x?x?xf32>
 // CHECK:               }
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_19:.*]] = memref.tensor_load %[[VAL_10]] : memref<?x?x?xf32>
+// CHECK:           %[[VAL_19:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<?x?x?xf32>
 // CHECK:           return %[[VAL_19]] : tensor<?x?x?xf32>
 // CHECK:         }
 func @sparse_dynamic_dims(%arga: tensor<?x?x?xf32, #X>,
                           %argx: tensor<?x?x?xf32>) -> tensor<?x?x?xf32> {
   %0 = linalg.generic #trait
     ins(%arga: tensor<?x?x?xf32, #X>)
     outs(%argx: tensor<?x?x?xf32>) {
       ^bb(%a : f32, %x: f32):
         linalg.yield %a : f32
   } -> tensor<?x?x?xf32>
   return %0 : tensor<?x?x?xf32>
 }

mlir/test/Dialect/SparseTensor/sparse_perm_lower.mlir

 // CHECK-HIR-SAME:      %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK-HIR-DAG:       %[[VAL_2:.*]] = arith.constant 1 : index
 // CHECK-HIR-DAG:       %[[VAL_3:.*]] = arith.constant 0 : index
 // CHECK-HIR-DAG:       %[[VAL_4:.*]] = arith.constant 2 : index
 // CHECK-HIR:           %[[VAL_5:.*]] = tensor.dim %[[VAL_0]], %[[VAL_4]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_6:.*]] = tensor.dim %[[VAL_0]], %[[VAL_3]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_7:.*]] = tensor.dim %[[VAL_0]], %[[VAL_2]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK-HIR:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?x?xf32, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK-HIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK-HIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK-HIR:           %[[VAL_10:.*]] = memref.alloc() : memref<f32>
 // CHECK-HIR:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<f32> to memref<f32>
 // CHECK-HIR:           %[[VAL_11:.*]] = memref.load %[[VAL_10]][] : memref<f32>
 // CHECK-HIR:           %[[VAL_12:.*]] = scf.for %[[VAL_13:.*]] = %[[VAL_3]] to %[[VAL_5]] step %[[VAL_2]] iter_args(%[[VAL_14:.*]] = %[[VAL_11]]) -> (f32) {
 // CHECK-HIR:             %[[VAL_15:.*]] = scf.for %[[VAL_16:.*]] = %[[VAL_3]] to %[[VAL_6]] step %[[VAL_2]] iter_args(%[[VAL_17:.*]] = %[[VAL_14]]) -> (f32) {
 // CHECK-HIR:               %[[VAL_18:.*]] = arith.muli %[[VAL_6]], %[[VAL_13]] : index
 // CHECK-HIR:               %[[VAL_19:.*]] = arith.addi %[[VAL_18]], %[[VAL_16]] : index
 // CHECK-HIR:               %[[VAL_20:.*]] = scf.for %[[VAL_21:.*]] = %[[VAL_3]] to %[[VAL_7]] step %[[VAL_2]] iter_args(%[[VAL_22:.*]] = %[[VAL_17]]) -> (f32) {
 // CHECK-HIR:                 %[[VAL_23:.*]] = arith.muli %[[VAL_7]], %[[VAL_19]] : index
 // CHECK-HIR:                 %[[VAL_24:.*]] = arith.addi %[[VAL_23]], %[[VAL_21]] : index
 // CHECK-HIR:                 %[[VAL_25:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK-HIR:                 %[[VAL_26:.*]] = arith.addf %[[VAL_22]], %[[VAL_25]] : f32
 // CHECK-HIR:                 scf.yield %[[VAL_26]] : f32
 // CHECK-HIR:               }
 // CHECK-HIR:               scf.yield %[[VAL_20]] : f32
 // CHECK-HIR:             }
 // CHECK-HIR:             scf.yield %[[VAL_15]] : f32
 // CHECK-HIR:           }
 // CHECK-HIR:           memref.store %[[VAL_12]], %[[VAL_10]][] : memref<f32>
-// CHECK-HIR:           %[[VAL_30:.*]] = memref.tensor_load %[[VAL_10]] : memref<f32>
+// CHECK-HIR:           %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<f32>
 // CHECK-HIR:           return %[[VAL_30]] : tensor<f32>
 // CHECK-HIR:         }
 //
 // CHECK-MIR-LABEL:   func @sparse_dynamic_dims(
 // CHECK-MIR-SAME:      %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-MIR-SAME:      %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK-MIR-DAG:       %[[VAL_2:.*]] = arith.constant 2 : index
 // CHECK-MIR-DAG:       %[[VAL_3:.*]] = arith.constant 1 : index
 // CHECK-MIR-DAG:       %[[VAL_4:.*]] = arith.constant 0 : index
 // CHECK-MIR:           %[[VAL_5:.*]] = call @sparseDimSize(%[[VAL_0]], %[[VAL_4]]) : (!llvm.ptr<i8>, index) -> index
 // CHECK-MIR:           %[[VAL_6:.*]] = call @sparseDimSize(%[[VAL_0]], %[[VAL_3]]) : (!llvm.ptr<i8>, index) -> index
 // CHECK-MIR:           %[[VAL_7:.*]] = call @sparseDimSize(%[[VAL_0]], %[[VAL_2]]) : (!llvm.ptr<i8>, index) -> index
 // CHECK-MIR:           %[[VAL_8:.*]] = call @sparseValuesF32(%[[VAL_0]]) : (!llvm.ptr<i8>) -> memref<?xf32>
-// CHECK-MIR:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
+// CHECK-MIR:           %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
 // CHECK-MIR:           %[[VAL_10:.*]] = memref.alloc() : memref<f32>
 // CHECK-MIR:           memref.copy %[[VAL_9]], %[[VAL_10]] : memref<f32> to memref<f32>
 // CHECK-MIR:           %[[VAL_11:.*]] = memref.load %[[VAL_10]][] : memref<f32>
 // CHECK-MIR:           %[[VAL_12:.*]] = scf.for %[[VAL_13:.*]] = %[[VAL_4]] to %[[VAL_5]] step %[[VAL_3]] iter_args(%[[VAL_14:.*]] = %[[VAL_11]]) -> (f32) {
 // CHECK-MIR:             %[[VAL_15:.*]] = scf.for %[[VAL_16:.*]] = %[[VAL_4]] to %[[VAL_6]] step %[[VAL_3]] iter_args(%[[VAL_17:.*]] = %[[VAL_14]]) -> (f32) {
 // CHECK-MIR:               %[[VAL_18:.*]] = arith.muli %[[VAL_6]], %[[VAL_13]] : index
 // CHECK-MIR:               %[[VAL_19:.*]] = arith.addi %[[VAL_18]], %[[VAL_16]] : index
 // CHECK-MIR:               %[[VAL_20:.*]] = scf.for %[[VAL_21:.*]] = %[[VAL_4]] to %[[VAL_7]] step %[[VAL_3]] iter_args(%[[VAL_22:.*]] = %[[VAL_17]]) -> (f32) {
 // CHECK-MIR:                 %[[VAL_23:.*]] = arith.muli %[[VAL_7]], %[[VAL_19]] : index
 // CHECK-MIR:                 %[[VAL_24:.*]] = arith.addi %[[VAL_23]], %[[VAL_21]] : index
 // CHECK-MIR:                 %[[VAL_25:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK-MIR:                 %[[VAL_26:.*]] = arith.addf %[[VAL_22]], %[[VAL_25]] : f32
 // CHECK-MIR:                 scf.yield %[[VAL_26]] : f32
 // CHECK-MIR:               }
 // CHECK-MIR:               scf.yield %[[VAL_20]] : f32
 // CHECK-MIR:             }
 // CHECK-MIR:             scf.yield %[[VAL_15]] : f32
 // CHECK-MIR:           }
 // CHECK-MIR:           memref.store %[[VAL_12]], %[[VAL_10]][] : memref<f32>
-// CHECK-MIR:           %[[VAL_30:.*]] = memref.tensor_load %[[VAL_10]] : memref<f32>
+// CHECK-MIR:           %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<f32>
 // CHECK-MIR:           return %[[VAL_30]] : tensor<f32>
 // CHECK-MIR:         }
 func @sparse_dynamic_dims(%arga: tensor<?x?x?xf32, #X>,
                           %argx: tensor<f32>) -> tensor<f32> {
   %0 = linalg.generic #trait
     ins(%arga: tensor<?x?x?xf32, #X>)
     outs(%argx: tensor<f32>) {
       ^bb(%a : f32, %x: f32):
         %0 = arith.addf %x, %a : f32
         linalg.yield %0 : f32
   } -> tensor<f32>
   return %0 : tensor<f32>
 }

mlir/test/Dialect/SparseTensor/sparse_scalars.mlir

 // CHECK:           %[[VAL_6:.*]] = arith.constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = arith.addf %[[VAL_2]], %[[VAL_3]] : f32
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xindex>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xindex>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xindex>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xindex>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32, #sparse_tensor.encoding<{{.*}}>> to memref<?xf32>
-// CHECK:           %[[VAL_14:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_4]] : memref<32x16xf32>
+// CHECK:           %[[VAL_14:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_4]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_14]][] : memref<f32>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_6]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
 // CHECK:           scf.for %[[VAL_19:.*]] = %[[VAL_17]] to %[[VAL_18]] step %[[VAL_7]] {
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_19]]] : memref<?xindex>
 // CHECK:             %[[VAL_22:.*]] = arith.addi %[[VAL_19]], %[[VAL_7]] : index
 // CHECK:             %[[VAL_23:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_22]]] : memref<?xindex>
 // CHECK:             scf.for %[[VAL_24:.*]] = %[[VAL_21]] to %[[VAL_23]] step %[[VAL_7]] {
 // CHECK:               %[[VAL_25:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_24]]] : memref<?xindex>
 // CHECK:               %[[VAL_26:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_24]]] : memref<?xf32>
 // CHECK:               %[[VAL_27:.*]] = arith.mulf %[[VAL_26]], %[[VAL_16]] : f32
 // CHECK:               %[[VAL_28:.*]] = arith.mulf %[[VAL_27]], %[[VAL_2]] : f32
 // CHECK:               %[[VAL_29:.*]] = arith.mulf %[[VAL_28]], %[[VAL_3]] : f32
 // CHECK:               %[[VAL_30:.*]] = arith.mulf %[[VAL_29]], %[[VAL_8]] : f32
 // CHECK:               %[[VAL_31:.*]] = arith.mulf %[[VAL_30]], %[[VAL_5]] : f32
 // CHECK:               %[[VAL_32:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_20]], %[[VAL_25]]] : memref<32x16xf32>
 // CHECK:               %[[VAL_33:.*]] = arith.addf %[[VAL_31]], %[[VAL_32]] : f32
 // CHECK:               memref.store %[[VAL_33]], %[[VAL_15]]{{\[}}%[[VAL_20]], %[[VAL_25]]] : memref<32x16xf32>
 // CHECK:             }
 // CHECK:           }
-// CHECK:           %[[VAL_34:.*]] = memref.tensor_load %[[VAL_15]] : memref<32x16xf32>
+// CHECK:           %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<32x16xf32>
 // CHECK:           return %[[VAL_34]] : tensor<32x16xf32>
 // CHECK:         }
 func @mul(%arga: tensor<32x16xf32, #SparseMatrix>,
           %argp: tensor<f32>,
           %argq: f32,
           %argr: f32,
           %argx: tensor<32x16xf32> {linalg.inplaceable = true}) -> tensor<32x16xf32> {
   %s = arith.addf %argq, %argr : f32

mlir/test/Dialect/SparseTensor/sparse_vector_chain.mlir

 // CHECK-DAG:       %[[VAL_7:.*]] = arith.constant 64 : index
 // CHECK-DAG:       %[[VAL_8:.*]] = arith.constant 1 : index
 // CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_8]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_8]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_8]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_8]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
 // CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<64x32xf64, #sparse_tensor.encoding<{{{.*}}}>>
-// CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_0]] : memref<f64>
+// CHECK:           %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_0]] : memref<f64>
 // CHECK:           %[[VAL_16:.*]] = tensor.extract %[[VAL_0]][] : tensor<f64>
 // CHECK:           %[[VAL_17:.*]] = scf.for %[[VAL_18:.*]] = %[[VAL_6]] to %[[VAL_7]] step %[[VAL_8]] iter_args(%[[VAL_19:.*]] = %[[VAL_16]]) -> (f64) {
 // CHECK:             %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_21:.*]] = arith.addi %[[VAL_18]], %[[VAL_8]] : index
 // CHECK:             %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_21]]] : memref<?xindex>
 // CHECK:             %[[VAL_23:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_18]]] : memref<?xindex>
 // CHECK:             %[[VAL_24:.*]] = arith.addi %[[VAL_18]], %[[VAL_8]] : index
 // CHECK:             %[[VAL_25:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_24]]] : memref<?xindex>

 // CHECK:               %[[VAL_82:.*]] = arith.addf %[[VAL_77]], %[[VAL_81]] : vector<8xf64>
 // CHECK:               %[[VAL_83:.*]] = select %[[VAL_80]], %[[VAL_82]], %[[VAL_77]] : vector<8xi1>, vector<8xf64>
 // CHECK:               scf.yield %[[VAL_83]] : vector<8xf64>
 // CHECK:             }
 // CHECK:             %[[VAL_84:.*]] = vector.reduction "add", %[[VAL_85:.*]] : vector<8xf64> into f64
 // CHECK:             scf.yield %[[VAL_84]] : f64
 // CHECK:           }
 // CHECK:           memref.store %[[VAL_86:.*]], %[[VAL_15]][] : memref<f64>
-// CHECK:           %[[VAL_87:.*]] = memref.tensor_load %[[VAL_15]] : memref<f64>
+// CHECK:           %[[VAL_87:.*]] = bufferization.to_tensor %[[VAL_15]] : memref<f64>
 // CHECK:           return %[[VAL_87]] : tensor<f64>
 // CHECK:         }
 func @sparse_matrix_sum(%argx: tensor<f64> {linalg.inplaceable = true},
                          %arga: tensor<64x32xf64, #SparseMatrix>,
                          %argb: tensor<64x32xf64, #SparseMatrix>) -> tensor<f64> {
   %0 = linalg.generic #trait
      ins(%arga, %argb: tensor<64x32xf64, #SparseMatrix>,
                        tensor<64x32xf64, #SparseMatrix>)
       outs(%argx: tensor<f64>) {
       ^bb(%a: f64, %b: f64, %x: f64):
         %m = arith.addf %a, %b : f64
         %t = arith.addf %x, %m : f64
         linalg.yield %t : f64
   } -> tensor<f64>
   return %0 : tensor<f64>
 }

mlir/test/Dialect/Standard/bufferize.mlir

 // RUN: mlir-opt %s -std-bufferize | FileCheck %s
 
 // CHECK-LABEL:   func @select(
 // CHECK-SAME:                 %[[PRED:.*]]: i1,
 // CHECK-SAME:                 %[[TRUE_VAL:.*]]: tensor<f32>,
 // CHECK-SAME:                 %[[FALSE_VAL:.*]]: tensor<f32>) -> tensor<f32> {
-// CHECK-DAG:           %[[TRUE_VAL_MEMREF:.*]] = memref.buffer_cast %[[TRUE_VAL]] : memref<f32>
-// CHECK-DAG:           %[[FALSE_VAL_MEMREF:.*]] = memref.buffer_cast %[[FALSE_VAL]] : memref<f32>
+// CHECK-DAG:           %[[TRUE_VAL_MEMREF:.*]] = bufferization.to_memref %[[TRUE_VAL]] : memref<f32>
+// CHECK-DAG:           %[[FALSE_VAL_MEMREF:.*]] = bufferization.to_memref %[[FALSE_VAL]] : memref<f32>
 // CHECK:           %[[RET_MEMREF:.*]] = select %[[PRED]], %[[TRUE_VAL_MEMREF]], %[[FALSE_VAL_MEMREF]] : memref<f32>
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[RET_MEMREF]] : memref<f32>
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[RET_MEMREF]] : memref<f32>
 // CHECK:           return %[[RET]] : tensor<f32>
 func @select(%arg0: i1, %arg1: tensor<f32>, %arg2: tensor<f32>) -> tensor<f32> {
   %0 = select %arg0, %arg1, %arg2 : tensor<f32>
   return %0 : tensor<f32>
 }

mlir/test/Dialect/Standard/func-bufferize.mlir

 func private @sink(tensor<f32>)
 func @call_sink(%arg0: tensor<f32>) {
   call @sink(%arg0) : (tensor<f32>) -> ()
   return
 }
 
 // CHECK-LABEL:   func @unconverted_op_in_body() -> memref<f32> {
 // CHECK:           %[[TENSOR:.*]] = "test.source"() : () -> tensor<f32>
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<f32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<f32>
 // CHECK:           return %[[MEMREF]] : memref<f32>
 func @unconverted_op_in_body() -> tensor<f32> {
   %0 = "test.source"() : () -> tensor<f32>
   return %0 : tensor<f32>
 }
 
 // -----
 

mlir/test/Dialect/Standard/tensor-constant-bufferize.mlir

 // CHECK-NOT: alignment
 
 // ALIGNED: memref.global "private" constant @__constant_3x4xf32 : memref<3x4xf32> = dense<7.000000e+00>
 // ALIGNED-SAME: {alignment = 64 : i64}
 
 // CHECK: @basic
 func @basic() -> tensor<3x4xf32> {
   // CHECK: %[[MEMREF:.*]] = memref.get_global @__constant_3x4xf32 : memref<3x4xf32>
-  // CHECK: %[[TENSOR:.*]] = memref.tensor_load %[[MEMREF]]
+  // CHECK: %[[TENSOR:.*]] = bufferization.to_tensor %[[MEMREF]]
   %0 = arith.constant dense<7.0> : tensor<3x4xf32>
   // CHECK: return %[[TENSOR]]
   return %0 : tensor<3x4xf32>
 }
 
 // CHECK: }
 
 // -----

mlir/test/Dialect/Tensor/bufferize.mlir

 // RUN: mlir-opt %s -tensor-bufferize | FileCheck %s
 
 // CHECK-LABEL:   func @dim(
 // CHECK-SAME:              %[[TENSOR:.*]]: tensor<f32>,
 // CHECK-SAME:              %[[INDEX:.*]]: index) -> index {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<f32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<f32>
 // CHECK:           %[[EXTENT:.*]] = memref.dim %[[MEMREF]], %[[INDEX]] : memref<f32>
 // CHECK:           return %[[EXTENT]] : index
 func @dim(%arg0: tensor<f32>, %arg1: index) -> index {
   %0 = tensor.dim %arg0, %arg1 : tensor<f32>
   return %0 : index
 }
 
 // CHECK-LABEL:   func @tensor.cast(
 // CHECK-SAME:                      %[[TENSOR:.*]]: tensor<?xindex>) -> tensor<2xindex> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]]
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]]
 // CHECK:           %[[CASTED:.*]] = memref.cast %[[MEMREF]] : memref<?xindex> to memref<2xindex>
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[CASTED]]
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[CASTED]]
 // CHECK:           return %[[RET]] : tensor<2xindex>
 func @tensor.cast(%arg0: tensor<?xindex>) -> tensor<2xindex> {
   %0 = tensor.cast %arg0 : tensor<?xindex> to tensor<2xindex>
   return %0 : tensor<2xindex>
 }
 
 // CHECK-LABEL:   func @tensor.cast_from_unranked(
 // CHECK-SAME:                                    %[[TENSOR:.*]]: tensor<*xf32>) -> tensor<2xf32> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<*xf32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<*xf32>
 // CHECK:           %[[CASTED_MEMREF:.*]] = memref.cast %[[MEMREF]] : memref<*xf32> to memref<2xf32>
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[CASTED_MEMREF]] : memref<2xf32>
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[CASTED_MEMREF]] : memref<2xf32>
 // CHECK:           return %[[RET]] : tensor<2xf32>
 func @tensor.cast_from_unranked(%arg0: tensor<*xf32>) -> tensor<2xf32> {
   %0 = tensor.cast %arg0 : tensor<*xf32> to tensor<2xf32>
   return %0 : tensor<2xf32>
 }
 
 // CHECK-LABEL:   func @tensor.cast_to_unranked(
 // CHECK-SAME:                                  %[[TENSOR:.*]]: tensor<2xf32>) -> tensor<*xf32> {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<2xf32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<2xf32>
 // CHECK:           %[[CASTED_MEMREF:.*]] = memref.cast %[[MEMREF]] : memref<2xf32> to memref<*xf32>
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[CASTED_MEMREF]] : memref<*xf32>
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[CASTED_MEMREF]] : memref<*xf32>
 // CHECK:           return %[[RET]] : tensor<*xf32>
 func @tensor.cast_to_unranked(%arg0: tensor<2xf32>) -> tensor<*xf32> {
   %0 = tensor.cast %arg0 : tensor<2xf32> to tensor<*xf32>
   return %0 : tensor<*xf32>
 }
 
 // CHECK-LABEL:   func @tensor.extract(
 // CHECK-SAME:                  %[[TENSOR:.*]]: tensor<?xf32>,
 // CHECK-SAME:                  %[[IDX:.*]]: index) -> f32 {
-// CHECK:           %[[MEMREF:.*]] = memref.buffer_cast %[[TENSOR]] : memref<?xf32>
+// CHECK:           %[[MEMREF:.*]] = bufferization.to_memref %[[TENSOR]] : memref<?xf32>
 // CHECK:           %[[RET:.*]] = memref.load %[[MEMREF]][%[[IDX]]] : memref<?xf32>
 // CHECK:           return %[[RET]] : f32
 // CHECK:         }
 func @tensor.extract(%arg0: tensor<?xf32>, %arg1: index) -> f32 {
   %0 = tensor.extract %arg0[%arg1] : tensor<?xf32>
   return %0 : f32
 }
 
 // CHECK-LABEL:   func @tensor.from_elements(
 // CHECK-SAME:                               %[[ELEM0:.*]]: index,
 // CHECK-SAME:                               %[[ELEM1:.*]]: index) -> tensor<2xindex> {
 // CHECK:           %[[MEMREF:.*]] = memref.alloc()
 // CHECK:           %[[C0:.*]] = arith.constant 0 : index
 // CHECK:           store %[[ELEM0]], %[[MEMREF]][%[[C0]]]
 // CHECK:           %[[C1:.*]] = arith.constant 1 : index
 // CHECK:           store %[[ELEM1]], %[[MEMREF]][%[[C1]]]
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[MEMREF]]
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[MEMREF]]
 // CHECK:           return %[[RET]] : tensor<2xindex>
 func @tensor.from_elements(%arg0: index, %arg1: index) -> tensor<2xindex> {
   %0 = tensor.from_elements %arg0, %arg1 : tensor<2xindex>
   return %0 : tensor<2xindex>
 }
 
 // CHECK-LABEL:   func @tensor.generate(
 // CHECK-SAME:                                       %[[ARG:.*]]: tensor<*xf32>,
 // CHECK-SAME:                                       %[[DYNAMIC_EXTENT:.*]]: index) -> tensor<?xindex> {
-// CHECK:           %[[CASTED:.*]] = memref.buffer_cast %[[ARG]] : memref<*xf32>
+// CHECK:           %[[CASTED:.*]] = bufferization.to_memref %[[ARG]] : memref<*xf32>
 // CHECK:           %[[MEMREF:.*]] = memref.alloc(%[[DYNAMIC_EXTENT]]) : memref<?xindex>
 // CHECK:           %[[C0:.*]] = arith.constant 0 : index
 // CHECK:           %[[C1:.*]] = arith.constant 1 : index
 // CHECK:           scf.parallel (%[[I:.*]]) = (%[[C0]]) to (%[[DYNAMIC_EXTENT]]) step (%[[C1]]) {
 // CHECK:             %[[ELEM:.*]] = memref.dim %[[CASTED]], %[[I]] : memref<*xf32>
 // CHECK:             store %[[ELEM]], %[[MEMREF]][%[[I]]] : memref<?xindex>
 // CHECK:             scf.yield
 // CHECK:           }
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[MEMREF]] : memref<?xindex>
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[MEMREF]] : memref<?xindex>
 // CHECK:           return %[[RET]] : tensor<?xindex>
 // CHECK:         }
 func @tensor.generate(%arg: tensor<*xf32>, %dynamic_extent: index) -> tensor<?xindex> {
   %result = tensor.generate %dynamic_extent {
   ^bb0(%i : index):
     %elem = tensor.dim %arg, %i : tensor<*xf32>
     tensor.yield %elem : index
   } : tensor<?xindex>

 // CHECK:           %[[C0:.*]] = arith.constant 0 : index
 // CHECK:           %[[C1:.*]] = arith.constant 1 : index
 // CHECK:           %[[C16:.*]] = arith.constant 16 : index
 // CHECK:           scf.parallel (%[[I:.*]], %[[J:.*]]) = (%[[C0]], %[[C0]]) to (%[[C16]], %[[DYNAMIC_EXTENT]]) step (%[[C1]], %[[C1]]) {
 // CHECK:             %[[VAL_7:.*]] = arith.addi %[[I]], %[[J]] : index
 // CHECK:             store %[[VAL_7]], %[[MEMREF]][%[[I]], %[[J]]] : memref<16x?xindex>
 // CHECK:             scf.yield
 // CHECK:           }
-// CHECK:           %[[RET:.*]] = memref.tensor_load %[[MEMREF]] : memref<16x?xindex>
+// CHECK:           %[[RET:.*]] = bufferization.to_tensor %[[MEMREF]] : memref<16x?xindex>
 // CHECK:           return %[[RET]] : tensor<16x?xindex>
 // CHECK:         }
 func @tensor.generate_static_and_dynamic(%arg0: index) -> tensor<16x?xindex> {
   %result = tensor.generate %arg0 {
   ^bb0(%i: index, %j: index):
     %sum = arith.addi %i, %j : index
     tensor.yield %sum : index
   } : tensor<16x?xindex>

mlir/test/IR/core-ops.mlir

func @test_splat_op(%s : f32) {
   %t = splat %s : tensor<8xf32>
   // CHECK: splat [[S]] : tensor<8xf32>
   %u = "std.splat"(%s) : (f32) -> vector<4xf32>
   // CHECK: splat [[S]] : vector<4xf32>
   return
 }
 
 // CHECK-LABEL: func @tensor_load_store
-func @tensor_load_store(%0 : memref<4x4xi32>) {
-  // CHECK: %[[TENSOR:.*]] = memref.tensor_load %[[MEMREF:.*]] : memref<4x4xi32>
-  %1 = memref.tensor_load %0 : memref<4x4xi32>
+func @tensor_load_store(%0 : memref<4x4xi32>, %1 : tensor<4x4xi32>) {
+  // CHECK-SAME: (%[[MEMREF:.*]]: memref<4x4xi32>,
+  // CHECK-SAME:  %[[TENSOR:.*]]: tensor<4x4xi32>)
   // CHECK: memref.tensor_store %[[TENSOR]], %[[MEMREF]] : memref<4x4xi32>
   memref.tensor_store %1, %0 : memref<4x4xi32>
   return
 }
 
 // CHECK-LABEL: func @unranked_tensor_load_store
-func @unranked_tensor_load_store(%0 : memref<*xi32>) {
-  // CHECK: %[[TENSOR:.*]] = memref.tensor_load %[[MEMREF:.*]] : memref<*xi32>
-  %1 = memref.tensor_load %0 : memref<*xi32>
+func @unranked_tensor_load_store(%0 : memref<*xi32>, %1 : tensor<*xi32>) {
+  // CHECK-SAME: (%[[MEMREF:.*]]: memref<*xi32>,
+  // CHECK-SAME:  %[[TENSOR:.*]]: tensor<*xi32>)
   // CHECK: memref.tensor_store %[[TENSOR]], %[[MEMREF]] : memref<*xi32>
   memref.tensor_store %1, %0 : memref<*xi32>
   return
 }
 
 // CHECK-LABEL: func @atomic_rmw
 // CHECK-SAME: ([[BUF:%.*]]: memref<10xf32>, [[VAL:%.*]]: f32, [[I:%.*]]: index)
 func @atomic_rmw(%I: memref<10xf32>, %val: f32, %i : index) {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_cast.mlir

func @entry() {
     %6 = arith.constant dense<[ 4294967295.0, 4294967294.0, 4294967293.0, 4294967292.0,
                           0.0, 1.1, 2.2, 3.3, 4.4, 305.5 ]> : tensor<10xf64>
     %7 = sparse_tensor.convert %6 : tensor<10xf64> to tensor<10xf64, #SV>
 
     //
     // CHECK: ( -4, -3, -2, -1, 0, 1, 2, 3, 4, 305 )
     //
     %c0 = call @sparse_cast_s32_to_f32(%1) : (tensor<10xi32, #SV>) -> tensor<10xf32>
-    %m0 = memref.buffer_cast %c0 : memref<10xf32>
+    %m0 = bufferization.to_memref %c0 : memref<10xf32>
     %v0 = vector.transfer_read %m0[%z], %f: memref<10xf32>, vector<10xf32>
     vector.print %v0 : vector<10xf32>
 
     //
     // CHECK: ( 4.29497e+09, 4.29497e+09, 4.29497e+09, 4.29497e+09, 0, 1, 2, 3, 4, 305 )
     //
     %c1 = call @sparse_cast_u32_to_f32(%1) : (tensor<10xi32, #SV>) -> tensor<10xf32>
-    %m1 = memref.buffer_cast %c1 : memref<10xf32>
+    %m1 = bufferization.to_memref %c1 : memref<10xf32>
     %v1 = vector.transfer_read %m1[%z], %f: memref<10xf32>, vector<10xf32>
     vector.print %v1 : vector<10xf32>
 
     //
     // CHECK: ( -4, -3, -2, -1, 0, 1, 2, 3, 4, 305 )
     //
     %c2 = call @sparse_cast_f32_to_s32(%3) : (tensor<10xf32, #SV>) -> tensor<10xi32>
-    %m2 = memref.buffer_cast %c2 : memref<10xi32>
+    %m2 = bufferization.to_memref %c2 : memref<10xi32>
     %v2 = vector.transfer_read %m2[%z], %i: memref<10xi32>, vector<10xi32>
     vector.print %v2 : vector<10xi32>
 
     //
     // CHECK: ( 4294967295, 4294967294, 4294967293, 4294967292, 0, 1, 2, 3, 4, 305 )
     //
     %c3 = call @sparse_cast_f64_to_u32(%7) : (tensor<10xf64, #SV>) -> tensor<10xi32>
-    %m3 = memref.buffer_cast %c3 : memref<10xi32>
+    %m3 = bufferization.to_memref %c3 : memref<10xi32>
     %v3 = vector.transfer_read %m3[%z], %i: memref<10xi32>, vector<10xi32>
     %vu = vector.bitcast %v3 : vector<10xi32> to vector<10xui32>
     vector.print %vu : vector<10xui32>
 
     //
     // CHECK: ( -4.4, -3.3, -2.2, -1.1, 0, 1.1, 2.2, 3.3, 4.4, 305.5 )
     //
     %c4 = call @sparse_cast_f32_to_f64(%3) : (tensor<10xf32, #SV>) -> tensor<10xf64>
-    %m4 = memref.buffer_cast %c4 : memref<10xf64>
+    %m4 = bufferization.to_memref %c4 : memref<10xf64>
     %v4 = vector.transfer_read %m4[%z], %d: memref<10xf64>, vector<10xf64>
     vector.print %v4 : vector<10xf64>
 
     //
     // CHECK: ( -4.4, -3.3, -2.2, -1.1, 0, 1.1, 2.2, 3.3, 4.4, 305.5 )
     //
     %c5 = call @sparse_cast_f64_to_f32(%5) : (tensor<10xf64, #SV>) -> tensor<10xf32>
-    %m5 = memref.buffer_cast %c5 : memref<10xf32>
+    %m5 = bufferization.to_memref %c5 : memref<10xf32>
     %v5 = vector.transfer_read %m5[%z], %f: memref<10xf32>, vector<10xf32>
     vector.print %v5 : vector<10xf32>
 
     //
     // CHECK: ( -4, -3, -2, -1, 0, 1, 2, 3, 4, 305 )
     //
     %c6 = call @sparse_cast_s32_to_u64(%1) : (tensor<10xi32, #SV>) -> tensor<10xi64>
-    %m6 = memref.buffer_cast %c6 : memref<10xi64>
+    %m6 = bufferization.to_memref %c6 : memref<10xi64>
     %v6 = vector.transfer_read %m6[%z], %l: memref<10xi64>, vector<10xi64>
     vector.print %v6 : vector<10xi64>
 
     //
     // CHECK: ( 4294967292, 4294967293, 4294967294, 4294967295, 0, 1, 2, 3, 4, 305 )
     //
     %c7 = call @sparse_cast_u32_to_s64(%1) : (tensor<10xi32, #SV>) -> tensor<10xi64>
-    %m7 = memref.buffer_cast %c7 : memref<10xi64>
+    %m7 = bufferization.to_memref %c7 : memref<10xi64>
     %v7 = vector.transfer_read %m7[%z], %l: memref<10xi64>, vector<10xi64>
     vector.print %v7 : vector<10xi64>
 
     //
     // CHECK: ( -4, -3, -2, -1, 0, 1, 2, 3, 4, 49 )
     //
     %c8 = call @sparse_cast_i32_to_i8(%1) : (tensor<10xi32, #SV>) -> tensor<10xi8>
-    %m8 = memref.buffer_cast %c8 : memref<10xi8>
+    %m8 = bufferization.to_memref %c8 : memref<10xi8>
     %v8 = vector.transfer_read %m8[%z], %b: memref<10xi8>, vector<10xi8>
     vector.print %v8 : vector<10xi8>
 
     //
     // CHECK: ( -1064514355, -1068289229, -1072902963, -1081291571, 0, 1066192077, 1074580685, 1079194419, 1082969293, 1134084096 )
     //
     %c9 = call @sparse_cast_f32_as_s32(%3) : (tensor<10xf32, #SV>) -> tensor<10xi32>
-    %m9 = memref.buffer_cast %c9 : memref<10xi32>
+    %m9 = bufferization.to_memref %c9 : memref<10xi32>
     %v9 = vector.transfer_read %m9[%z], %i: memref<10xi32>, vector<10xi32>
     vector.print %v9 : vector<10xi32>
 
     // Release the resources.
     sparse_tensor.release %1 : tensor<10xi32, #SV>
     sparse_tensor.release %3 : tensor<10xf32, #SV>
     sparse_tensor.release %5 : tensor<10xf64, #SV>
     sparse_tensor.release %7 : tensor<10xf64, #SV>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_conversion_sparse2dense.mlir

func @dump(%arg0: tensor<2x3x4xf64>) {
     %c0 = arith.constant 0 : index
     %d0 = arith.constant -1.0 : f64
     %0 = vector.transfer_read %arg0[%c0, %c0, %c0], %d0: tensor<2x3x4xf64>, vector<2x3x4xf64>
     vector.print %0 : vector<2x3x4xf64>
     return
   }
   func @dumpAndRelease_234(%arg0: tensor<2x3x4xf64>) {
     call @dump(%arg0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<2x3x4xf64>
+    %1 = bufferization.to_memref %arg0 : memref<2x3x4xf64>
     memref.dealloc %1 : memref<2x3x4xf64>
     return
   }
   func @dumpAndRelease_p34(%arg0: tensor<?x3x4xf64>) {
     %0 = tensor.cast %arg0 : tensor<?x3x4xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<?x3x4xf64>
+    %1 = bufferization.to_memref %arg0 : memref<?x3x4xf64>
     memref.dealloc %1 : memref<?x3x4xf64>
     return
   }
   func @dumpAndRelease_2p4(%arg0: tensor<2x?x4xf64>) {
     %0 = tensor.cast %arg0 : tensor<2x?x4xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<2x?x4xf64>
+    %1 = bufferization.to_memref %arg0 : memref<2x?x4xf64>
     memref.dealloc %1 : memref<2x?x4xf64>
     return
   }
   func @dumpAndRelease_23p(%arg0: tensor<2x3x?xf64>) {
     %0 = tensor.cast %arg0 : tensor<2x3x?xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<2x3x?xf64>
+    %1 = bufferization.to_memref %arg0 : memref<2x3x?xf64>
     memref.dealloc %1 : memref<2x3x?xf64>
     return
   }
   func @dumpAndRelease_2pp(%arg0: tensor<2x?x?xf64>) {
     %0 = tensor.cast %arg0 : tensor<2x?x?xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<2x?x?xf64>
+    %1 = bufferization.to_memref %arg0 : memref<2x?x?xf64>
     memref.dealloc %1 : memref<2x?x?xf64>
     return
   }
   func @dumpAndRelease_p3p(%arg0: tensor<?x3x?xf64>) {
     %0 = tensor.cast %arg0 : tensor<?x3x?xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<?x3x?xf64>
+    %1 = bufferization.to_memref %arg0 : memref<?x3x?xf64>
     memref.dealloc %1 : memref<?x3x?xf64>
     return
   }
   func @dumpAndRelease_pp4(%arg0: tensor<?x?x4xf64>) {
     %0 = tensor.cast %arg0 : tensor<?x?x4xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<?x?x4xf64>
+    %1 = bufferization.to_memref %arg0 : memref<?x?x4xf64>
     memref.dealloc %1 : memref<?x?x4xf64>
     return
   }
   func @dumpAndRelease_ppp(%arg0: tensor<?x?x?xf64>) {
     %0 = tensor.cast %arg0 : tensor<?x?x?xf64> to tensor<2x3x4xf64>
     call @dump(%0) : (tensor<2x3x4xf64>) -> ()
-    %1 = memref.buffer_cast %arg0 : memref<?x?x?xf64>
+    %1 = bufferization.to_memref %arg0 : memref<?x?x?xf64>
     memref.dealloc %1 : memref<?x?x?xf64>
     return
   }
 
   //
   // Main driver.
   //
   func @entry() {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_filter_conv2d.mlir

func @entry() {
     //
     // CHECK:    ( ( 0, 0, -1, -6, -1, 6 ),
     // CHECK-SAME: ( -1, 0, 1, 0, 1, 0 ),
     // CHECK-SAME: ( 0, -1, 1, 0, 0, 0 ),
     // CHECK-SAME: ( -1, 0, 0, 0, 0, 0 ),
     // CHECK-SAME: ( 0, 0, 3, 6, -3, -6 ),
     // CHECK-SAME: ( 2, -1, 3, 0, -3, 0 ) )
     //
-    %m = memref.buffer_cast %0 : memref<6x6xi32>
+    %m = bufferization.to_memref %0 : memref<6x6xi32>
     %v = vector.transfer_read %m[%c0, %c0], %i0
       : memref<6x6xi32>, vector<6x6xi32>
     vector.print %v : vector<6x6xi32>
 
     // Release the resources.
     sparse_tensor.release %sparse_filter : tensor<3x3xi32, #DCSR>
     memref.dealloc %m : memref<6x6xi32>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_flatten.mlir

func @entry() {
 
     // Setup matrix memory that is initialized to zero.
     %xdata = memref.alloc() : memref<7x3xf64>
     scf.for %i = %c0 to %c7 step %c1 {
       scf.for %j = %c0 to %c3 step %c1 {
         memref.store %d0, %xdata[%i, %j] : memref<7x3xf64>
       }
     }
-    %x = memref.tensor_load %xdata : memref<7x3xf64>
+    %x = bufferization.to_tensor %xdata : memref<7x3xf64>
 
     // Read the sparse tensor from file, construct sparse storage.
     %fileName = call @getTensorFilename(%c0) : (index) -> (!Filename)
     %a = sparse_tensor.new %fileName : !Filename to tensor<7x3x3x3x3x3x5x3xf64, #SparseTensor>
 
     // Call the kernel.
     %0 = call @kernel_flatten(%a, %x)
       : (tensor<7x3x3x3x3x3x5x3xf64, #SparseTensor>, tensor<7x3xf64>) -> tensor<7x3xf64>
 
     // Print the result for verification.
     //
     // CHECK: ( 6.25, 0, 0 )
     // CHECK: ( 4.224, 6.21, 0 )
     // CHECK: ( 0, 0, 15.455 )
     // CHECK: ( 0, 0, 0 )
     // CHECK: ( 0, 0, 0 )
     // CHECK: ( 0, 0, 0 )
     // CHECK: ( 7, 0, 0 )
     //
-    %r = memref.buffer_cast %0 : memref<7x3xf64>
+    %r = bufferization.to_memref %0 : memref<7x3xf64>
     scf.for %i = %c0 to %c7 step %c1 {
       %v = vector.transfer_read %r[%i, %c0], %d0: memref<7x3xf64>, vector<3xf64>
       vector.print %v : vector<3xf64>
     }
 
     // Release the resources.
     memref.dealloc %xdata : memref<7x3xf64>
     sparse_tensor.release %a : tensor<7x3x3x3x3x3x5x3xf64, #SparseTensor>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matrix_ops.mlir

func @matrix_mul(%arga: tensor<?x?xf64, #DCSR>,
     return %0 : tensor<?x?xf64, #DCSR>
   }
 
   // Dump a sparse matrix.
   func @dump(%arg0: tensor<?x?xf64, #DCSR>) {
     %d0 = arith.constant 0.0 : f64
     %c0 = arith.constant 0 : index
     %dm = sparse_tensor.convert %arg0 : tensor<?x?xf64, #DCSR> to tensor<?x?xf64>
-    %0 = memref.buffer_cast %dm : memref<?x?xf64>
+    %0 = bufferization.to_memref %dm : memref<?x?xf64>
     %1 = vector.transfer_read %0[%c0, %c0], %d0: memref<?x?xf64>, vector<4x8xf64>
     vector.print %1 : vector<4x8xf64>
     memref.dealloc %0 : memref<?x?xf64>
     return
   }
 
   // Driver method to call and verify matrix kernels.
   func @entry() {

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_matvec.mlir

func @entry() {
     scf.for %i = %c0 to %c256 step %c1 {
       %k = arith.addi %i, %c1 : index
       %j = arith.index_cast %k : index to i32
       memref.store %j, %bdata[%i] : memref<?xi32>
     }
     scf.for %i = %c0 to %c4 step %c1 {
       memref.store %i0, %xdata[%i] : memref<?xi32>
     }
-    %b = memref.tensor_load %bdata : memref<?xi32>
-    %x = memref.tensor_load %xdata : memref<?xi32>
+    %b = bufferization.to_tensor %bdata : memref<?xi32>
+    %x = bufferization.to_tensor %xdata : memref<?xi32>
 
     // Call kernel.
     %0 = call @kernel_matvec(%a, %b, %x)
       : (tensor<?x?xi32, #SparseMatrix>, tensor<?xi32>, tensor<?xi32>) -> tensor<?xi32>
 
     // Print the result for verification.
     //
     // CHECK: ( 889, 1514, -21, -3431 )
     //
-    %m = memref.buffer_cast %0 : memref<?xi32>
+    %m = bufferization.to_memref %0 : memref<?xi32>
     %v = vector.transfer_read %m[%c0], %i0: memref<?xi32>, vector<4xi32>
     vector.print %v : vector<4xi32>
 
     // Release the resources.
     memref.dealloc %bdata : memref<?xi32>
     memref.dealloc %xdata : memref<?xi32>
     sparse_tensor.release %a : tensor<?x?xi32, #SparseMatrix>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_mttkrp.mlir

scf.for %i = %c0 to %c3 step %c1 {
       scf.for %j = %c0 to %c5 step %c1 {
         %k0 = arith.muli %i, %c5 : index
         %k1 = arith.addi %k0, %j : index
         %k2 = arith.index_cast %k1 : index to i32
         %k = arith.sitofp %k2 : i32 to f64
         memref.store %k, %cdata[%i, %j] : memref<?x?xf64>
       }
     }
-    %c = memref.tensor_load %cdata : memref<?x?xf64>
+    %c = bufferization.to_tensor %cdata : memref<?x?xf64>
 
     %ddata = memref.alloc(%c4, %c5) : memref<?x?xf64>
     scf.for %i = %c0 to %c4 step %c1 {
       scf.for %j = %c0 to %c5 step %c1 {
         %k0 = arith.muli %i, %c5 : index
         %k1 = arith.addi %k0, %j : index
         %k2 = arith.index_cast %k1 : index to i32
         %k = arith.sitofp %k2 : i32 to f64
         memref.store %k, %ddata[%i, %j] : memref<?x?xf64>
       }
     }
-    %d = memref.tensor_load %ddata : memref<?x?xf64>
+    %d = bufferization.to_tensor %ddata : memref<?x?xf64>
 
     %adata = memref.alloc(%c2, %c5) : memref<?x?xf64>
     scf.for %i = %c0 to %c2 step %c1 {
       scf.for %j = %c0 to %c5 step %c1 {
         memref.store %i0, %adata[%i, %j] : memref<?x?xf64>
       }
     }
-    %a = memref.tensor_load %adata : memref<?x?xf64>
+    %a = bufferization.to_tensor %adata : memref<?x?xf64>
 
     // Call kernel.
     %0 = call @kernel_mttkrp(%b, %c, %d, %a)
       : (tensor<?x?x?xf64, #SparseMatrix>,
         tensor<?x?xf64>, tensor<?x?xf64>, tensor<?x?xf64>) -> tensor<?x?xf64>
 
     // Print the result for verification.
     //
     // CHECK: ( ( 16075, 21930, 28505, 35800, 43815 ),
     // CHECK:   ( 10000, 14225, 19180, 24865, 31280 ) )
     //
-    %m = memref.buffer_cast %0 : memref<?x?xf64>
+    %m = bufferization.to_memref %0 : memref<?x?xf64>
     %v = vector.transfer_read %m[%c0, %c0], %i0
           : memref<?x?xf64>, vector<2x5xf64>
     vector.print %v : vector<2x5xf64>
 
     // Release the resources.
     memref.dealloc %adata : memref<?x?xf64>
     memref.dealloc %cdata : memref<?x?xf64>
     memref.dealloc %ddata : memref<?x?xf64>
     sparse_tensor.release %b : tensor<?x?x?xf64, #SparseMatrix>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_quantized_matmul.mlir

func @entry() {
     // Verify the output.
     //
     // CHECK:    ( ( -16510, 0, 16640, 12500, 1250, -390 ),
     // CHECK-SAME: ( -254, 0, 256, -200, -20, -6 ),
     // CHECK-SAME: ( 1143, 0, -1152, -200, -20, 27 ),
     // CHECK-SAME: ( -254, 0, 256, -300, -30, -6 ),
     // CHECK-SAME: ( 1397, 0, -1408, 100, 10, 33 ) )
     //
-    %m = memref.buffer_cast %0 : memref<5x6xi32>
+    %m = bufferization.to_memref %0 : memref<5x6xi32>
     %v = vector.transfer_read %m[%c0, %c0], %i0
       : memref<5x6xi32>, vector<5x6xi32>
     vector.print %v : vector<5x6xi32>
 
     // Release the resources.
     sparse_tensor.release %sparse_input2 : tensor<3x6xi8, #DCSR>
     memref.dealloc %m : memref<5x6xi32>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_reductions.mlir

func @entry() {
     // CHECK: 26
     // CHECK: 27.5
     // CHECK: 3087
     // CHECK: 168
     // CHECK: 1
     // CHECK: 15
     // CHECK: 10
     //
-    %m0 = memref.buffer_cast %0 : memref<i32>
+    %m0 = bufferization.to_memref %0 : memref<i32>
     call @dump_i32(%m0) : (memref<i32>) -> ()
-    %m1 = memref.buffer_cast %1 : memref<f32>
+    %m1 = bufferization.to_memref %1 : memref<f32>
     call @dump_f32(%m1) : (memref<f32>) -> ()
-    %m2 = memref.buffer_cast %2 : memref<i32>
+    %m2 = bufferization.to_memref %2 : memref<i32>
     call @dump_i32(%m2) : (memref<i32>) -> ()
-    %m3 = memref.buffer_cast %3 : memref<f32>
+    %m3 = bufferization.to_memref %3 : memref<f32>
     call @dump_f32(%m3) : (memref<f32>) -> ()
-    %m4 = memref.buffer_cast %4 : memref<i32>
+    %m4 = bufferization.to_memref %4 : memref<i32>
     call @dump_i32(%m4) : (memref<i32>) -> ()
-    %m5 = memref.buffer_cast %5 : memref<i32>
+    %m5 = bufferization.to_memref %5 : memref<i32>
     call @dump_i32(%m5) : (memref<i32>) -> ()
-    %m6 = memref.buffer_cast %6 : memref<i32>
+    %m6 = bufferization.to_memref %6 : memref<i32>
     call @dump_i32(%m6) : (memref<i32>) -> ()
 
     // Release the resources.
     sparse_tensor.release %sparse_input_i32 : tensor<32xi32, #SV>
     sparse_tensor.release %sparse_input_f32 : tensor<32xf32, #SV>
     sparse_tensor.release %dense_input_i32  : tensor<32xi32, #DV>
     sparse_tensor.release %dense_input_f32  : tensor<32xf32, #DV>
     memref.dealloc %m0 : memref<i32>

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_matmul.mlir

scf.for %i = %c0 to %c5 step %c1 {
       %p = arith.addi %i, %c1 : index
       %q = arith.index_cast %p : index to i32
       %d = arith.sitofp %q : i32 to f32
       scf.for %j = %c0 to %c10 step %c1 {
         memref.store %d, %adata[%i, %j] : memref<?x?xf32>
         memref.store %d, %bdata[%j, %i] : memref<?x?xf32>
       }
     }
-    %a = memref.tensor_load %adata : memref<?x?xf32>
-    %b = memref.tensor_load %bdata : memref<?x?xf32>
-    %x = memref.tensor_load %xdata : memref<?x?xf32>
+    %a = bufferization.to_tensor %adata : memref<?x?xf32>
+    %b = bufferization.to_tensor %bdata : memref<?x?xf32>
+    %x = bufferization.to_tensor %xdata : memref<?x?xf32>
 
     // Read the sparse matrix from file, construct sparse storage.
     %fileName = call @getTensorFilename(%c0) : (index) -> (!Filename)
     %s = sparse_tensor.new %fileName : !Filename to tensor<?x?xf32, #SparseMatrix>
 
     // Call the kernel.
     %0 = call @sampled_dense_dense(%s, %a, %b, %x)
        : (tensor<?x?xf32, #SparseMatrix>,
           tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>) -> tensor<?x?xf32>
 
     // Print the result for verification.
     //
     // CHECK: ( 10, 0, 0, 56, 0 )
     // CHECK: ( 0, 80, 0, 0, 250 )
     // CHECK: ( 0, 0, 270, 0, 0 )
     // CHECK: ( 164, 0, 0, 640, 0 )
     // CHECK: ( 0, 520, 0, 0, 1250 )
     //
-    %r = memref.buffer_cast %0 : memref<?x?xf32>
+    %r = bufferization.to_memref %0 : memref<?x?xf32>
     scf.for %i = %c0 to %c5 step %c1 {
       %v = vector.transfer_read %r[%i, %c0], %d0: memref<?x?xf32>, vector<5xf32>
       vector.print %v : vector<5xf32>
     }
 
     // Release the resources.
     memref.dealloc %adata : memref<?x?xf32>
     memref.dealloc %bdata : memref<?x?xf32>
     memref.dealloc %xdata : memref<?x?xf32>
     sparse_tensor.release %s : tensor<?x?xf32, #SparseMatrix>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sampled_mm_fusion.mlir

func @entry() {
     // CHECK-SAME: ( 0, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 0 ),
     // CHECK-SAME: ( 0, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 192 ) )
     //
     // CHECK:    ( ( 96, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 0 ),
     // CHECK-SAME: ( 0, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 0 ),
     // CHECK-SAME: ( 0, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 0 ),
     // CHECK-SAME: ( 0, 0, 0, 0, 0, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 0, 0, 192 ) )
     //
-    %m0 = memref.buffer_cast %0 : memref<8x8xf64>
-    %m1 = memref.buffer_cast %1 : memref<8x8xf64>
+    %m0 = bufferization.to_memref %0 : memref<8x8xf64>
+    %m1 = bufferization.to_memref %1 : memref<8x8xf64>
     %v0 = vector.transfer_read %m0[%c0, %c0], %d0
         : memref<8x8xf64>, vector<8x8xf64>
     %v1 = vector.transfer_read %m1[%c0, %c0], %d0
         : memref<8x8xf64>, vector<8x8xf64>
     vector.print %v0 : vector<8x8xf64>
     vector.print %v1 : vector<8x8xf64>
 
     // Release the resources.
     sparse_tensor.release %s : tensor<8x8xf64, #SM>
     memref.dealloc %m0 : memref<8x8xf64>
     memref.dealloc %m1 : memref<8x8xf64>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_spmm.mlir

scf.for %i = %c0 to %c256 step %c1 {
         memref.store %k, %bdata[%i, %j] : memref<?x?xf64>
       }
     }
     scf.for %i = %c0 to %c4 step %c1 {
       scf.for %j = %c0 to %c4 step %c1 {
         memref.store %i0, %xdata[%i, %j] : memref<?x?xf64>
       }
     }
-    %b = memref.tensor_load %bdata : memref<?x?xf64>
-    %x = memref.tensor_load %xdata : memref<?x?xf64>
+    %b = bufferization.to_tensor %bdata : memref<?x?xf64>
+    %x = bufferization.to_tensor %xdata : memref<?x?xf64>
 
     // Call kernel.
     %0 = call @kernel_spmm(%a, %b, %x)
       : (tensor<?x?xf64, #SparseMatrix>, tensor<?x?xf64>, tensor<?x?xf64>) -> tensor<?x?xf64>
 
     // Print the result for verification.
     //
     // CHECK: ( ( 3548, 3550, 3552, 3554 ), ( 6052, 6053, 6054, 6055 ), ( -56, -63, -70, -77 ), ( -13704, -13709, -13714, -13719 ) )
     //
-    %m = memref.buffer_cast %0 : memref<?x?xf64>
+    %m = bufferization.to_memref %0 : memref<?x?xf64>
     %v = vector.transfer_read %m[%c0, %c0], %i0: memref<?x?xf64>, vector<4x4xf64>
     vector.print %v : vector<4x4xf64>
 
     // Release the resources.
     memref.dealloc %bdata : memref<?x?xf64>
     memref.dealloc %xdata : memref<?x?xf64>
     sparse_tensor.release %a : tensor<?x?xf64, #SparseMatrix>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum.mlir

module {
   func @entry() {
     %d0 = arith.constant 0.0 : f64
     %c0 = arith.constant 0 : index
 
     // Setup memory for a single reduction scalar,
     // initialized to zero.
     %xdata = memref.alloc() : memref<f64>
     memref.store %d0, %xdata[] : memref<f64>
-    %x = memref.tensor_load %xdata : memref<f64>
+    %x = bufferization.to_tensor %xdata : memref<f64>
 
     // Read the sparse matrix from file, construct sparse storage.
     %fileName = call @getTensorFilename(%c0) : (index) -> (!Filename)
     %a = sparse_tensor.new %fileName : !Filename to tensor<?x?xf64, #SparseMatrix>
 
     // Call the kernel.
     %0 = call @kernel_sum_reduce(%a, %x)
       : (tensor<?x?xf64, #SparseMatrix>, tensor<f64>) -> tensor<f64>
 
     // Print the result for verification.
     //
     // CHECK: 30.2
     //
-    %m = memref.buffer_cast %0 : memref<f64>
+    %m = bufferization.to_memref %0 : memref<f64>
     %v = memref.load %m[] : memref<f64>
     vector.print %v : f64
 
     // Release the resources.
     memref.dealloc %xdata : memref<f64>
     sparse_tensor.release %a : tensor<?x?xf64, #SparseMatrix>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_vector_ops.mlir

func @dump(%arg0: tensor<?xf64, #SparseVector>) {
     // Dump the values array to verify only sparse contents are stored.
     %c0 = arith.constant 0 : index
     %d0 = arith.constant -1.0 : f64
     %0 = sparse_tensor.values %arg0 : tensor<?xf64, #SparseVector> to memref<?xf64>
     %1 = vector.transfer_read %0[%c0], %d0: memref<?xf64>, vector<16xf64>
     vector.print %1 : vector<16xf64>
     // Dump the dense vector to verify structure is correct.
     %dv = sparse_tensor.convert %arg0 : tensor<?xf64, #SparseVector> to tensor<?xf64>
-    %2 = memref.buffer_cast %dv : memref<?xf64>
+    %2 = bufferization.to_memref %dv : memref<?xf64>
     %3 = vector.transfer_read %2[%c0], %d0: memref<?xf64>, vector<32xf64>
     vector.print %3 : vector<32xf64>
     memref.dealloc %2 : memref<?xf64>
     return
   }
 
   // Driver method to call and verify vector kernels.
   func @entry() {
%v2 = arith.constant sparse<
          [11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0, 20.0 ]
     > : tensor<32xf64>
     %sv1 = sparse_tensor.convert %v1 : tensor<32xf64> to tensor<?xf64, #SparseVector>
     %sv2 = sparse_tensor.convert %v2 : tensor<32xf64> to tensor<?xf64, #SparseVector>
 
     // Setup memory for a single reduction scalar.
     %xdata = memref.alloc() : memref<f64>
     memref.store %d1, %xdata[] : memref<f64>
-    %x = memref.tensor_load %xdata : memref<f64>
+    %x = bufferization.to_tensor %xdata : memref<f64>
 
     // Call sparse vector kernels.
     %0 = call @vector_scale(%sv1)
        : (tensor<?xf64, #SparseVector>) -> tensor<?xf64, #SparseVector>
     %1 = call @vector_scale_inplace(%sv1)
        : (tensor<?xf64, #SparseVector>) -> tensor<?xf64, #SparseVector>
     %2 = call @vector_add(%sv1, %sv2)
        : (tensor<?xf64, #SparseVector>,
func @entry() {
     call @dump(%sv2) : (tensor<?xf64, #SparseVector>) -> ()
     call @dump(%0) : (tensor<?xf64, #SparseVector>) -> ()
     call @dump(%1) : (tensor<?xf64, #SparseVector>) -> ()
     call @dump(%2) : (tensor<?xf64, #SparseVector>) -> ()
     call @dump(%3) : (tensor<?xf64, #SparseVector>) -> ()
     %m4 = sparse_tensor.values %4 : tensor<?xf64, #DenseVector> to memref<?xf64>
     %v4 = vector.load %m4[%c0]: memref<?xf64>, vector<32xf64>
     vector.print %v4 : vector<32xf64>
-    %m5 = memref.buffer_cast %5 : memref<f64>
+    %m5 = bufferization.to_memref %5 : memref<f64>
     %v5 = memref.load %m5[] : memref<f64>
     vector.print %v5 : f64
 
     // Release the resources.
     sparse_tensor.release %sv1 : tensor<?xf64, #SparseVector>
     sparse_tensor.release %sv2 : tensor<?xf64, #SparseVector>
     sparse_tensor.release %0 : tensor<?xf64, #SparseVector>
     sparse_tensor.release %2 : tensor<?xf64, #SparseVector>
     sparse_tensor.release %3 : tensor<?xf64, #SparseVector>
     sparse_tensor.release %4 : tensor<?xf64, #DenseVector>
     memref.dealloc %xdata : memref<f64>
     return
   }
 }

mlir/test/Transforms/buffer-deallocation.mlir

^bb2:
   test.buffer_based in(%arg1: memref<2xf32>) out(%0: memref<2xf32>)
   br ^bb3(%0 : memref<2xf32>)
 ^bb3(%1: memref<2xf32>):
   test.copy(%1, %arg2) : (memref<2xf32>, memref<2xf32>)
   return
 }
 
 // CHECK-NEXT: cond_br
-//      CHECK: %[[ALLOC0:.*]] = memref.clone
+//      CHECK: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: br ^bb3(%[[ALLOC0]]
 //      CHECK: %[[ALLOC1:.*]] = memref.alloc
 // CHECK-NEXT: test.buffer_based
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone %[[ALLOC1]]
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone %[[ALLOC1]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 // CHECK-NEXT: br ^bb3(%[[ALLOC2]]
 //      CHECK: test.copy
 // CHECK-NEXT: memref.dealloc
 // CHECK-NEXT: return
 
 // -----
 
^bb2(%0: index):
   test.buffer_based in(%arg1: memref<?xf32>) out(%1: memref<?xf32>)
   br ^bb3(%1 : memref<?xf32>)
 ^bb3(%2: memref<?xf32>):
   test.copy(%2, %arg2) : (memref<?xf32>, memref<?xf32>)
   return
 }
 
 // CHECK-NEXT: cond_br
-//      CHECK: %[[ALLOC0:.*]] = memref.clone
+//      CHECK: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: br ^bb3(%[[ALLOC0]]
 //      CHECK: ^bb2(%[[IDX:.*]]:{{.*}})
 // CHECK-NEXT: %[[ALLOC1:.*]] = memref.alloc(%[[IDX]])
 // CHECK-NEXT: test.buffer_based
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 // CHECK-NEXT: br ^bb3
 // CHECK-NEXT: ^bb3(%[[ALLOC3:.*]]:{{.*}})
 //      CHECK: test.copy(%[[ALLOC3]],
 // CHECK-NEXT: memref.dealloc %[[ALLOC3]]
 // CHECK-NEXT: return
 
 // -----
^bb2(%0: index):
   test.buffer_based in(%arg1: memref<*xf32>) out(%2: memref<*xf32>)
   br ^bb3(%2 : memref<*xf32>)
 ^bb3(%3: memref<*xf32>):
   test.copy(%3, %arg2) : (memref<*xf32>, memref<*xf32>)
   return
 }
 
 // CHECK-NEXT: cond_br
-//      CHECK: %[[ALLOC0:.*]] = memref.clone
+//      CHECK: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: br ^bb3(%[[ALLOC0]]
 //      CHECK: ^bb2(%[[IDX:.*]]:{{.*}})
 // CHECK-NEXT: %[[ALLOC1:.*]] = memref.alloc(%[[IDX]])
 //      CHECK: test.buffer_based
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 // CHECK-NEXT: br ^bb3
 // CHECK-NEXT: ^bb3(%[[ALLOC3:.*]]:{{.*}})
 //      CHECK: test.copy(%[[ALLOC3]],
 // CHECK-NEXT: memref.dealloc %[[ALLOC3]]
 // CHECK-NEXT: return
 
 // -----
^bb6(%3: memref<?xf32>):
   br ^bb7(%3 : memref<?xf32>)
 ^bb7(%4: memref<?xf32>):
   test.copy(%4, %arg2) : (memref<?xf32>, memref<?xf32>)
   return
 }
 
 // CHECK-NEXT: cond_br{{.*}}
 // CHECK-NEXT: ^bb1
-// CHECK-NEXT: %[[ALLOC0:.*]] = memref.clone
+// CHECK-NEXT: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: br ^bb6(%[[ALLOC0]]
 //      CHECK: ^bb2(%[[IDX:.*]]:{{.*}})
 // CHECK-NEXT: %[[ALLOC1:.*]] = memref.alloc(%[[IDX]])
 // CHECK-NEXT: test.buffer_based
 //      CHECK: cond_br
 //      CHECK: ^bb3:
 // CHECK-NEXT: br ^bb5(%[[ALLOC1]]{{.*}})
 //      CHECK: ^bb4:
 // CHECK-NEXT: br ^bb5(%[[ALLOC1]]{{.*}})
 // CHECK-NEXT: ^bb5(%[[ALLOC2:.*]]:{{.*}})
-// CHECK-NEXT: %[[ALLOC3:.*]] = memref.clone %[[ALLOC2]]
+// CHECK-NEXT: %[[ALLOC3:.*]] = bufferization.clone %[[ALLOC2]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 // CHECK-NEXT: br ^bb6(%[[ALLOC3]]{{.*}})
 // CHECK-NEXT: ^bb6(%[[ALLOC4:.*]]:{{.*}})
 // CHECK-NEXT: br ^bb7(%[[ALLOC4]]{{.*}})
 // CHECK-NEXT: ^bb7(%[[ALLOC5:.*]]:{{.*}})
 //      CHECK: test.copy(%[[ALLOC5]],
 // CHECK-NEXT: memref.dealloc %[[ALLOC4]]
 // CHECK-NEXT: return
^bb1:
   %0 = memref.alloc() : memref<2xf32>
   test.buffer_based in(%arg1: memref<2xf32>) out(%0: memref<2xf32>)
   br ^bb2(%0 : memref<2xf32>)
 ^bb2(%1: memref<2xf32>):
   test.copy(%1, %arg2) : (memref<2xf32>, memref<2xf32>)
   return
 }
 
-// CHECK-NEXT: %[[ALLOC0:.*]] = memref.clone
+// CHECK-NEXT: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: cond_br
 //      CHECK: %[[ALLOC1:.*]] = memref.alloc()
 // CHECK-NEXT: test.buffer_based
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone %[[ALLOC1]]
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone %[[ALLOC1]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 //      CHECK: test.copy
 // CHECK-NEXT: memref.dealloc
 // CHECK-NEXT: return
 
 // -----
 
 // Test Case:
^exit(%arg2: memref<2xf32>):
   test.copy(%arg2, %arg1) : (memref<2xf32>, memref<2xf32>)
   return
 }
 
 // CHECK-NEXT: cond_br{{.*}}
 // CHECK-NEXT: ^bb1
 //      CHECK: %[[ALLOC0:.*]] = memref.alloc()
 // CHECK-NEXT: test.buffer_based
-// CHECK-NEXT: %[[ALLOC1:.*]] = memref.clone %[[ALLOC0]]
+// CHECK-NEXT: %[[ALLOC1:.*]] = bufferization.clone %[[ALLOC0]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC0]]
 // CHECK-NEXT: br ^bb3(%[[ALLOC1]]
 // CHECK-NEXT: ^bb2
 // CHECK-NEXT: %[[ALLOC2:.*]] = memref.alloc()
 // CHECK-NEXT: test.buffer_based
-// CHECK-NEXT: %[[ALLOC3:.*]] = memref.clone %[[ALLOC2]]
+// CHECK-NEXT: %[[ALLOC3:.*]] = bufferization.clone %[[ALLOC2]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC2]]
 // CHECK-NEXT: br ^bb3(%[[ALLOC3]]
 // CHECK-NEXT: ^bb3(%[[ALLOC4:.*]]:{{.*}})
 //      CHECK: test.copy
 // CHECK-NEXT: memref.dealloc %[[ALLOC4]]
 // CHECK-NEXT: return
 
 // -----
^bb2:
   }
   br ^bb3(%0 : memref<2xf32>)
 ^bb3(%1: memref<2xf32>):
   test.copy(%1, %arg2) : (memref<2xf32>, memref<2xf32>)
   return
 }
 //      CHECK: (%[[cond:.*]]: {{.*}}, %[[ARG1:.*]]: {{.*}}, %{{.*}}: {{.*}})
 // CHECK-NEXT:   cond_br %[[cond]], ^[[BB1:.*]], ^[[BB2:.*]]
-//      CHECK:   %[[ALLOC0:.*]] = memref.clone %[[ARG1]]
+//      CHECK:   %[[ALLOC0:.*]] = bufferization.clone %[[ARG1]]
 //      CHECK: ^[[BB2]]:
 //      CHECK:   %[[ALLOC1:.*]] = memref.alloc()
 // CHECK-NEXT:   test.region_buffer_based in(%[[ARG1]]{{.*}}out(%[[ALLOC1]]
 //      CHECK:     %[[ALLOC2:.*]] = memref.alloc()
 // CHECK-NEXT:     test.buffer_based in(%[[ARG1]]{{.*}}out(%[[ALLOC2]]
 //      CHECK:     memref.dealloc %[[ALLOC2]]
 // CHECK-NEXT:     %{{.*}} = math.exp
-//      CHECK:   %[[ALLOC3:.*]] = memref.clone %[[ALLOC1]]
+//      CHECK:   %[[ALLOC3:.*]] = bufferization.clone %[[ALLOC1]]
 // CHECK-NEXT:   memref.dealloc %[[ALLOC1]]
 //      CHECK:  ^[[BB3:.*]]({{.*}}):
 //      CHECK:  test.copy
 // CHECK-NEXT:  memref.dealloc
 
 // -----
 
 // Test Case: buffer deallocation escaping
%2 = scf.if %0 -> (memref<?x?xf32>) {
     %3 = memref.alloc(%arg0, %arg1) : memref<?x?xf32>
     scf.yield %3 : memref<?x?xf32>
   }
   return %2 : memref<?x?xf32>
 }
 
 //      CHECK: %[[ALLOC0:.*]] = memref.alloc(%arg0, %arg0)
 // CHECK-NEXT: %[[ALLOC1:.*]] = scf.if
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone %[[ALLOC0]]
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone %[[ALLOC0]]
 //      CHECK: scf.yield %[[ALLOC2]]
 //      CHECK: %[[ALLOC3:.*]] = memref.alloc(%arg0, %arg1)
-// CHECK-NEXT: %[[ALLOC4:.*]] = memref.clone %[[ALLOC3]]
+// CHECK-NEXT: %[[ALLOC4:.*]] = bufferization.clone %[[ALLOC3]]
 //      CHECK: memref.dealloc %[[ALLOC3]]
 //      CHECK: scf.yield %[[ALLOC4]]
 //      CHECK: memref.dealloc %[[ALLOC0]]
 // CHECK-NEXT: return %[[ALLOC1]]
 
 // -----
 
 // Test Case: nested region control flow within a region interface.
^bb2:
   br ^bb3(%0 : memref<2xf32>)
 ^bb3(%1: memref<2xf32>):
   test.copy(%1, %arg2) : (memref<2xf32>, memref<2xf32>)
   return
 }
 //      CHECK: (%[[cond:.*]]: {{.*}}, %[[ARG1:.*]]: {{.*}}, %{{.*}}: {{.*}})
 // CHECK-NEXT:   cond_br %[[cond]], ^[[BB1:.*]], ^[[BB2:.*]]
 //      CHECK: ^[[BB1]]:
-//      CHECK: %[[ALLOC0:.*]] = memref.clone
+//      CHECK: %[[ALLOC0:.*]] = bufferization.clone
 //      CHECK: ^[[BB2]]:
 //      CHECK:   %[[ALLOC1:.*]] = memref.alloc()
 // CHECK-NEXT:   test.region_buffer_based in(%[[ARG1]]{{.*}}out(%[[ALLOC1]]
 //      CHECK:     %[[ALLOCA:.*]] = memref.alloca()
 // CHECK-NEXT:     test.buffer_based in(%[[ARG1]]{{.*}}out(%[[ALLOCA]]
 //      CHECK:     %{{.*}} = math.exp
-//      CHECK:  %[[ALLOC2:.*]] = memref.clone %[[ALLOC1]]
+//      CHECK:  %[[ALLOC2:.*]] = bufferization.clone %[[ALLOC1]]
 // CHECK-NEXT:  memref.dealloc %[[ALLOC1]]
 //      CHECK:  ^[[BB3:.*]]({{.*}}):
 //      CHECK:  test.copy
 // CHECK-NEXT:  memref.dealloc
 
 // -----
 
 // CHECK-LABEL: func @nestedRegionControlFlowAlloca
%1 = scf.for %i = %lb to %ub step %step
     scf.yield %3 : memref<2xf32>
   }
   test.copy(%1, %res) : (memref<2xf32>, memref<2xf32>)
   return
 }
 
 //      CHECK: %[[ALLOC0:.*]] = memref.alloc()
 // CHECK-NEXT: memref.dealloc %[[ALLOC0]]
-// CHECK-NEXT: %[[ALLOC1:.*]] = memref.clone %arg3
+// CHECK-NEXT: %[[ALLOC1:.*]] = bufferization.clone %arg3
 //      CHECK: %[[ALLOC2:.*]] = scf.for {{.*}} iter_args
 // CHECK-SAME: (%[[IALLOC:.*]] = %[[ALLOC1]]
 //      CHECK:    arith.cmpi
 //      CHECK:    memref.dealloc %[[IALLOC]]
 //      CHECK:    %[[ALLOC3:.*]] = memref.alloc()
-//      CHECK:    %[[ALLOC4:.*]] = memref.clone %[[ALLOC3]]
+//      CHECK:    %[[ALLOC4:.*]] = bufferization.clone %[[ALLOC3]]
 //      CHECK:    memref.dealloc %[[ALLOC3]]
 //      CHECK:    scf.yield %[[ALLOC4]]
 //      CHECK: }
 //      CHECK: test.copy(%[[ALLOC2]], %arg4)
 // CHECK-NEXT: memref.dealloc %[[ALLOC2]]
 
 // -----
 
%3 = scf.if %2 -> (memref<2xf32>) {
       scf.yield %0 : memref<2xf32>
     }
     scf.yield %3 : memref<2xf32>
   }
   return %1 : memref<2xf32>
 }
 
 //      CHECK: %[[ALLOC0:.*]] = memref.alloc()
-// CHECK-NEXT: %[[ALLOC1:.*]] = memref.clone %arg3
+// CHECK-NEXT: %[[ALLOC1:.*]] = bufferization.clone %arg3
 // CHECK-NEXT: %[[ALLOC2:.*]] = scf.for {{.*}} iter_args
 // CHECK-SAME: (%[[IALLOC:.*]] = %[[ALLOC1]]
 //      CHECK: memref.dealloc %[[IALLOC]]
 //      CHECK: %[[ALLOC3:.*]] = scf.if
 
 //      CHECK: %[[ALLOC4:.*]] = memref.alloc()
-// CHECK-NEXT: %[[ALLOC5:.*]] = memref.clone %[[ALLOC4]]
+// CHECK-NEXT: %[[ALLOC5:.*]] = bufferization.clone %[[ALLOC4]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC4]]
 // CHECK-NEXT: scf.yield %[[ALLOC5]]
 
-//      CHECK: %[[ALLOC6:.*]] = memref.clone %[[ALLOC0]]
+//      CHECK: %[[ALLOC6:.*]] = bufferization.clone %[[ALLOC0]]
 // CHECK-NEXT: scf.yield %[[ALLOC6]]
 
-//      CHECK: %[[ALLOC7:.*]] = memref.clone %[[ALLOC3]]
+//      CHECK: %[[ALLOC7:.*]] = bufferization.clone %[[ALLOC3]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC3]]
 // CHECK-NEXT: scf.yield %[[ALLOC7]]
 
 //      CHECK: memref.dealloc %[[ALLOC0]]
 // CHECK-NEXT: return %[[ALLOC2]]
 
 // -----
 
%1 = scf.for %i = %lb to %ub step %step
     scf.yield %2 : memref<2xf32>
   }
   test.copy(%1, %res) : (memref<2xf32>, memref<2xf32>)
   return
 }
 
 //      CHECK: %[[ALLOC0:.*]] = memref.alloc()
 // CHECK-NEXT: memref.dealloc %[[ALLOC0]]
-// CHECK-NEXT: %[[ALLOC1:.*]] = memref.clone %arg3
+// CHECK-NEXT: %[[ALLOC1:.*]] = bufferization.clone %arg3
 // CHECK-NEXT: %[[VAL_7:.*]] = scf.for {{.*}} iter_args
 // CHECK-SAME: (%[[IALLOC0:.*]] = %[[ALLOC1]])
-// CHECK-NEXT: %[[ALLOC2:.*]] = memref.clone %[[IALLOC0]]
+// CHECK-NEXT: %[[ALLOC2:.*]] = bufferization.clone %[[IALLOC0]]
 // CHECK-NEXT: memref.dealloc %[[IALLOC0]]
 // CHECK-NEXT: %[[ALLOC3:.*]] = scf.for {{.*}} iter_args
 // CHECK-SAME: (%[[IALLOC1:.*]] = %[[ALLOC2]])
-// CHECK-NEXT: %[[ALLOC5:.*]] = memref.clone %[[IALLOC1]]
+// CHECK-NEXT: %[[ALLOC5:.*]] = bufferization.clone %[[IALLOC1]]
 // CHECK-NEXT: memref.dealloc %[[IALLOC1]]
 
 //      CHECK: %[[ALLOC6:.*]] = scf.for {{.*}} iter_args
 // CHECK-SAME: (%[[IALLOC2:.*]] = %[[ALLOC5]])
 //      CHECK: %[[ALLOC8:.*]] = memref.alloc()
 // CHECK-NEXT: memref.dealloc %[[ALLOC8]]
 //      CHECK: %[[ALLOC9:.*]] = scf.if
 
 //      CHECK: %[[ALLOC11:.*]] = memref.alloc()
-// CHECK-NEXT: %[[ALLOC12:.*]] = memref.clone %[[ALLOC11]]
+// CHECK-NEXT: %[[ALLOC12:.*]] = bufferization.clone %[[ALLOC11]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC11]]
 // CHECK-NEXT: scf.yield %[[ALLOC12]]
 
-//      CHECK: %[[ALLOC13:.*]] = memref.clone %[[IALLOC2]]
+//      CHECK: %[[ALLOC13:.*]] = bufferization.clone %[[IALLOC2]]
 // CHECK-NEXT: scf.yield %[[ALLOC13]]
 
 //      CHECK: memref.dealloc %[[IALLOC2]]
-// CHECK-NEXT: %[[ALLOC10:.*]] = memref.clone %[[ALLOC9]]
+// CHECK-NEXT: %[[ALLOC10:.*]] = bufferization.clone %[[ALLOC9]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC9]]
 // CHECK-NEXT: scf.yield %[[ALLOC10]]
 
-//      CHECK: %[[ALLOC7:.*]] = memref.clone %[[ALLOC6]]
+//      CHECK: %[[ALLOC7:.*]] = bufferization.clone %[[ALLOC6]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC6]]
 // CHECK-NEXT: scf.yield %[[ALLOC7]]
 
-//      CHECK: %[[ALLOC4:.*]] = memref.clone %[[ALLOC3]]
+//      CHECK: %[[ALLOC4:.*]] = bufferization.clone %[[ALLOC3]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC3]]
 // CHECK-NEXT: scf.yield %[[ALLOC4]]
 
 //      CHECK: test.copy(%[[VAL_7]], %arg4)
 // CHECK-NEXT: memref.dealloc %[[VAL_7]]
 
 // -----
 

 // CHECK-SAME: %[[ARG1:.*]]: {{.*}},
 // CHECK-SAME: %[[ARG2:.*]]: {{.*}}
 //      CHECK: %[[UNUSED_RESULT:.*]] = shape.assuming %[[ARG0]]
 // CHECK-NEXT:    %[[ALLOC0:.*]] = memref.alloc()
 // CHECK-NEXT:    memref.dealloc %[[ALLOC0]]
 // CHECK-NEXT:    shape.assuming_yield %[[ARG1]]
 //      CHECK: %[[ASSUMING_RESULT:.*]] = shape.assuming %[[ARG0]]
 // CHECK-NEXT:    %[[TMP_ALLOC:.*]] = memref.alloc()
-// CHECK-NEXT:    %[[RETURNING_ALLOC:.*]] = memref.clone %[[TMP_ALLOC]]
+// CHECK-NEXT:    %[[RETURNING_ALLOC:.*]] = bufferization.clone %[[TMP_ALLOC]]
 // CHECK-NEXT:    memref.dealloc %[[TMP_ALLOC]]
 // CHECK-NEXT:    shape.assuming_yield %[[RETURNING_ALLOC]]
 //      CHECK: test.copy(%[[ASSUMING_RESULT:.*]], %[[ARG2]])
 // CHECK-NEXT: memref.dealloc %[[ASSUMING_RESULT]]
 
 // -----
 
 // Test Case: The op "test.bar" does not implement the RegionBranchOpInterface.
// expected-error@+1 {{All operations with attached regions need to implement the RegionBranchOpInterface.}}
   }) : () -> (i32)
   "test.terminator"() : () -> ()
 }
 
 // -----
 
 // CHECK-LABEL: func @dealloc_existing_clones
 // CHECK: (%[[ARG0:.*]]: memref<?x?xf64>, %[[ARG1:.*]]: memref<?x?xf64>)
-// CHECK: %[[RES0:.*]] = memref.clone %[[ARG0]]
-// CHECK: %[[RES1:.*]] = memref.clone %[[ARG1]]
+// CHECK: %[[RES0:.*]] = bufferization.clone %[[ARG0]]
+// CHECK: %[[RES1:.*]] = bufferization.clone %[[ARG1]]
 // CHECK-NOT: memref.dealloc %[[RES0]]
 // CHECK: memref.dealloc %[[RES1]]
 // CHECK: return %[[RES0]]
 func @dealloc_existing_clones(%arg0: memref<?x?xf64>, %arg1: memref<?x?xf64>) -> memref<?x?xf64> {
-  %0 = memref.clone %arg0 : memref<?x?xf64> to memref<?x?xf64>
-  %1 = memref.clone %arg1 : memref<?x?xf64> to memref<?x?xf64>
+  %0 = bufferization.clone %arg0 : memref<?x?xf64> to memref<?x?xf64>
+  %1 = bufferization.clone %arg1 : memref<?x?xf64> to memref<?x?xf64>
   return %0 : memref<?x?xf64>
 }

mlir/test/Transforms/canonicalize.mlir

func @fold_trunci_sexti(%arg0: i1) -> i1 attributes {} {
   %0 = arith.extsi %arg0 : i1 to i8
   %1 = arith.trunci %0 : i8 to i1
   return %1 : i1
 }
 
 // CHECK-LABEL: func @simple_clone_elimination
 func @simple_clone_elimination() -> memref<5xf32> {
   %ret = memref.alloc() : memref<5xf32>
-  %temp = memref.clone %ret : memref<5xf32> to memref<5xf32>
+  %temp = bufferization.clone %ret : memref<5xf32> to memref<5xf32>
   memref.dealloc %temp : memref<5xf32>
   return %ret : memref<5xf32>
 }
 // CHECK-NEXT: %[[ret:.*]] = memref.alloc()
-// CHECK-NOT: %{{.*}} = memref.clone
+// CHECK-NOT: %{{.*}} = bufferization.clone
 // CHECK-NOT: memref.dealloc %{{.*}}
 // CHECK: return %[[ret]]
 
 // -----
 
 // CHECK-LABEL: func @clone_loop_alloc
 func @clone_loop_alloc(%arg0: index, %arg1: index, %arg2: index, %arg3: memref<2xf32>, %arg4: memref<2xf32>) {
   %0 = memref.alloc() : memref<2xf32>
   memref.dealloc %0 : memref<2xf32>
-  %1 = memref.clone %arg3 : memref<2xf32> to memref<2xf32>
+  %1 = bufferization.clone %arg3 : memref<2xf32> to memref<2xf32>
   %2 = scf.for %arg5 = %arg0 to %arg1 step %arg2 iter_args(%arg6 = %1) -> (memref<2xf32>) {
     %3 = arith.cmpi eq, %arg5, %arg1 : index
     memref.dealloc %arg6 : memref<2xf32>
     %4 = memref.alloc() : memref<2xf32>
-    %5 = memref.clone %4 : memref<2xf32> to memref<2xf32>
+    %5 = bufferization.clone %4 : memref<2xf32> to memref<2xf32>
     memref.dealloc %4 : memref<2xf32>
-    %6 = memref.clone %5 : memref<2xf32> to memref<2xf32>
+    %6 = bufferization.clone %5 : memref<2xf32> to memref<2xf32>
     memref.dealloc %5 : memref<2xf32>
     scf.yield %6 : memref<2xf32>
   }
   linalg.copy(%2, %arg4) : memref<2xf32>, memref<2xf32>
   memref.dealloc %2 : memref<2xf32>
   return
 }
 
-// CHECK-NEXT: %[[ALLOC0:.*]] = memref.clone
+// CHECK-NEXT: %[[ALLOC0:.*]] = bufferization.clone
 // CHECK-NEXT: %[[ALLOC1:.*]] = scf.for
 // CHECK-NEXT: memref.dealloc
 // CHECK-NEXT: %[[ALLOC2:.*]] = memref.alloc
 // CHECK-NEXT: scf.yield %[[ALLOC2]]
 // CHECK: linalg.copy(%[[ALLOC1]]
 // CHECK-NEXT: memref.dealloc %[[ALLOC1]]
 
 // -----
 
 // CHECK-LABEL: func @clone_nested_region
 func @clone_nested_region(%arg0: index, %arg1: index, %arg2: index) -> memref<?x?xf32> {
   %cmp = arith.cmpi eq, %arg0, %arg1 : index
   %0 = arith.cmpi eq, %arg0, %arg1 : index
   %1 = memref.alloc(%arg0, %arg0) : memref<?x?xf32>
   %2 = scf.if %0 -> (memref<?x?xf32>) {
     %3 = scf.if %cmp -> (memref<?x?xf32>) {
-      %9 = memref.clone %1 : memref<?x?xf32> to memref<?x?xf32>
+      %9 = bufferization.clone %1 : memref<?x?xf32> to memref<?x?xf32>
       scf.yield %9 : memref<?x?xf32>
     } else {
       %7 = memref.alloc(%arg0, %arg1) : memref<?x?xf32>
-      %10 = memref.clone %7 : memref<?x?xf32> to memref<?x?xf32>
+      %10 = bufferization.clone %7 : memref<?x?xf32> to memref<?x?xf32>
       memref.dealloc %7 : memref<?x?xf32>
       scf.yield %10 : memref<?x?xf32>
     }
-    %6 = memref.clone %3 : memref<?x?xf32> to memref<?x?xf32>
+    %6 = bufferization.clone %3 : memref<?x?xf32> to memref<?x?xf32>
     memref.dealloc %3 : memref<?x?xf32>
     scf.yield %6 : memref<?x?xf32>
   } else {
     %3 = memref.alloc(%arg1, %arg1) : memref<?x?xf32>
-    %6 = memref.clone %3 : memref<?x?xf32> to memref<?x?xf32>
+    %6 = bufferization.clone %3 : memref<?x?xf32> to memref<?x?xf32>
     memref.dealloc %3 : memref<?x?xf32>
     scf.yield %6 : memref<?x?xf32>
   }
   memref.dealloc %1 : memref<?x?xf32>
   return %2 : memref<?x?xf32>
 }
 
 //      CHECK: %[[ALLOC1:.*]] = memref.alloc
 // CHECK-NEXT: %[[ALLOC2:.*]] = scf.if
 // CHECK-NEXT: %[[ALLOC3_1:.*]] = scf.if
-// CHECK-NEXT: %[[ALLOC4_1:.*]] = memref.clone %[[ALLOC1]]
+// CHECK-NEXT: %[[ALLOC4_1:.*]] = bufferization.clone %[[ALLOC1]]
 // CHECK-NEXT: scf.yield %[[ALLOC4_1]]
 //      CHECK: %[[ALLOC4_2:.*]] = memref.alloc
 // CHECK-NEXT: scf.yield %[[ALLOC4_2]]
 //      CHECK: scf.yield %[[ALLOC3_1]]
 //      CHECK: %[[ALLOC3_2:.*]] = memref.alloc
 // CHECK-NEXT: scf.yield %[[ALLOC3_2]]
 //      CHECK: memref.dealloc %[[ALLOC1]]
 // CHECK-NEXT: return %[[ALLOC2]]

mlir/test/Transforms/finalizing-bufferize.mlir

 // RUN: mlir-opt %s -finalizing-bufferize -split-input-file -verify-diagnostics | FileCheck %s
 
 // CHECK-LABEL:   func @eliminate_materializations(
 // CHECK-SAME:                                     %[[ARG:.*]]: memref<f32>) -> memref<f32> {
 // CHECK:           return %[[ARG]] : memref<f32>
 func @eliminate_materializations(%arg0: memref<f32>) -> memref<f32> {
-  %0 = memref.tensor_load %arg0 : memref<f32>
-  %1 = memref.buffer_cast %0 : memref<f32>
+  %0 = bufferization.to_tensor %arg0 : memref<f32>
+  %1 = bufferization.to_memref %0 : memref<f32>
   return %1 : memref<f32>
 }
 
 // -----
 
 func @unable_to_convert_lone_buffer_cast() -> memref<f32> {
   // expected-error @+1 {{failed to legalize operation 'test.source'}}
   %0 = "test.source"() : () -> tensor<f32>
-  %1 = memref.buffer_cast %0 : memref<f32>
+  %1 = bufferization.to_memref %0 : memref<f32>
   return %1 : memref<f32>
 }
 
 // -----
 
 func @unable_to_convert_lone_tensor_load(%arg0: memref<f32>) {
-  %0 = memref.tensor_load %arg0 : memref<f32>
+  %0 = bufferization.to_tensor %arg0 : memref<f32>
   // expected-error @+1 {{failed to legalize operation 'test.sink'}}
   "test.sink"(%0) : (tensor<f32>) -> ()
   return
 }

mlir/test/lib/Dialect/Vector/TestVectorTransforms.cpp

StringRef getDescription() const final {
     return "Test optimization transformations for transfer ops";
   }
   void runOnFunction() override { transferOpflowOpt(getFunction()); }
 };
 
 struct TestVectorTransferLoweringPatterns
     : public PassWrapper<TestVectorTransferLoweringPatterns, FunctionPass> {
   void getDependentDialects(DialectRegistry &registry) const override {
-    registry.insert<memref::MemRefDialect>();
+    registry.insert<tensor::TensorDialect, memref::MemRefDialect>();
   }
   StringRef getArgument() const final {
     return "test-vector-transfer-lowering-patterns";
   }
   StringRef getDescription() const final {
     return "Test lowering patterns to lower transfer ops to other vector ops";
   }
   void runOnFunction() override {

mlir/test/mlir-opt/commandline.mlir

 // RUN: mlir-opt --show-dialects | FileCheck %s
 // CHECK: Available Dialects:
 // CHECK-NEXT: acc
 // CHECK-NEXT: affine
 // CHECK-NEXT: amx
 // CHECK-NEXT: arith
 // CHECK-NEXT: arm_neon
 // CHECK-NEXT: arm_sve
 // CHECK-NEXT: async
+// CHECK-NEXT: bufferization
 // CHECK-NEXT: builtin
 // CHECK-NEXT: complex
 // CHECK-NEXT: dlti
 // CHECK-NEXT: emitc
 // CHECK-NEXT: gpu
 // CHECK-NEXT: linalg
 // CHECK-NEXT: llvm
 // CHECK-NEXT: math

utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

hdrs = [
         "include/mlir/Dialect/SCF/Passes.h",
         "include/mlir/Dialect/SCF/Transforms.h",
     ],
     includes = ["include"],
     deps = [
         ":Affine",
         ":Analysis",
         ":ArithmeticDialect",
+        ":BufferizationDialect",
         ":DialectUtils",
         ":IR",
         ":MemRefDialect",
         ":Pass",
         ":SCFDialect",
         ":SCFPassIncGen",
         ":StandardOps",
         ":Support",
hdrs = [
         "include/mlir/Dialect/SparseTensor/Transforms/Passes.h",
         "include/mlir/ExecutionEngine/SparseTensorUtils.h",
     ],
     includes = ["include"],
     deps = [
         ":Affine",
         ":ArithmeticDialect",
         ":BufferizableOpInterface",
+        ":BufferizationDialect",
         ":IR",
         ":LLVMDialect",
         ":LinalgOps",
         ":LinalgTransforms",
         ":MemRefDialect",
         ":Pass",
         ":SCFDialect",
         ":SparseTensor",
deps = [
         ":AffineOpsIncGen",
         ":ArithmeticDialect",
         ":IR",
         ":LoopLikeInterface",
         ":MemRefDialect",
         ":SideEffectInterfaces",
         ":StandardOps",
         ":Support",
+        ":TensorDialect",
         "//llvm:Support",
     ],
 )
 
 cc_library(
     name = "EmitC",
     srcs = glob([
         "lib/Dialect/EmitC/IR/*.cpp",
srcs = glob(
         ],
     ),
     hdrs = glob([
         "include/mlir/Dialect/SCF/*.h",
     ]),
     includes = ["include"],
     deps = [
         ":ArithmeticDialect",
+        ":BufferizationDialect",
         ":ControlFlowInterfaces",
         ":IR",
         ":LoopLikeInterface",
         ":MemRefDialect",
         ":Pass",
         ":SCFIncGen",
         ":SCFPassIncGen",
         ":StandardOps",
deps = [
         ":Affine",
         ":ArithmeticDialect",
         ":DialectUtils",
         ":IR",
         ":InferTypeOpInterface",
         ":LinalgInterfacesIncGen",
         ":LinalgStructuredOpsIncGen",
         ":MemRefDialect",
+        ":TensorDialect",
         ":ViewLikeInterface",
         "//llvm:Support",
     ],
 )
 
 cc_library(
     name = "DataLayoutInterfaces",
     srcs = ["lib/Interfaces/DataLayoutInterfaces.cpp"],
cc_library(
     srcs = glob([
         "lib/Dialect/Shape/Transforms/*.cpp",
         "lib/Dialect/Shape/Transforms/*.h",
     ]),
     hdrs = ["include/mlir/Dialect/Shape/Transforms/Passes.h"],
     includes = ["include"],
     deps = [
         ":ArithmeticDialect",
+        ":BufferizationDialect",
         ":IR",
         ":MemRefDialect",
         ":Pass",
         ":Shape",
         ":ShapeTransformsPassIncGen",
         ":StandardOps",
         ":Transforms",
     ],
srcs = glob([
         "lib/Dialect/StandardOps/Transforms/*.h",
     ]),
     hdrs = glob(["include/mlir/Dialect/StandardOps/Transforms/*.h"]),
     includes = ["include"],
     deps = [
         ":Affine",
         ":ArithmeticDialect",
         ":ArithmeticTransforms",
+        ":BufferizationDialect",
         ":IR",
         ":MemRefDialect",  # TODO: Remove dependency on MemRef dialect
         ":Pass",
         ":SCFDialect",
         ":StandardOps",
         ":StandardOpsTransformsPassIncGen",
         ":Support",
         ":TensorDialect",
srcs = glob(
             "lib/Dialect/Tensor/Transforms/*.h",
         ],
     ),
     hdrs = ["include/mlir/Dialect/Tensor/Transforms/Passes.h"],
     includes = ["include"],
     deps = [
         ":ArithmeticDialect",
         ":Async",
+        ":BufferizationDialect",
         ":IR",
         ":MemRefDialect",
         ":ParallelLoopMapperAttrGen",
         ":Pass",
         ":SCFDialect",
         ":StandardOps",
         ":Support",
         ":TensorDialect",
cc_library(
     ]),
     hdrs = glob(["include/mlir/Transforms/*.h"]),
     includes = ["include"],
     deps = [
         ":Affine",
         ":AllocationOpInterface",
         ":Analysis",
         ":ArithmeticDialect",
+        ":BufferizationDialect",
         ":ControlFlowInterfaces",
         ":CopyOpInterface",
         ":IR",
         ":LinalgOps",
         ":LoopLikeInterface",
         ":MemRefDialect",
         ":Pass",
         ":Rewrite",
hdrs = glob(
             "include/mlir/Analysis/*/*.h",
         ],
         exclude = ["include/mlir/Analysis/Vector*.h"],
     ),
     includes = ["include"],
     deps = [
         ":Affine",
         ":ArithmeticDialect",
+        ":BufferizationDialect",
         ":CallOpInterfaces",
         ":ControlFlowInterfaces",
         ":DataLayoutInterfaces",
         ":IR",
         ":LinalgOps",
         ":SCFDialect",
         ":SideEffectInterfaces",
         ":StandardOps",
deps = [
         ":ArithmeticTransforms",
         ":ArmNeon",
         ":ArmSVE",
         ":ArmSVETransforms",
         ":Async",
         ":AsyncPassIncGen",
         ":AsyncToLLVM",
         ":AsyncTransforms",
+        ":BufferizationDialect",
         ":ComplexDialect",
         ":ComplexToLLVM",
         ":ConversionPasses",
         ":DLTIDialect",
         ":EmitC",
         ":GPUDialect",
         ":GPUPassIncGen",
         ":GPUToGPURuntimeTransforms",
srcs = [
         "lib/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.cpp",
     ],
     hdrs = [
         "include/mlir/Dialect/Linalg/ComprehensiveBufferize/BufferizableOpInterface.h",
     ],
     includes = ["include"],
     deps = [
         ":BufferizableOpInterfaceIncGen",
+        ":BufferizationDialect",
         ":IR",
         ":MemRefDialect",
         ":Support",
         "//llvm:Support",
     ],
 )
 
 cc_library(
cc_library(
     ],
     includes = ["include"],
     deps = [
         ":Affine",
         ":AffineUtils",
         ":Analysis",
         ":ArithmeticDialect",
         ":BufferizableOpInterface",
+        ":BufferizationDialect",
         ":ComplexDialect",
         ":ComprehensiveBufferize",
         ":DialectUtils",
         ":IR",
         ":InferTypeOpInterface",
         ":LinalgBufferizableOpInterfaceImpl",
         ":LinalgOps",
         ":LinalgPassIncGen",
cc_library(
     hdrs = [
         "include/mlir/Dialect/Linalg/ComprehensiveBufferize/ComprehensiveBufferize.h",
     ],
     includes = ["include"],
     deps = [
         ":Affine",
         ":ArithmeticDialect",
         ":BufferizableOpInterface",
+        ":BufferizationDialect",
         ":DialectUtils",
         ":IR",
         ":InferTypeOpInterface",
         ":MemRefDialect",
         ":Pass",
         ":SCFDialect",
         ":StandardOps",
         ":Support",
srcs = glob([
         "lib/Dialect/Arithmetic/Transforms/*.cpp",
         "lib/Dialect/Arithmetic/Transforms/*.h",
     ]),
     hdrs = ["include/mlir/Dialect/Arithmetic/Transforms/Passes.h"],
     includes = ["include"],
     deps = [
         ":ArithmeticDialect",
         ":ArithmeticPassIncGen",
+        ":BufferizationDialect",
         ":IR",
         ":MemRefDialect",
         ":Pass",
         ":Transforms",
     ],
 )
 
 td_library(
deps = [
         ":ControlFlowInterfaces",
         ":CopyOpInterface",
         ":DialectUtils",
         ":IR",
         ":InferTypeOpInterface",
         ":MemRefBaseIncGen",
         ":MemRefOpsIncGen",
         ":StandardOps",
-        ":TensorDialect",
         ":ViewLikeInterface",
         "//llvm:Support",
     ],
 )
 
 gentbl_cc_library(
     name = "MemRefPassIncGen",
     strip_include_prefix = "include",

 cc_library(
     name = "AllocationOpInterface",
     srcs = ["lib/Dialect/Bufferization/IR/AllocationOpInterface.cpp"],
     hdrs = ["include/mlir/Dialect/Bufferization/IR/AllocationOpInterface.h"],
     includes = ["include"],
     deps = [
         ":AllocationOpInterfaceIncGen",
         ":IR",
+    ],
+)
+
+td_library(
+    name = "BufferizationOpsTdFiles",
+    srcs = [
+        "include/mlir/Dialect/Bufferization/IR/BufferizationBase.td",
+        "include/mlir/Dialect/Bufferization/IR/BufferizationOps.td",
+    ],
+    includes = ["include"],
+    deps = [
+        ":AllocationOpInterfaceTdFiles",
+        ":CopyOpInterfaceTdFiles",
+        ":OpBaseTdFiles",
+        ":SideEffectInterfacesTdFiles",
+    ],
+)
+
+gentbl_cc_library(
+    name = "BufferizationBaseIncGen",
+    strip_include_prefix = "include",
+    tbl_outs = [
+        (
+            [
+                "-gen-dialect-decls",
+                "-dialect=bufferization",
+            ],
+            "include/mlir/Dialect/Bufferization/IR/BufferizationOpsDialect.h.inc",
+        ),
+        (
+            [
+                "-gen-dialect-defs",
+                "-dialect=bufferization",
+            ],
+            "include/mlir/Dialect/Bufferization/IR/BufferizationOpsDialect.cpp.inc",
+        ),
+    ],
+    tblgen = ":mlir-tblgen",
+    td_file = "include/mlir/Dialect/Bufferization/IR/BufferizationBase.td",
+    deps = [":BufferizationOpsTdFiles"],
+)
+
+gentbl_cc_library(
+    name = "BufferizationOpsIncGen",
+    strip_include_prefix = "include",
+    tbl_outs = [
+        (
+            ["-gen-op-decls"],
+            "include/mlir/Dialect/Bufferization/IR/BufferizationOps.h.inc",
+        ),
+        (
+            ["-gen-op-defs"],
+            "include/mlir/Dialect/Bufferization/IR/BufferizationOps.cpp.inc",
+        ),
+    ],
+    tblgen = ":mlir-tblgen",
+    td_file = "include/mlir/Dialect/Bufferization/IR/BufferizationOps.td",
+    deps = [":BufferizationOpsTdFiles"],
+)
+
+cc_library(
+    name = "BufferizationDialect",
+    srcs = glob(
+        [
+            "lib/Dialect/Bufferization/IR/Bufferization*.h",
+            "lib/Dialect/Bufferization/IR/Bufferization*.cpp",
+        ],
+    ),
+    hdrs = ["include/mlir/Dialect/Bufferization/IR/Bufferization.h"],
+    includes = ["include"],
+    deps = [
+        ":BufferizationBaseIncGen",
+        ":BufferizationOpsIncGen",
+        ":ControlFlowInterfaces",
+        ":CopyOpInterface",
+        ":AllocationOpInterface",
+        #":DialectUtils",
+        ":IR",
+        ":InferTypeOpInterface",
+        ":StandardOps",
+        ":TensorDialect",
         ":MemRefDialect",
+        ":ViewLikeInterface",
+        "//llvm:Support",
     ],
 )
 
 td_library(
     name = "DLTIDialectTdFiles",
     srcs = [
         "include/mlir/Dialect/DLTI/DLTI.td",
         "include/mlir/Dialect/DLTI/DLTIBase.td",