[mlir][sparse] migrate sparse operations into new sparse tensor dialect

[aartbik]

Summary

This is the very first step toward removing the glue and clutter from linalg and
replace it with proper sparse tensor types. This revision migrates the LinalgSparseOps
into SparseTensorOps of a sparse tensor dialect. This also provides a new home for
sparse tensor related transformation.

NOTE: the actual replacement with sparse tensor types (and removal of linalg glue/clutter) will follow but I am trying to keep the amount of changes per revision manageable.

[rriddle]

mlir/include/mlir/Dialect/Sparse/IR/Sparse.h

1 ↗

(On Diff #341315)

Why not SparseTensor dialect?

[aartbik]

mlir/include/mlir/Dialect/Sparse/IR/Sparse.h

1 ↗

(On Diff #341315)

do you prefer that? SGMT...

[rriddle]

mlir/include/mlir/Dialect/Sparse/IR/Sparse.h

1 ↗

(On Diff #341315)

Just seems like a more natural fit, given that this is supposed to be a sister of the Tensor dialect right? Or does this go beyond tensors?

[aartbik]

mlir/include/mlir/Dialect/Sparse/IR/Sparse.h

1 ↗

(On Diff #341315)

You are absolutely right. I spend too much time referring to the stuff as just "sparse", but it really should be sparse tensors. Coming up!

[aartbik]

renamed sparse -> sparse tensor

[aartbik]

lint

[rriddle]

Does Linalg need a dependency on the SparseTensor dialect? It would seem so from reading the code (or at least, some of the passes do).

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td

23–24

Is there more you can add here?

[aartbik]

In D101488#2726620, @rriddle wrote:

Does Linalg need a dependency on the SparseTensor dialect? It would seem so from reading the code (or at least, some of the passes do).

Did I miss something? Note that the "pass" right now lives in testing only (and will migrate to SparseTensorDialect after that).

Nicolas kindly provided me a temporary home in "Linalg", but I overstayed my welcome, so it is time to get my own home ;-)

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td

23–24

next CL will remove this op completely, hence the TODO
(in a nutshell, this was "glue" to bring a void pointer back to a tensor type; with the new sparse tensor type, this will be properly typed \O/)

[rriddle]

In D101488#2726646, @aartbik wrote:

In D101488#2726620, @rriddle wrote:

Does Linalg need a dependency on the SparseTensor dialect? It would seem so from reading the code (or at least, some of the passes do).

Did I miss something? Note that the "pass" right now lives in testing only (and will migrate to SparseTensorDialect after that).

Nicolas kindly provided me a temporary home in "Linalg", but I overstayed my welcome, so it is time to get my own home ;-)

It seems that Linalg/Transforms/Sparsification.cpp is creating operations in the SparseTensor dialect. I don't know where that is used to know where the dependency declaration would go, I just saw the Linalg/ directory creating SparseTensor ops. If that is just in the test pass, I suppose no dependency is necessary.

[aartbik]

In D101488#2726655, @rriddle wrote:

In D101488#2726646, @aartbik wrote:

In D101488#2726620, @rriddle wrote:

Does Linalg need a dependency on the SparseTensor dialect? It would seem so from reading the code (or at least, some of the passes do).

Did I miss something? Note that the "pass" right now lives in testing only (and will migrate to SparseTensorDialect after that).

Nicolas kindly provided me a temporary home in "Linalg", but I overstayed my welcome, so it is time to get my own home ;-)

It seems that Linalg/Transforms/Sparsification.cpp is creating operations in the SparseTensor dialect. I don't know where that is used to know where the dependency declaration would go, I just saw the Linalg/ directory creating SparseTensor ops. If that is just in the test pass, I suppose no dependency is necessary.

Yes, apologies for the shuffling (you will see one ore two more...). With a new permanent home, things can finally move into the right place (and stay there ;-)

[mehdi_amini]

Sorry I had to revert in 086e0f05bfc2 because this broke the build with -DBUILD_SHARED_LIBS=ON

[aartbik]

In D101488#2726835, @mehdi_amini wrote:

Sorry I had to revert in 086e0f05bfc2 because this broke the build with -DBUILD_SHARED_LIBS=ON

I was just constructing a fix for that ;-(

mlir/include/mlir/Dialect/CMakeLists.txt

 add_subdirectory(MemRef)
 add_subdirectory(OpenACC)
 add_subdirectory(OpenMP)
 add_subdirectory(PDL)
 add_subdirectory(PDLInterp)
 add_subdirectory(Quant)
 add_subdirectory(SCF)
 add_subdirectory(Shape)
+add_subdirectory(SparseTensor)
 add_subdirectory(SPIRV)
 add_subdirectory(StandardOps)
 add_subdirectory(Tensor)
 add_subdirectory(Tosa)
 add_subdirectory(Vector)
 add_subdirectory(X86Vector)

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

 
 set(LLVM_TARGET_DEFINITIONS LinalgStructuredOps.td)
 mlir_tablegen(LinalgStructuredOps.h.inc -gen-op-decls)
 mlir_tablegen(LinalgStructuredOps.cpp.inc -gen-op-defs)
 add_public_tablegen_target(MLIRLinalgStructuredOpsIncGen)
 add_dependencies(MLIRLinalgStructuredOpsIncGen LinalgOdsGen)
 add_dependencies(mlir-headers MLIRLinalgStructuredOpsIncGen)
 
-set(LLVM_TARGET_DEFINITIONS LinalgSparseOps.td)
-mlir_tablegen(LinalgSparseOps.h.inc -gen-op-decls)
-mlir_tablegen(LinalgSparseOps.cpp.inc -gen-op-defs)
-add_public_tablegen_target(MLIRLinalgSparseOpsIncGen)
-add_dependencies(mlir-headers MLIRLinalgSparseOpsIncGen)
-
 set(LLVM_TARGET_DEFINITIONS LinalgInterfaces.td)
 mlir_tablegen(LinalgInterfaces.h.inc -gen-op-interface-decls)
 mlir_tablegen(LinalgInterfaces.cpp.inc -gen-op-interface-defs)
 add_public_tablegen_target(MLIRLinalgInterfacesIncGen)
 add_dependencies(mlir-headers MLIRLinalgInterfacesIncGen)

mlir/include/mlir/Dialect/Linalg/IR/LinalgOps.h

 #include "mlir/Dialect/Linalg/IR/LinalgInterfaces.h"
 
 #define GET_OP_CLASSES
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h.inc"
 
 #define GET_OP_CLASSES
 #include "mlir/Dialect/Linalg/IR/LinalgStructuredOps.h.inc"
 
-#define GET_OP_CLASSES
-#include "mlir/Dialect/Linalg/IR/LinalgSparseOps.h.inc"
-
 #endif // MLIR_DIALECT_LINALG_LINALGOPS_H_

mlir/include/mlir/Dialect/Linalg/Transforms/Transforms.h

struct SparsificationOptions {
   bool fastOutput; // experimental: fast output buffers
 };
 
 /// Sets up sparsification rewriting rules with the given options.
 void populateSparsificationPatterns(
     RewritePatternSet &patterns,
     const SparsificationOptions &options = SparsificationOptions());
 
-/// Sets up sparsification conversion rules with the given options.
-void populateSparsificationConversionPatterns(RewritePatternSet &patterns);
-
 } // namespace linalg
 } // namespace mlir
 
 #endif // DIALECT_LINALG_TRANSFORMS_TRANSFORMS_H_

mlir/include/mlir/Dialect/SparseTensor/CMakeLists.txt

+add_subdirectory(IR)

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

+add_mlir_dialect(SparseTensorOps sparse_tensor)
+add_mlir_doc(SparseTensorOps SparseTensorOps Dialects/ -gen-dialect-doc)

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensor.h

+//===- SparseTensor.h - SparseTensor 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_SPARSETENSOR_IR_SPARSE_H_
+#define MLIR_DIALECT_SPARSETENSOR_IR_SPARSE_H_
+
+#include "mlir/IR/BuiltinTypes.h"
+#include "mlir/IR/Dialect.h"
+#include "mlir/IR/OpDefinition.h"
+#include "mlir/IR/OpImplementation.h"
+#include "mlir/Interfaces/SideEffectInterfaces.h"
+
+#define GET_OP_CLASSES
+#include "mlir/Dialect/SparseTensor/IR/SparseTensorOps.h.inc"
+
+#include "mlir/Dialect/SparseTensor/IR/SparseTensorOpsDialect.h.inc"
+
+#endif // MLIR_DIALECT_SPARSETENSOR_IR_SPARSE_H_

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorBase.td

+//===- SparseTensorBase.td - Sparse tensor 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 SPARSETENSOR_BASE
+#define SPARSETENSOR_BASE
+
+include "mlir/IR/OpBase.td"
+
+def SparseTensor_Dialect : Dialect {
+  let name = "sparse_tensor";
+  let cppNamespace = "::mlir::sparse_tensor";
+  let description = [{
+    The `sparse tensor` dialect is intended to hold primitives that
+    form a bridge between high-level operations on sparse tensors
+    and lower-level operations on the actual sparse storage schemes
+    consisting of pointers, indices, and values. This bridge
+    simplifies a `sparse compiler` pass by postponing actual
+    code generation for the supported primitives to a later phase,
+    either by generating calls into a runtime support library
+    or by further lowering the primitives into actual code.
+  }];
+}
+
+#endif // SPARSETENSOR_BASE

mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td

-//===- LinalgSparseOps.td - Linalg dialect sparse ops ------*- tablegen -*-===//
+//===- SparseTensorOps.td - SparseTensor dialect ops -------*- 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
 //
 //===----------------------------------------------------------------------===//
-//
-// The following operations bootstrap working with sparse tensors solely
-// within the Linalg dialect. They provide temporary bridges between a
-// future SparseTensorType (now an opaque pointer), the actual TensorType,
-// and MemRef arrays underlying an actual sparse storage scheme in memory.
-//
-// Lacking a proper sparse tensor type, the 'sparse_tensor' operation
-// provides a bridge between an opaque pointer and a regular tensor type
-// just to simplify feeding the value into a Linalg op. The operation
-// simply disappears during lowering.
-//
-// The other operations form the bridge between the opaque pointer and
-// the actual storage of pointers, indices, and values. These operations
-// resemble 'buffer_cast' in the sense that they map tensors to
-// their bufferized memrefs, but they lower into actual calls since
-// sparse storage does not bufferize into a single memrefs, as dense
-// tensors do, but into a hierarchical storage scheme where pointers
-// access memrefs with indices and eventually into values.
-//
-// TODO: introduce SparseTensorType as first class citizen in MLIR
-//
-//===----------------------------------------------------------------------===//
 
-#ifndef LINALG_SPARSE_OPS
-#define LINALG_SPARSE_OPS
+#ifndef SPARSETENSOR_OPS
+#define SPARSETENSOR_OPS
 
-include "mlir/Dialect/Linalg/IR/LinalgBase.td"
+include "mlir/Dialect/SparseTensor/IR/SparseTensorBase.td"
 include "mlir/Interfaces/SideEffectInterfaces.td"
 
 // Base class.
-class Linalg_SparseOp<string mnemonic, list<OpTrait> traits = []>
-  : Op<Linalg_Dialect, mnemonic, traits> {
+class SparseTensor_Op<string mnemonic, list<OpTrait> traits = []>
+  : Op<SparseTensor_Dialect, mnemonic, traits> {
   let printer = [{ return ::print(p, *this); }];
   let verifier = ?;
   let parser = [{ return ::parse$cppClass(parser, result); }];
 }
 
-def Linalg_SparseTensorFromPointerOp :
-    Linalg_SparseOp<"sparse_tensor">,
+// TODO: remove me
+def SparseTensor_FromPointerOp : SparseTensor_Op<"fromPtr">,
     Arguments<(ins AnyType:$ptr)>,
     Results<(outs AnyTensor:$result)> {
   let summary = "Views an opaque sparse tensor pointer as a tensor";
   let description = [{
      Lacking a first class citizen type for sparse tensors, this operation
      forms the glue between a sparse storage scheme (behind an opaque
      pointer) and the (dense) tensors used in the kernel definitions.
      This operation merely provides a way to assign a proper tensor
      type and shape to the incoming opaque pointer. It disappears
      completely during lowering.
 
      Example:
 
     ```mlir
      !SparseTensor = type !llvm.ptr<i8>
 
-     %0 = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<64x64xf64>
+     %0 = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<64x64xf64>
     ```
   }];
   let assemblyFormat = "$ptr attr-dict `:` type($ptr) `to` type($result)";
 }
 
-def Linalg_SparseTensorToPointersMemRefOp :
-    Linalg_SparseOp<"sparse_pointers", [NoSideEffect]>,
+def SparseTensor_ToPointersOp : SparseTensor_Op<"pointers", [NoSideEffect]>,
     Arguments<(ins AnyTensor:$tensor, Index:$dim)>,
     Results<(outs AnyStridedMemRefOfRank<1>:$result)> {
   let summary = "Extract pointers array at given dimension from a tensor";
   let description = [{
      Returns the pointers array of the sparse storage scheme at the
      given dimension for the given tensor. This is similar to the
      `buffer_cast` operation in the sense that it provides a bridge
      between a tensor world view and a bufferized world view. Unlike the
      `buffer_cast` operation, however, this sparse operation actually
      lowers into a call into a support library to obtain access to the
      pointers array.
 
      Example:
 
     ```mlir
-    %1 = linalg.sparse_pointers %0, %c1 : tensor<64x64xf64> to memref<?xindex>
+    %1 = sparse_tensor.pointers %0, %c1 : tensor<64x64xf64> to memref<?xindex>
     ```
   }];
   let assemblyFormat = "$tensor `,` $dim attr-dict `:` type($tensor)"
       " `to` type($result)";
 }
 
-def Linalg_SparseTensorToIndicesMemRefOp :
-    Linalg_SparseOp<"sparse_indices", [NoSideEffect]>,
+def SparseTensor_ToIndicesOp : SparseTensor_Op<"indices", [NoSideEffect]>,
     Arguments<(ins AnyTensor:$tensor, Index:$dim)>,
     Results<(outs AnyStridedMemRefOfRank<1>:$result)> {
   let summary = "Extract indices array at given dimension from a tensor";
   let description = [{
      Returns the indices array of the sparse storage scheme at the
      given dimension for the given tensor. This is similar to the
      `buffer_cast` operation in the sense that it provides a bridge
      between a tensor world view and a bufferized world view. Unlike the
      `buffer_cast` operation, however, this sparse operation actually
      lowers into a call into a support library to obtain access to the
      indices array.
 
      Example:
 
     ```mlir
-    %1 = linalg.sparse_indices %0, %c1 : tensor<64x64xf64> to memref<?xindex>
+    %1 = sparse_tensor.indices %0, %c1 : tensor<64x64xf64> to memref<?xindex>
     ```
   }];
   let assemblyFormat = "$tensor `,` $dim attr-dict `:` type($tensor)"
       " `to` type($result)";
 }
 
-def Linalg_SparseTensorToValuesMemRefOp :
-    Linalg_SparseOp<"sparse_values", [NoSideEffect]>,
+def SparseTensor_ToValuesOp : SparseTensor_Op<"values", [NoSideEffect]>,
     Arguments<(ins AnyTensor:$tensor)>,
     Results<(outs AnyStridedMemRefOfRank<1>:$result)> {
   let summary = "Extract numerical values array from a tensor";
   let description = [{
      Returns the values array of the sparse storage scheme for the given
      tensor, independent of the actual dimension. This is similar to the
      `buffer_cast` operation in the sense that it provides a bridge
      between a tensor world view and a bufferized world view. Unlike the
      `buffer_cast` operation, however, this sparse operation actually
      lowers into a call into a support library to obtain access to the
      values array.
 
      Example:
 
     ```mlir
-    %1 = linalg.sparse_values %0 : tensor<64x64xf64> to memref<?xf64>
+    %1 = sparse_tensor.values %0 : tensor<64x64xf64> to memref<?xf64>
     ```
   }];
   let assemblyFormat = "$tensor attr-dict `:` type($tensor) `to` type($result)";
 }
 
-#endif // LINALG_SPARSE_OPS
+#endif // SPARSETENSOR_OPS

mlir/include/mlir/Dialect/SparseTensor/Transforms/Transforms.h

+//===- Transforms.h - Sparse transformations --------------------*- 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_SPARSETENSOR_TRANSFORMS_TRANSFORMS_H_
+#define MLIR_DIALECT_SPARSETENSOR_TRANSFORMS_TRANSFORMS_H_
+
+#include "mlir/IR/PatternMatch.h"
+
+namespace mlir {
+namespace sparse_tensor {
+
+/// Sets up sparsification conversion rules with the given options.
+void populateSparsificationConversionPatterns(RewritePatternSet &patterns);
+
+} // namespace sparse_tensor
+} // namespace mlir
+
+#endif // MLIR_DIALECT_SPARSETENSOR_TRANSFORMS_TRANSFORMS_H_

mlir/include/mlir/InitAllDialects.h

 #include "mlir/Dialect/OpenMP/OpenMPDialect.h"
 #include "mlir/Dialect/PDL/IR/PDL.h"
 #include "mlir/Dialect/PDLInterp/IR/PDLInterp.h"
 #include "mlir/Dialect/Quant/QuantOps.h"
 #include "mlir/Dialect/SCF/SCF.h"
 #include "mlir/Dialect/SDBM/SDBMDialect.h"
 #include "mlir/Dialect/SPIRV/IR/SPIRVDialect.h"
 #include "mlir/Dialect/Shape/IR/Shape.h"
+#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/Dialect/Tensor/IR/Tensor.h"
 #include "mlir/Dialect/Tosa/IR/TosaOps.h"
 #include "mlir/Dialect/Vector/VectorOps.h"
 #include "mlir/Dialect/X86Vector/X86VectorDialect.h"
 #include "mlir/IR/Dialect.h"
 
 namespace mlir {
registry.insert<acc::OpenACCDialect,
                   spirv::SPIRVDialect,
                   StandardOpsDialect,
                   arm_sve::ArmSVEDialect,
                   vector::VectorDialect,
                   NVVM::NVVMDialect,
                   ROCDL::ROCDLDialect,
                   SDBMDialect,
                   shape::ShapeDialect,
+                  sparse_tensor::SparseTensorDialect,
                   tensor::TensorDialect,
                   tosa::TosaDialect,
                   x86vector::X86VectorDialect>();
   // clang-format on
 }
 
 /// Append all the MLIR dialects to the registry contained in the given context.
 inline void registerAllDialects(MLIRContext &context) {
   DialectRegistry registry;
   registerAllDialects(registry);
   context.appendDialectRegistry(registry);
 }
 
 } // namespace mlir
 
 #endif // MLIR_INITALLDIALECTS_H_

mlir/lib/Dialect/CMakeLists.txt

 add_subdirectory(OpenACC)
 add_subdirectory(OpenMP)
 add_subdirectory(PDL)
 add_subdirectory(PDLInterp)
 add_subdirectory(Quant)
 add_subdirectory(SCF)
 add_subdirectory(SDBM)
 add_subdirectory(Shape)
+add_subdirectory(SparseTensor)
 add_subdirectory(SPIRV)
 add_subdirectory(StandardOps)
 add_subdirectory(Tensor)
 add_subdirectory(Tosa)
 add_subdirectory(Utils)
 add_subdirectory(Vector)
 add_subdirectory(X86Vector)
 

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

 #include "mlir/Dialect/Linalg/IR/LinalgNamedStructuredOps.yamlgen.cpp.inc"
 
 #define GET_OP_CLASSES
 #include "mlir/Dialect/Linalg/IR/LinalgOps.cpp.inc"
 
 #define GET_OP_CLASSES
 #include "mlir/Dialect/Linalg/IR/LinalgStructuredOps.cpp.inc"
 
-#define GET_OP_CLASSES
-#include "mlir/Dialect/Linalg/IR/LinalgSparseOps.cpp.inc"
-
 /// Return the dims that are `iteratorTypeName` loops in the LinalgOp `op`.
 /// Assumes `op` is a LinalgOp.
 void mlir::linalg::getDimsOfType(Operation *op, StringRef iteratorTypeName,
                                  SmallVectorImpl<AffineExpr> &res) {
   if (!cast<LinalgOp>(op).iterator_types())
     return;
 
   unsigned dim = 0;

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

void mlir::linalg::LinalgDialect::initialize() {
   addOperations<
 #define GET_OP_LIST
 #include "mlir/Dialect/Linalg/IR/LinalgOps.cpp.inc"
       >();
   addOperations<
 #define GET_OP_LIST
 #include "mlir/Dialect/Linalg/IR/LinalgStructuredOps.cpp.inc"
       >();
-  addOperations<
-#define GET_OP_LIST
-#include "mlir/Dialect/Linalg/IR/LinalgSparseOps.cpp.inc"
-      >();
 
   // Fill the Linalg-specific OpName to RegionBuilder map.
   addNamedOpBuilders<
 #define GET_OP_LIST
 #include "mlir/Dialect/Linalg/IR/LinalgStructuredOps.cpp.inc"
       >(namedStructuredOpRegionBuilders);
 
   addInterfaces<LinalgInlinerInterface>();

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

 add_mlir_dialect_library(MLIRLinalgTransforms
   Bufferize.cpp
   CodegenStrategy.cpp
   Detensorize.cpp
   DropUnitDims.cpp
   ElementwiseToLinalg.cpp
   Fusion.cpp
   FusionOnTensors.cpp
   Generalization.cpp
   Hoisting.cpp
   Interchange.cpp
   Loops.cpp
   Promotion.cpp
-  SparseLowering.cpp
   Sparsification.cpp
   Tiling.cpp
   Transforms.cpp
   Vectorization.cpp
 
   ADDITIONAL_HEADER_DIRS
   ${MLIR_MAIN_INCLUDE_DIR}/mlir/Dialect/Linalg
 

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

 // Implementation detail: We use llvm::SmallVector for vectors with
 // variable lengths and std::vector for vectors with fixed lengths.
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/Linalg/IR/LinalgOps.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
 #include "mlir/Dialect/Linalg/Utils/Utils.h"
 #include "mlir/Dialect/SCF/SCF.h"
+#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
 #include "mlir/Dialect/StandardOps/IR/Ops.h"
 #include "mlir/IR/Matchers.h"
 
 using namespace mlir;
 
 namespace {
 
 enum class Kind { kTensor, kInvariant, kMulF, kMulI, kAddF, kAddI };
for (unsigned d = 0; d < rank; d++) {
       }
     }
   }
 }
 
 /// Returns true if tensor was set up with sparse storage scheme.
 static bool linkedSparse(linalg::GenericOp op, unsigned tensor) {
   if (tensor < op.getNumInputs())
-    return isa_and_nonnull<linalg::SparseTensorFromPointerOp>(
+    return isa_and_nonnull<sparse_tensor::FromPointerOp>(
         op.getInput(tensor).getDefiningOp());
   return false;
 }
 
 /// A DFS helper to compute a topological sort. Note that recursion is
 /// bounded by the number of implicit loops, which is always small.
 /// Returns false when a cycle is detected.
 static bool topSortDFS(unsigned i, std::vector<unsigned> &visit,
for (unsigned d = 0, rank = shape.size(); d < rank; d++) {
         dense = false;
         auto dynShape = {ShapedType::kDynamicSize};
         auto ptrTp = MemRefType::get(
             dynShape, genIntType(rewriter, codegen.options.ptrType));
         auto indTp = MemRefType::get(
             dynShape, genIntType(rewriter, codegen.options.indType));
         Value dim = rewriter.create<ConstantIndexOp>(loc, d);
         // Generate sparse primitives to obtains pointer and indices.
-        codegen.pointers[t][i] =
-            rewriter.create<linalg::SparseTensorToPointersMemRefOp>(
-                loc, ptrTp, tensor, dim);
-        codegen.indices[t][i] =
-            rewriter.create<linalg::SparseTensorToIndicesMemRefOp>(loc, indTp,
-                                                                   tensor, dim);
+        codegen.pointers[t][i] = rewriter.create<sparse_tensor::ToPointersOp>(
+        codegen.indices[t][i] = rewriter.create<sparse_tensor::ToIndicesOp>(
+            loc, indTp, tensor, dim);
       }
       // Find lower and upper bound in current dimension.
       Value up;
       if (shape[d] == MemRefType::kDynamicSize) {
         up = rewriter.create<memref::DimOp>(loc, tensor, d);
         args.push_back(up);
       } else {
         up = rewriter.create<ConstantIndexOp>(loc, shape[d]);
if (dense) {
             rewriter.create<memref::BufferCastOp>(loc, denseTp, tensor);
       else
         codegen.buffers[t] =
             genOutputBuffer(codegen, rewriter, op, denseTp, args);
     } else {
       auto dynShape = {ShapedType::kDynamicSize};
       auto sparseTp = MemRefType::get(dynShape, tensorType.getElementType());
       codegen.buffers[t] =
-          rewriter.create<linalg::SparseTensorToValuesMemRefOp>(loc, sparseTp,
-                                                                tensor);
+          rewriter.create<sparse_tensor::ToValuesOp>(loc, sparseTp, tensor);
     }
   }
 }
 
 /// Constructs vector type.
 static VectorType vectorType(CodeGen &codegen, Type etp) {
   return VectorType::get(codegen.curVecLength, etp);
 }

mlir/lib/Dialect/SparseTensor/CMakeLists.txt

+add_subdirectory(IR)
+add_subdirectory(Transforms)

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

+add_mlir_dialect_library(MLIRSparseTensor
+  SparseTensorDialect.cpp
+
+  ADDITIONAL_HEADER_DIRS
+  ${MLIR_MAIN_INCLUDE_DIR}/mlir/Dialect/SparseTensor
+
+  DEPENDS
+  MLIRSparseTensorOpsIncGen
+
+  LINK_LIBS PUBLIC
+  MLIRDialect
+  MLIRIR
+  )

mlir/lib/Dialect/SparseTensor/IR/SparseTensorDialect.cpp

+//===- SparseTensorDialect.cpp - Sparse tensor dialect 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/SparseTensor/IR/SparseTensor.h"
+
+#include "mlir/IR/Builders.h"
+#include "mlir/IR/OpImplementation.h"
+
+using namespace mlir;
+using namespace mlir::sparse_tensor;
+
+void SparseTensorDialect::initialize() {
+  addOperations<
+#define GET_OP_LIST
+#include "mlir/Dialect/SparseTensor/IR/SparseTensorOps.cpp.inc"
+      >();
+}
+
+#define GET_OP_CLASSES
+#include "mlir/Dialect/SparseTensor/IR/SparseTensorOps.cpp.inc"

mlir/lib/Dialect/SparseTensor/Transforms/CMakeLists.txt

+add_mlir_dialect_library(MLIRSparseTensorTransforms
+  SparseTensorLowering.cpp
+
+  ADDITIONAL_HEADER_DIRS
+  ${MLIR_MAIN_INCLUDE_DIR}/mlir/Dialect/SparseTensor
+
+  LINK_LIBS PUBLIC
+  MLIRIR
+  MLIRLLVMIR
+  MLIRPass
+  MLIRStandard
+  MLIRSparseTensor
+  MLIRTransforms
+)

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

-//===- SparseLowering.cpp - Lowers sparse primitives to library calls.  ---===//
+//===- SparseTensorLowering.cpp - Sparse tensor primitives lowering
+//--------===//
 //
 // 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
 //
 //===----------------------------------------------------------------------===//
+//
+// Lower sparse tensor primitives to calls into a runtime support library.
+// Note that this is a current implementation choice to keep the lowering
+// simple. In principle, these primitives could also be lowered to actual
+// elaborate IR code that implements the primitives on the selected sparse
+// tensor storage schemes.
+//
+//===----------------------------------------------------------------------===//
 
 #include "mlir/Dialect/LLVMIR/LLVMTypes.h"
-#include "mlir/Dialect/Linalg/IR/LinalgOps.h"
-#include "mlir/Dialect/Linalg/Transforms/Transforms.h"
+#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
+#include "mlir/Dialect/SparseTensor/Transforms/Transforms.h"
+#include "mlir/Dialect/StandardOps/IR/Ops.h"
+#include "mlir/Transforms/DialectConversion.h"
 
 using namespace mlir;
 
 namespace {
 
 /// Returns function reference (first hit also inserts into module).
 static FlatSymbolRefAttr getFunc(Operation *op, StringRef name, Type result,
                                  ValueRange operands) {
   MLIRContext *context = op->getContext();
   auto module = op->getParentOfType<ModuleOp>();
   auto func = module.lookupSymbol<FuncOp>(name);
   if (!func) {
     OpBuilder moduleBuilder(module.getBodyRegion());
     moduleBuilder
         .create<FuncOp>(op->getLoc(), name,
                         FunctionType::get(context, operands.getTypes(), result))
         .setPrivate();
   }
   return SymbolRefAttr::get(context, name);
 }
 
 /// Sparse conversion rule to remove opaque pointer cast.
 class TensorFromPointerConverter
-    : public OpConversionPattern<linalg::SparseTensorFromPointerOp> {
+    : public OpConversionPattern<sparse_tensor::FromPointerOp> {
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(linalg::SparseTensorFromPointerOp op,
-                  ArrayRef<Value> operands,
+  matchAndRewrite(sparse_tensor::FromPointerOp op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     rewriter.replaceOp(op, operands[0]);
     return success();
   }
 };
 
 /// Sparse conversion rule for dimension accesses.
 class TensorToDimSizeConverter : public OpConversionPattern<memref::DimOp> {
matchAndRewrite(memref::DimOp op, ArrayRef<Value> operands,
     rewriter.replaceOpWithNewOp<CallOp>(
         op, resType, getFunc(op, name, resType, operands), operands);
     return success();
   }
 };
 
 /// Sparse conversion rule for pointer accesses.
 class TensorToPointersConverter
-    : public OpConversionPattern<linalg::SparseTensorToPointersMemRefOp> {
+    : public OpConversionPattern<sparse_tensor::ToPointersOp> {
 public:
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(linalg::SparseTensorToPointersMemRefOp op,
-                  ArrayRef<Value> operands,
+  matchAndRewrite(sparse_tensor::ToPointersOp op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     Type resType = op.getType();
     Type eltType = resType.cast<ShapedType>().getElementType();
     StringRef name;
     if (eltType.isIndex() || eltType.isInteger(64))
       name = "sparsePointers64";
     else if (eltType.isInteger(32))
       name = "sparsePointers32";
     else if (eltType.isInteger(16))
       name = "sparsePointers16";
     else if (eltType.isInteger(8))
       name = "sparsePointers8";
     else
       return failure();
     rewriter.replaceOpWithNewOp<CallOp>(
         op, resType, getFunc(op, name, resType, operands), operands);
     return success();
   }
 };
 
 /// Sparse conversion rule for index accesses.
 class TensorToIndicesConverter
-    : public OpConversionPattern<linalg::SparseTensorToIndicesMemRefOp> {
+    : public OpConversionPattern<sparse_tensor::ToIndicesOp> {
 public:
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(linalg::SparseTensorToIndicesMemRefOp op,
-                  ArrayRef<Value> operands,
+  matchAndRewrite(sparse_tensor::ToIndicesOp op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     Type resType = op.getType();
     Type eltType = resType.cast<ShapedType>().getElementType();
     StringRef name;
     if (eltType.isIndex() || eltType.isInteger(64))
       name = "sparseIndices64";
     else if (eltType.isInteger(32))
       name = "sparseIndices32";
     else if (eltType.isInteger(16))
       name = "sparseIndices16";
     else if (eltType.isInteger(8))
       name = "sparseIndices8";
     else
       return failure();
     rewriter.replaceOpWithNewOp<CallOp>(
         op, resType, getFunc(op, name, resType, operands), operands);
     return success();
   }
 };
 
 /// Sparse conversion rule for value accesses.
 class TensorToValuesConverter
-    : public OpConversionPattern<linalg::SparseTensorToValuesMemRefOp> {
+    : public OpConversionPattern<sparse_tensor::ToValuesOp> {
 public:
   using OpConversionPattern::OpConversionPattern;
   LogicalResult
-  matchAndRewrite(linalg::SparseTensorToValuesMemRefOp op,
-                  ArrayRef<Value> operands,
+  matchAndRewrite(sparse_tensor::ToValuesOp op, ArrayRef<Value> operands,
                   ConversionPatternRewriter &rewriter) const override {
     Type resType = op.getType();
     Type eltType = resType.cast<ShapedType>().getElementType();
     StringRef name;
     if (eltType.isF64())
       name = "sparseValuesF64";
     else if (eltType.isF32())
       name = "sparseValuesF32";
matchAndRewrite(sparse_tensor::ToValuesOp op, ArrayRef<Value> operands,
     return success();
   }
 };
 
 } // namespace
 
 /// Populates the given patterns list with conversion rules required for
 /// the sparsification of linear algebra operations.
-void linalg::populateSparsificationConversionPatterns(
+void sparse_tensor::populateSparsificationConversionPatterns(
     RewritePatternSet &patterns) {
   patterns.add<TensorFromPointerConverter, TensorToDimSizeConverter,
                TensorToPointersConverter, TensorToIndicesConverter,
                TensorToValuesConverter>(patterns.getContext());
 }

mlir/test/Dialect/Linalg/sparse_1d.mlir

 // CHECK-LABEL:   func @add_s(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant true
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_10]], %[[VAL_11]]) : memref<32xf32>, 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:.*]] = cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
 // CHECK:             scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
func @add_s(%arga: tensor<32xf32>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
   return %0 : tensor<32xf32>
 }
 
 // CHECK-LABEL:   func @repeated_add_s(
 // CHECK-SAME:                         %[[VAL_0:.*]]: tensor<32xf32>,
 // CHECK-SAME:                         %[[VAL_1:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_2:.*]] = constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = constant 1 : index
-// CHECK:           %[[VAL_4:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<32xf32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_7]], %[[VAL_8]]) : memref<32xf32>, 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-LABEL:   func @mul_s(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<32xf32>,
 // CHECK-SAME:                %[[VAL_1:.*]]: f32,
 // CHECK-SAME:                %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_8]], %[[VAL_9]]) : memref<32xf32>, 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-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant true
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
 // CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32xf32>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32xf32>, 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:.*]] = cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
 // CHECK:             scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index

 
 // CHECK-LABEL:   func @mul_ds(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
 // CHECK:           %[[VAL_5:.*]] = memref.buffer_cast %[[VAL_0]] : memref<32xf32>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_9]], %[[VAL_10]]) : memref<32xf32>, 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-LABEL:   func @add_sd(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant true
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> 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_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32xf32>, 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:.*]] = cmpi ult, %[[VAL_16]], %[[VAL_14]] : index

 }
 
 // CHECK-LABEL:   func @mul_sd(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> 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_10:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_9]], %[[VAL_10]]) : memref<32xf32>, 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-LABEL:   func @add_ss(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32xf32>, 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-LABEL:   func @mul_ss(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32xf32>, 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-LABEL:   func @two_way_inv(
 // CHECK-SAME:                      %[[VAL_0:.*0]]: tensor<16xf32>,
 // CHECK-SAME:                      %[[VAL_1:.*1]]: tensor<16xf32>,
 // CHECK-SAME:                      %[[VAL_2:.*2]]: f32,
 // CHECK-SAME:                      %[[VAL_3:.*3]]: tensor<16xf32>) -> tensor<16xf32> {
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<16xf32>, 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-LABEL:   func @two_way_inv_alt(
 // CHECK-SAME:                          %[[VAL_0:.*0]]: tensor<16xf32>,
 // CHECK-SAME:                          %[[VAL_1:.*1]]: tensor<16xf32>,
 // CHECK-SAME:                          %[[VAL_2:.*2]]: f32,
 // CHECK-SAME:                          %[[VAL_3:.*3]]: tensor<16xf32>) -> tensor<16xf32> {
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_3]] : memref<16xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<16xf32>, 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) {
#trait_sum_reduction = {
   doc = "x += SUM_i a(i)"
 }
 
 // CHECK-LABEL:   func @sum_reduction(
 // CHECK-SAME:                        %[[VAL_0:.*]]: tensor<?xf32>,
 // CHECK-SAME:                        %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_2:.*]] = constant 0 : index
 // CHECK:           %[[VAL_3:.*]] = constant 1 : index
-// CHECK:           %[[VAL_4:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_2]] : tensor<?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf32> to memref<?xf32>
 // CHECK:           %[[VAL_6:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_7:.*]] = memref.alloc() : memref<f32>
 // CHECK:           linalg.copy(%[[VAL_6]], %[[VAL_7]]) : memref<f32>, memref<f32>
 // CHECK:           %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
 // CHECK:           %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
 // CHECK:           %[[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-LABEL:   func @sum_reduction_ss(
 // CHECK-SAME:                           %[[VAL_0:.*0]]: tensor<16xf32>,
 // CHECK-SAME:                           %[[VAL_1:.*1]]: tensor<16xf32>,
 // CHECK-SAME:                           %[[VAL_2:.*2]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<f32>
 // CHECK:           %[[VAL_12:.*]] = memref.alloc() : memref<f32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<f32>, memref<f32>
 // 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-LABEL:   func @sum_reduction_inv(
 // CHECK-SAME:                            %[[VAL_0:.*0]]: tensor<16xf32>,
 // CHECK-SAME:                            %[[VAL_1:.*1]]: tensor<f32>,
 // CHECK-SAME:                            %[[VAL_2:.*2]]: tensor<16xf32>,
 // CHECK-SAME:                            %[[VAL_3:.*3]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_2]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_2]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_values %[[VAL_2]] : tensor<16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_3]] : memref<f32>
 // CHECK:           %[[VAL_14:.*]] = memref.alloc() : memref<f32>
 // CHECK:           linalg.copy(%[[VAL_13]], %[[VAL_14]]) : memref<f32>, memref<f32>
 // CHECK:           %[[VAL_15:.*]] = memref.load %[[VAL_9]][] : memref<f32>
 // CHECK:           %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_18:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_5]]] : memref<?xindex>

 // CHECK-SAME:                          %[[VAL_1:.*1]]: tensor<?xf64>,
 // CHECK-SAME:                          %[[VAL_2:.*2]]: tensor<?xf64>,
 // CHECK-SAME:                          %[[VAL_3:.*3]]: tensor<?xf64>,
 // CHECK-SAME:                          %[[VAL_4:.*4]]: tensor<?xf64>) -> tensor<?xf64> {
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant true
 // CHECK:           %[[VAL_7:.*]] = constant 1 : index
 // CHECK:           %[[VAL_8:.*]] = memref.buffer_cast %[[VAL_0]] : memref<?xf64>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64> to memref<?xf64>
 // CHECK:           %[[VAL_12:.*]] = memref.buffer_cast %[[VAL_2]] : memref<?xf64>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_pointers %[[VAL_3]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_indices %[[VAL_3]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_15:.*]] = linalg.sparse_values %[[VAL_3]] : tensor<?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.pointers %[[VAL_3]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.indices %[[VAL_3]], %[[VAL_5]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_15:.*]] = sparse_tensor.values %[[VAL_3]] : tensor<?xf64> to memref<?xf64>
 // CHECK:           %[[VAL_16:.*]] = memref.dim %[[VAL_4]], %[[VAL_5]] : tensor<?xf64>
 // CHECK:           %[[VAL_17:.*]] = memref.buffer_cast %[[VAL_4]] : memref<?xf64>
 // CHECK:           %[[VAL_18:.*]] = memref.alloc(%[[VAL_16]]) : memref<?xf64>
 // CHECK:           linalg.copy(%[[VAL_17]], %[[VAL_18]]) : memref<?xf64>, 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-LABEL:   func @red3s(
 // CHECK-SAME:                %[[VAL_0:.*0]]: tensor<?xf64>,
 // CHECK-SAME:                %[[VAL_1:.*1]]: tensor<?xf64>,
 // CHECK-SAME:                %[[VAL_2:.*2]]: tensor<?xf64>,
 // CHECK-SAME:                %[[VAL_3:.*3]]: tensor<f64>) -> tensor<f64> {
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<?xf64> to memref<?xf64>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<?xf64> to memref<?xf64>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_2]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_2]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_2]] : tensor<?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<?xf64> to memref<?xf64>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_3]] : memref<f64>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<f64>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<f64>, memref<f64>
 // 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_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
 // CHECK:           %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
 // CHECK:           %[[VAL_21:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_4]]] : memref<?xindex>

mlir/test/Dialect/Linalg/sparse_2d.mlir

 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant true
 // CHECK:           %[[VAL_7:.*]] = constant 1 : index
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_13:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<32x16xf32>, 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:.*]] = addi %[[VAL_14]], %[[VAL_7]] : index
 // CHECK:             %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_16]]] : memref<?xindex>

 
 // CHECK-LABEL:   func @mul_ds(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_11:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_10]], %[[VAL_11]]) : memref<32x16xf32>, 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:.*]] = addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>

 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant true
 // CHECK:           %[[VAL_6:.*]] = constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = constant 1 : index
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_13:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<32x16xf32>, 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:.*]] = cmpi ult, %[[VAL_17]], %[[VAL_15]] : index

 
 // CHECK-LABEL:   func @mul_sd(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 16 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_11:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_10]], %[[VAL_11]]) : memref<32x16xf32>, 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-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant true
 // CHECK:           %[[VAL_6:.*]] = constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = constant 1 : index
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_15:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_14]], %[[VAL_15]]) : memref<32x16xf32>, 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:.*]] = cmpi ult, %[[VAL_19]], %[[VAL_17]] : index

 }
 
 // CHECK-LABEL:   func @mul_ss(
 // CHECK-SAME:                 %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> 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_12:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32x16xf32>, 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-LABEL:   func @add_ss_ss(
 // CHECK-SAME:                    %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<32x16xf32>, 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-LABEL:   func @mul_ss_ss(
 // CHECK-SAME:                    %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<32x16xf32>, 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-LABEL:   func @add_sd_ds(
 // CHECK-SAME:                    %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<32x16xf32>, 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-LABEL:   func @mul_sd_ds(
 // CHECK-SAME:                    %[[VAL_0:.*0]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_1:.*1]]: tensor<32x16xf32>,
 // CHECK-SAME:                    %[[VAL_2:.*2]]: tensor<32x16xf32>) -> tensor<32x16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32x16xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32x16xf32> to memref<?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_2]] : memref<32x16xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc() : memref<32x16xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<32x16xf32>, 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-LABEL:   func @matvec(
 // CHECK-SAME:                 %[[VAL_0:.*]]: tensor<16x32xf32>,
 // CHECK-SAME:                 %[[VAL_1:.*]]: tensor<32xf32>,
 // CHECK-SAME:                 %[[VAL_2:.*]]: tensor<16xf32>) -> tensor<16xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 16 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<16x32xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<16x32xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16x32xf32> 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_11:.*]] = memref.alloc() : memref<16xf32>
 // CHECK:           linalg.copy(%[[VAL_10]], %[[VAL_11]]) : memref<16xf32>, memref<16xf32>
 // 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:.*]] = addi %[[VAL_12]], %[[VAL_5]] : index
 // CHECK:             %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_14]]] : memref<?xindex>

 }
 
 // CHECK-LABEL:   func @sum_reduction(
 // CHECK-SAME:                        %[[VAL_0:.*]]: tensor<10x20xf32>,
 // CHECK-SAME:                        %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_2:.*]] = constant 10 : index
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<10x20xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20xf32> to memref<?xf32>
 // CHECK:           %[[VAL_7:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_8:.*]] = memref.alloc() : memref<f32>
 // CHECK:           linalg.copy(%[[VAL_7]], %[[VAL_8]]) : memref<f32>, memref<f32>
 // CHECK:           scf.for %[[VAL_9:.*]] = %[[VAL_3]] to %[[VAL_2]] step %[[VAL_4]] {
 // CHECK:             %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_9]]] : memref<?xindex>
 // CHECK:             %[[VAL_11:.*]] = addi %[[VAL_9]], %[[VAL_4]] : index
 // CHECK:             %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_11]]] : memref<?xindex>
 // CHECK:             %[[VAL_13:.*]] = memref.load %[[VAL_8]][] : memref<f32>

 }
 
 // CHECK-LABEL:   func @scale(
 // CHECK-SAME:                %[[VAL_0:.*]]: tensor<?x?xf64>,
 // CHECK-SAME:                %[[VAL_1:.*]]: tensor<?x?xf64>) -> tensor<?x?xf64> {
 // CHECK-DAG:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK-DAG:           %[[VAL_4:.*]] = constant 1 : index
 // CHECK-DAG:           %[[VAL_2:.*]] = constant 2.000000e+00 : f64
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<?x?xf64> to memref<?xf64>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf64> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf64> to memref<?xf64>
 // CHECK:           %[[VAL_8:.*]] = memref.dim %[[VAL_1]], %[[VAL_3]] : tensor<?x?xf64>
 // CHECK:           %[[VAL_9:.*]] = memref.dim %[[VAL_1]], %[[VAL_4]] : tensor<?x?xf64>
 // CHECK:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<?x?xf64>
 // CHECK:           %[[VAL_11:.*]] = memref.alloc(%[[VAL_8]], %[[VAL_9]]) : memref<?x?xf64>
 // CHECK:           linalg.copy(%[[VAL_10]], %[[VAL_11]]) : memref<?x?xf64>, 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:.*]] = addi %[[VAL_12]], %[[VAL_4]] : index

 
 // CHECK-LABEL:   func @sampled_dense_dense(
 // CHECK-SAME:                              %[[VAL_0:.*0]]: tensor<?x?xf32>,
 // CHECK-SAME:                              %[[VAL_1:.*1]]: tensor<?x?xf32>,
 // CHECK-SAME:                              %[[VAL_2:.*2]]: tensor<?x?xf32>,
 // CHECK-SAME:                              %[[VAL_3:.*3]]: tensor<?x?xf32>) -> tensor<?x?xf32> {
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<?x?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf32> to memref<?xf32>
 // CHECK:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_1]] : memref<?x?xf32>
 // CHECK:           %[[VAL_12:.*]] = memref.dim %[[VAL_2]], %[[VAL_4]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_13:.*]] = memref.buffer_cast %[[VAL_2]] : memref<?x?xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.dim %[[VAL_3]], %[[VAL_4]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.dim %[[VAL_3]], %[[VAL_5]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.buffer_cast %[[VAL_3]] : memref<?x?xf32>
 // CHECK:           %[[VAL_17:.*]] = memref.alloc(%[[VAL_14]], %[[VAL_15]]) : memref<?x?xf32>
 // CHECK:           linalg.copy(%[[VAL_16]], %[[VAL_17]]) : memref<?x?xf32>, memref<?x?xf32>

 // CHECK-SAME:                              %[[VAL_1:.*1]]: tensor<?x?xf32>,
 // CHECK-SAME:                              %[[VAL_2:.*2]]: tensor<?x?xf32>,
 // CHECK-SAME:                              %[[VAL_3:.*3]]: tensor<?xf32>,
 // CHECK-SAME:                              %[[VAL_4:.*4]]: tensor<f32>,
 // CHECK-SAME:                              %[[VAL_5:.*5]]: tensor<?xf32>) -> tensor<?xf32> {
 // CHECK:           %[[VAL_6:.*]] = constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = constant true
 // CHECK:           %[[VAL_8:.*]] = constant 1 : index
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<?x?xf32> to memref<?xf32>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_15:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_16:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<?x?xf32> to memref<?xf32>
-// CHECK:           %[[VAL_17:.*]] = linalg.sparse_pointers %[[VAL_2]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_18:.*]] = linalg.sparse_indices %[[VAL_2]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_19:.*]] = linalg.sparse_values %[[VAL_2]] : tensor<?x?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?x?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_15:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_16:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?x?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_17:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_18:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_8]] : tensor<?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_19:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<?x?xf32> 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_22:.*]] = memref.dim %[[VAL_5]], %[[VAL_6]] : tensor<?xf32>
 // CHECK:           %[[VAL_23:.*]] = memref.buffer_cast %[[VAL_5]] : memref<?xf32>
 // CHECK:           %[[VAL_24:.*]] = memref.alloc(%[[VAL_22]]) : memref<?xf32>
 // CHECK:           linalg.copy(%[[VAL_23]], %[[VAL_24]]) : memref<?xf32>, memref<?xf32>
 // CHECK:           %[[VAL_25:.*]] = memref.load %[[VAL_21]][] : memref<f32>
 // CHECK:           %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_6]]] : memref<?xindex>

mlir/test/Dialect/Linalg/sparse_3d.mlir

 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = constant 8 : index
 // CHECK:           %[[VAL_7:.*]] = constant 0 : index
 // CHECK:           %[[VAL_8:.*]] = constant true
 // CHECK:           %[[VAL_9:.*]] = constant 1 : index
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_15:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_14]], %[[VAL_15]]) : memref<32x16x8xf32>, 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:.*]] = muli %[[VAL_16]], %[[VAL_5]] : index
 // CHECK:               %[[VAL_19:.*]] = addi %[[VAL_18]], %[[VAL_17]] : index

 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = constant 0 : index
 // CHECK:           %[[VAL_7:.*]] = constant 1 : index
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_13:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<32x16x8xf32>, 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:.*]] = muli %[[VAL_14]], %[[VAL_5]] : index
 // CHECK:               %[[VAL_17:.*]] = addi %[[VAL_16]], %[[VAL_15]] : index

 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant 8 : index
 // CHECK:           %[[VAL_6:.*]] = constant true
 // CHECK:           %[[VAL_7:.*]] = constant 0 : index
 // CHECK:           %[[VAL_8:.*]] = constant 1 : index
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_13]], %[[VAL_14]]) : memref<32x16x8xf32>, 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:.*]] = addi %[[VAL_15]], %[[VAL_8]] : index
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_17]]] : memref<?xindex>

 // CHECK-LABEL:   func @mul_dsd(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 8 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_12:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32x16x8xf32>, 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:.*]] = addi %[[VAL_13]], %[[VAL_6]] : index
 // CHECK:             %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>

 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = constant 8 : index
 // CHECK:           %[[VAL_7:.*]] = constant true
 // CHECK:           %[[VAL_8:.*]] = constant 0 : index
 // CHECK:           %[[VAL_9:.*]] = constant 1 : index
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_17:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_16]], %[[VAL_17]]) : memref<32x16x8xf32>, 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:.*]] = addi %[[VAL_18]], %[[VAL_9]] : index
 // CHECK:             %[[VAL_21:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_20]]] : memref<?xindex>

 // CHECK-LABEL:   func @mul_dss(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_6]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_13]], %[[VAL_14]]) : memref<32x16x8xf32>, 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:.*]] = addi %[[VAL_15]], %[[VAL_6]] : index
 // CHECK:             %[[VAL_18:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_17]]] : memref<?xindex>

 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant 8 : index
 // CHECK:           %[[VAL_6:.*]] = constant true
 // CHECK:           %[[VAL_7:.*]] = constant 0 : index
 // CHECK:           %[[VAL_8:.*]] = constant 1 : index
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_13]], %[[VAL_14]]) : memref<32x16x8xf32>, 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:.*]] = cmpi ult, %[[VAL_18]], %[[VAL_16]] : index

 // CHECK-LABEL:   func @mul_sdd(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 16 : index
 // CHECK:           %[[VAL_4:.*]] = constant 8 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_12:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<32x16x8xf32>, 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-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = constant 8 : index
 // CHECK:           %[[VAL_7:.*]] = constant true
 // CHECK:           %[[VAL_8:.*]] = constant 0 : index
 // CHECK:           %[[VAL_9:.*]] = constant 1 : index
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_17:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_16]], %[[VAL_17]]) : memref<32x16x8xf32>, 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:.*]] = cmpi ult, %[[VAL_21]], %[[VAL_19]] : index

 // CHECK-LABEL:   func @mul_sds(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_14:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_13]], %[[VAL_14]]) : memref<32x16x8xf32>, 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-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 32 : index
 // CHECK:           %[[VAL_4:.*]] = constant 16 : index
 // CHECK:           %[[VAL_5:.*]] = constant 8 : index
 // CHECK:           %[[VAL_6:.*]] = constant true
 // CHECK:           %[[VAL_7:.*]] = constant 0 : index
 // CHECK:           %[[VAL_8:.*]] = constant 1 : index
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_7]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_16:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<32x16x8xf32>, 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:.*]] = cmpi ult, %[[VAL_20]], %[[VAL_18]] : index

 
 // CHECK-LABEL:   func @mul_ssd(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 8 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_13:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_12]], %[[VAL_13]]) : memref<32x16x8xf32>, 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-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 32 : index
 // CHECK:           %[[VAL_5:.*]] = constant 16 : index
 // CHECK:           %[[VAL_6:.*]] = constant 8 : index
 // CHECK:           %[[VAL_7:.*]] = constant true
 // CHECK:           %[[VAL_8:.*]] = constant 0 : index
 // CHECK:           %[[VAL_9:.*]] = constant 1 : index
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_13:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_14:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_15:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_16:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_8]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_9]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_14:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_15:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_16:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_19:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_18]], %[[VAL_19]]) : memref<32x16x8xf32>, 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:.*]] = cmpi ult, %[[VAL_23]], %[[VAL_21]] : index

 
 // CHECK-LABEL:   func @mul_sss(
 // CHECK-SAME:                  %[[VAL_0:.*0]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_1:.*1]]: tensor<32x16x8xf32>,
 // CHECK-SAME:                  %[[VAL_2:.*2]]: tensor<32x16x8xf32>) -> tensor<32x16x8xf32> {
 // CHECK:           %[[VAL_3:.*]] = constant 2 : index
 // CHECK:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_10:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_11:.*]] = linalg.sparse_indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
-// CHECK:           %[[VAL_12:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<32x16x8xf32> to memref<?xf32>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_5]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32x16x8xf32> to memref<?xindex>
+// CHECK:           %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32x16x8xf32> 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_15:.*]] = memref.alloc() : memref<32x16x8xf32>
 // CHECK:           linalg.copy(%[[VAL_14]], %[[VAL_15]]) : memref<32x16x8xf32>, 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-LABEL:   func @kernel_3d(
 // CHECK-SAME:                    %[[VAL_0:.*0]]: tensor<?x?xf32>,
 // CHECK-SAME:                    %[[VAL_1:.*1]]: tensor<?x?x?xf32>,
 // CHECK-SAME:                    %[[VAL_2:.*2]]: tensor<?x?xf32>,
 // CHECK-SAME:                    %[[VAL_3:.*3]]: tensor<?x?xf32>) -> tensor<?x?xf32> {
 // CHECK:           %[[VAL_4:.*]] = constant 2 : index
 // CHECK:           %[[VAL_5:.*]] = constant 0 : index
 // CHECK:           %[[VAL_6:.*]] = constant 1 : index
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32> to memref<?xindex>
-// CHECK:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_1]] : tensor<?x?x?xf32> to memref<?xf32>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<?x?x?xf32> to memref<?xindex>
+// CHECK:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?x?x?xf32> to memref<?xf32>
 // CHECK:           %[[VAL_10:.*]] = memref.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_13:.*]] = memref.dim %[[VAL_0]], %[[VAL_5]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_14:.*]] = memref.dim %[[VAL_0]], %[[VAL_6]] : tensor<?x?xf32>
 // CHECK:           %[[VAL_15:.*]] = memref.buffer_cast %[[VAL_0]] : memref<?x?xf32>
 // CHECK:           %[[VAL_16:.*]] = memref.alloc(%[[VAL_13]], %[[VAL_14]]) : memref<?x?xf32>
 // CHECK:           linalg.copy(%[[VAL_15]], %[[VAL_16]]) : memref<?x?xf32>, memref<?x?xf32>

 }
 
 // CHECK-LABEL:   func @sum_reduction(
 // CHECK-SAME:                        %[[VAL_0:.*]]: tensor<10x20x30xf32>,
 // CHECK-SAME:                        %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
 // CHECK:           %[[VAL_2:.*]] = constant 2 : index
 // CHECK:           %[[VAL_3:.*]] = constant 0 : index
 // CHECK:           %[[VAL_4:.*]] = constant 1 : index
-// CHECK:           %[[VAL_5:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_3]] : tensor<10x20x30xf32> to memref<?xindex>
-// CHECK:           %[[VAL_6:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20x30xf32> to memref<?xindex>
-// CHECK:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_0]], %[[VAL_2]] : tensor<10x20x30xf32> to memref<?xindex>
-// CHECK:           %[[VAL_8:.*]] = linalg.sparse_values %[[VAL_0]] : tensor<10x20x30xf32> to memref<?xf32>
+// CHECK:           %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<10x20x30xf32> to memref<?xindex>
+// CHECK:           %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<10x20x30xf32> to memref<?xindex>
+// CHECK:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<10x20x30xf32> to memref<?xindex>
+// CHECK:           %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<10x20x30xf32> to memref<?xf32>
 // CHECK:           %[[VAL_9:.*]] = memref.buffer_cast %[[VAL_1]] : memref<f32>
 // CHECK:           %[[VAL_10:.*]] = memref.alloc() : memref<f32>
 // CHECK:           linalg.copy(%[[VAL_9]], %[[VAL_10]]) : memref<f32>, memref<f32>
 // 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:.*]] = addi %[[VAL_13]], %[[VAL_4]] : index

mlir/test/Dialect/Linalg/sparse_lower.mlir

 
 // CHECK-HIR-LABEL:   func @matvec(
 // CHECK-HIR-SAME:                 %[[VAL_0:.*]]: !llvm.ptr<i8>,
 // CHECK-HIR-SAME:                 %[[VAL_1:.*]]: tensor<64xf64>,
 // CHECK-HIR-SAME:                 %[[VAL_2:.*]]: tensor<64xf64>) -> tensor<64xf64> {
 // CHECK-HIR:           %[[VAL_3:.*]] = constant 64 : index
 // CHECK-HIR:           %[[VAL_4:.*]] = constant 0 : index
 // CHECK-HIR:           %[[VAL_5:.*]] = constant 1 : index
-// CHECK-HIR:           %[[VAL_6:.*]] = linalg.sparse_tensor %[[VAL_0]] : !llvm.ptr<i8> to tensor<64x64xf64>
-// CHECK-HIR:           %[[VAL_7:.*]] = linalg.sparse_pointers %[[VAL_6]], %[[VAL_5]] : tensor<64x64xf64> to memref<?xindex>
-// CHECK-HIR:           %[[VAL_8:.*]] = linalg.sparse_indices %[[VAL_6]], %[[VAL_5]] : tensor<64x64xf64> to memref<?xindex>
-// CHECK-HIR:           %[[VAL_9:.*]] = linalg.sparse_values %[[VAL_6]] : tensor<64x64xf64> to memref<?xf64>
+// CHECK-HIR:           %[[VAL_6:.*]] = sparse_tensor.fromPtr %[[VAL_0]] : !llvm.ptr<i8> to tensor<64x64xf64>
+// CHECK-HIR:           %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_6]], %[[VAL_5]] : tensor<64x64xf64> to memref<?xindex>
+// CHECK-HIR:           %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_6]], %[[VAL_5]] : tensor<64x64xf64> to memref<?xindex>
+// CHECK-HIR:           %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_6]] : tensor<64x64xf64> to memref<?xf64>
 // CHECK-HIR:           %[[VAL_10:.*]] = memref.buffer_cast %[[VAL_1]] : memref<64xf64>
 // CHECK-HIR:           %[[VAL_11:.*]] = memref.buffer_cast %[[VAL_2]] : memref<64xf64>
 // CHECK-HIR:           %[[VAL_12:.*]] = memref.alloc() : memref<64xf64>
 // CHECK-HIR:           linalg.copy(%[[VAL_11]], %[[VAL_12]]) : memref<64xf64>, memref<64xf64>
 // CHECK-HIR:           scf.for %[[VAL_13:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
 // CHECK-HIR:             %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xindex>
 // CHECK-HIR:             %[[VAL_15:.*]] = addi %[[VAL_13]], %[[VAL_5]] : index
 // CHECK-HIR:             %[[VAL_16:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_15]]] : memref<?xindex>

 // CHECK-FAST:             store %[[VAL_22:.*]], %[[VAL_2]]{{\[}}%[[VAL_9]]] : memref<64xf64>
 // CHECK-FAST:           }
 // CHECK-FAST:           return %[[VAL_2]] : memref<64xf64>
 // CHECK-FAST:         }
 
 !SparseTensor = type !llvm.ptr<i8>
 
 func @matvec(%argA: !SparseTensor, %argb: tensor<64xf64>, %argx: tensor<64xf64>) -> tensor<64xf64> {
-  %arga = linalg.sparse_tensor %argA : !SparseTensor to tensor<64x64xf64>
+  %arga = sparse_tensor.fromPtr %argA : !SparseTensor to tensor<64x64xf64>
   %0 = linalg.generic #trait_matvec
       ins(%arga, %argb : tensor<64x64xf64>, tensor<64xf64>)
       outs(%argx: tensor<64xf64>) {
     ^bb(%A: f64, %b: f64, %x: f64):
       %0 = mulf %A, %b : f64
       %1 = addf %x, %0 : f64
       linalg.yield %1 : f64
   } -> tensor<64xf64>
   return %0 : tensor<64xf64>
 }

mlir/test/Dialect/Linalg/sparse_nd.mlir

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

mlir/test/Dialect/Linalg/sparse_vector.mlir

 // CHECK-VEC3:         %[[m:.*]] = mulf %[[la]], %[[lb]] : vector<16xf32>
 // CHECK-VEC3:         vector.scatter %{{.*}}[%[[c0]]] [%[[li]]], %[[mask]], %[[m]] : memref<1024xf32>, vector<16xindex>, vector<16xi1>, vector<16xf32>
 // CHECK-VEC3:       }
 // CHECK-VEC3:       return
 //
 //
 !SparseTensor = type !llvm.ptr<i8>
 func @mul_s_alt(%argA: !SparseTensor, %argB: !SparseTensor, %argx: tensor<1024xf32>) -> tensor<1024xf32> {
-  %arga = linalg.sparse_tensor %argA : !SparseTensor to tensor<1024xf32>
-  %argb = linalg.sparse_tensor %argB : !SparseTensor to tensor<1024xf32>
+  %arga = sparse_tensor.fromPtr %argA : !SparseTensor to tensor<1024xf32>
+  %argb = sparse_tensor.fromPtr %argB : !SparseTensor to tensor<1024xf32>
   %0 = linalg.generic #trait_mul_s
     ins(%arga, %argb: tensor<1024xf32>, tensor<1024xf32>)
     outs(%argx: tensor<1024xf32>) {
       ^bb(%a: f32, %b: f32, %x: f32):
         %0 = mulf %a, %b : f32
         linalg.yield %0 : f32
   } -> tensor<1024xf32>
   return %0 : tensor<1024xf32>

mlir/test/Dialect/SparseTensor/lowering.mlir

 // RUN: mlir-opt --test-sparsification="lower" %s | FileCheck %s
 
 !SparseTensor = type !llvm.ptr<i8>
 
 // CHECK-LABEL: func @sparse_pointers(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[C:.*]] = constant 1 : index
 //       CHECK: %[[T:.*]] = call @sparsePointers64(%[[A]], %[[C]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 //       CHECK: return %[[T]] : memref<?xindex>
 func @sparse_pointers(%arg0: !SparseTensor) -> memref<?xindex> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
   %c = constant 1 : index
-  %0 = linalg.sparse_pointers %a, %c : tensor<128xf64> to memref<?xindex>
+  %0 = sparse_tensor.pointers %a, %c : tensor<128xf64> to memref<?xindex>
   return %0 : memref<?xindex>
 }
 
 // CHECK-LABEL: func @sparse_pointers32(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[C:.*]] = constant 1 : index
 //       CHECK: %[[T:.*]] = call @sparsePointers32(%[[A]], %[[C]]) : (!llvm.ptr<i8>, index) -> memref<?xi32>
 //       CHECK: return %[[T]] : memref<?xi32>
 func @sparse_pointers32(%arg0: !SparseTensor) -> memref<?xi32> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
   %c = constant 1 : index
-  %0 = linalg.sparse_pointers %a, %c : tensor<128xf64> to memref<?xi32>
+  %0 = sparse_tensor.pointers %a, %c : tensor<128xf64> to memref<?xi32>
   return %0 : memref<?xi32>
 }
 
 // CHECK-LABEL: func @sparse_indices(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[C:.*]] = constant 1 : index
 //       CHECK: %[[T:.*]] = call @sparseIndices64(%[[A]], %[[C]]) : (!llvm.ptr<i8>, index) -> memref<?xindex>
 //       CHECK: return %[[T]] : memref<?xindex>
 func @sparse_indices(%arg0: !SparseTensor) -> memref<?xindex> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
   %c = constant 1 : index
-  %0 = linalg.sparse_indices %a, %c : tensor<128xf64> to memref<?xindex>
+  %0 = sparse_tensor.indices %a, %c : tensor<128xf64> to memref<?xindex>
   return %0 : memref<?xindex>
 }
 
 // CHECK-LABEL: func @sparse_indices32(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[C:.*]] = constant 1 : index
 //       CHECK: %[[T:.*]] = call @sparseIndices32(%[[A]], %[[C]]) : (!llvm.ptr<i8>, index) -> memref<?xi32>
 //       CHECK: return %[[T]] : memref<?xi32>
 func @sparse_indices32(%arg0: !SparseTensor) -> memref<?xi32> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
   %c = constant 1 : index
-  %0 = linalg.sparse_indices %a, %c : tensor<128xf64> to memref<?xi32>
+  %0 = sparse_tensor.indices %a, %c : tensor<128xf64> to memref<?xi32>
   return %0 : memref<?xi32>
 }
 
 // CHECK-LABEL: func @sparse_valuesf64(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[T:.*]] = call @sparseValuesF64(%[[A]]) : (!llvm.ptr<i8>) -> memref<?xf64>
 //       CHECK: return %[[T]] : memref<?xf64>
 func @sparse_valuesf64(%arg0: !SparseTensor) -> memref<?xf64> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
-  %0 = linalg.sparse_values %a : tensor<128xf64> to memref<?xf64>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
+  %0 = sparse_tensor.values %a : tensor<128xf64> to memref<?xf64>
   return %0 : memref<?xf64>
 }
 
 // CHECK-LABEL: func @sparse_valuesf32(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
 //       CHECK: %[[T:.*]] = call @sparseValuesF32(%[[A]]) : (!llvm.ptr<i8>) -> memref<?xf32>
 //       CHECK: return %[[T]] : memref<?xf32>
 func @sparse_valuesf32(%arg0: !SparseTensor) -> memref<?xf32> {
-  %a = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf32>
-  %0 = linalg.sparse_values %a : tensor<128xf32> to memref<?xf32>
+  %a = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf32>
+  %0 = sparse_tensor.values %a : tensor<128xf32> to memref<?xf32>
   return %0 : memref<?xf32>
 }

mlir/test/Dialect/SparseTensor/roundtrip.mlir

 // RUN: mlir-opt -split-input-file %s | FileCheck %s
 
 !SparseTensor = type !llvm.ptr<i8>
 
 // CHECK-LABEL: func @sparse_tensor(
 //  CHECK-SAME: %[[A:.*]]: !llvm.ptr<i8>)
-//       CHECK: %[[T:.*]] = linalg.sparse_tensor %[[A]] : !llvm.ptr<i8> to tensor<128xf64>
+//       CHECK: %[[T:.*]] = sparse_tensor.fromPtr %[[A]] : !llvm.ptr<i8> to tensor<128xf64>
 //       CHECK: return %[[T]] : tensor<128xf64>
 func @sparse_tensor(%arg0: !SparseTensor) -> tensor<128xf64> {
-  %0 = linalg.sparse_tensor %arg0 : !SparseTensor to tensor<128xf64>
+  %0 = sparse_tensor.fromPtr %arg0 : !SparseTensor to tensor<128xf64>
   return %0 : tensor<128xf64>
 }
 
 // -----
 
 // CHECK-LABEL: func @sparse_pointers(
 //  CHECK-SAME: %[[A:.*]]: tensor<128xf64>)
 //       CHECK: %[[C:.*]] = constant 1 : index
-//       CHECK: %[[T:.*]] = linalg.sparse_pointers %[[A]], %[[C]] : tensor<128xf64> to memref<?xindex>
+//       CHECK: %[[T:.*]] = sparse_tensor.pointers %[[A]], %[[C]] : tensor<128xf64> to memref<?xindex>
 //       CHECK: return %[[T]] : memref<?xindex>
 func @sparse_pointers(%arg0: tensor<128xf64>) -> memref<?xindex> {
   %c = constant 1 : index
-  %0 = linalg.sparse_pointers %arg0, %c : tensor<128xf64> to memref<?xindex>
+  %0 = sparse_tensor.pointers %arg0, %c : tensor<128xf64> to memref<?xindex>
   return %0 : memref<?xindex>
 }
 
 // -----
 
 // CHECK-LABEL: func @sparse_indices(
 //  CHECK-SAME: %[[A:.*]]: tensor<128xf64>)
 //       CHECK: %[[C:.*]] = constant 1 : index
-//       CHECK: %[[T:.*]] = linalg.sparse_indices %[[A]], %[[C]] : tensor<128xf64> to memref<?xindex>
+//       CHECK: %[[T:.*]] = sparse_tensor.indices %[[A]], %[[C]] : tensor<128xf64> to memref<?xindex>
 //       CHECK: return %[[T]] : memref<?xindex>
 func @sparse_indices(%arg0: tensor<128xf64>) -> memref<?xindex> {
   %c = constant 1 : index
-  %0 = linalg.sparse_indices %arg0, %c : tensor<128xf64> to memref<?xindex>
+  %0 = sparse_tensor.indices %arg0, %c : tensor<128xf64> to memref<?xindex>
   return %0 : memref<?xindex>
 }
 
 // -----
 
 // CHECK-LABEL: func @sparse_values(
 //  CHECK-SAME: %[[A:.*]]: tensor<128xf64>)
-//       CHECK: %[[T:.*]] = linalg.sparse_values %[[A]] : tensor<128xf64> to memref<?xf64>
+//       CHECK: %[[T:.*]] = sparse_tensor.values %[[A]] : tensor<128xf64> to memref<?xf64>
 //       CHECK: return %[[T]] : memref<?xf64>
 func @sparse_values(%arg0: tensor<128xf64>) -> memref<?xf64> {
-  %0 = linalg.sparse_values %arg0 : tensor<128xf64> to memref<?xf64>
+  %0 = sparse_tensor.values %arg0 : tensor<128xf64> to memref<?xf64>
   return %0 : memref<?xf64>
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/frostt-example.mlir

 // RUN: mlir-opt %s \
 // RUN:  -convert-scf-to-std -convert-vector-to-scf \
 // RUN:  -convert-linalg-to-llvm -convert-vector-to-llvm -convert-std-to-llvm | \
 // RUN: TENSOR0="%mlir_integration_test_dir/data/test.tns" \
 // RUN: mlir-cpu-runner \
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 
 //
 // Use descriptive names for opaque pointers.
 //
 !Filename = type !llvm.ptr<i8>
 !Tensor   = type !llvm.ptr<i8>
 
 module {
   //
   // Example of using the sparse runtime support library to read a sparse tensor
   // in the FROSTT file format (http://frostt.io/tensors/file-formats.html).
   //
   func private @getTensorFilename(index) -> (!Filename)
   func private @openTensor(!Filename, memref<?xindex>) -> (!Tensor)
   func private @readTensorItem(!Tensor, memref<?xindex>, memref<?xf64>) -> ()
   func private @closeTensor(!Tensor) -> ()
 
   func @entry() {
     %d0  = constant 0.0 : f64
     %i0  = constant 0   : i64
     %c0  = constant 0   : index
     %c1  = constant 1   : index
     %c2  = constant 2   : index
     %c10 = constant 10  : index
 
     //
     // Setup memrefs to get meta data, indices and values.
     // The index array should provide sufficient space.
     //
     %idata = memref.alloc(%c10) : memref<?xindex>
     %ddata = memref.alloc(%c1)  : memref<?xf64>
 
     //
     // Obtain the sparse tensor filename through this test helper.
     //
     %fileName = call @getTensorFilename(%c0) : (index) -> (!Filename)
 
     //
     // Read a sparse tensor. The call yields a pointer to an opaque
     // memory-resident sparse tensor object that is only understood by
     // other methods in the sparse runtime support library. This call also
     // provides the rank and the number of nonzero elements (nnz) through
     // a memref array.
     //
     %tensor = call @openTensor(%fileName, %idata) : (!Filename, memref<?xindex>) -> (!Tensor)
 
     //
     // Print some meta data.
     //
     %rank = memref.load %idata[%c0] : memref<?xindex>
     %nnz  = memref.load %idata[%c1] : memref<?xindex>
     vector.print %rank : index
     vector.print %nnz  : index
     scf.for %r = %c2 to %c10 step %c1 {
       %d = memref.load %idata[%r] : memref<?xindex>
       vector.print %d : index
     }
 
     //
     // Now we are ready to read in the nonzero elements of the sparse tensor
     // and insert these into a sparse storage scheme. In this example, we
     // simply print the elements on the fly.
     //
     scf.for %k = %c0 to %nnz step %c1 {
       call @readTensorItem(%tensor, %idata, %ddata) : (!Tensor, memref<?xindex>, memref<?xf64>) -> ()
       //
       // Build index vector and print element (here, using the
       // knowledge that the read sparse tensor has rank 8).
       //
       %0 = vector.broadcast %i0 : i64 to vector<8xi64>
       %1 = scf.for %r = %c0 to %rank step %c1 iter_args(%in = %0) -> vector<8xi64> {
         %i  = memref.load %idata[%r] : memref<?xindex>
         %ii = index_cast %i : index to i64
         %ri = index_cast %r : index to i32
         %out = vector.insertelement %ii, %in[%ri : i32] : vector<8xi64>
         scf.yield %out : vector<8xi64>
       }
       %2 = memref.load %ddata[%c0] : memref<?xf64>
       vector.print %1 : vector<8xi64>
       vector.print %2 : f64
     }
 
     //
     // Since at this point we have processed the contents, make sure to
     // close the sparse tensor to release its memory resources.
     //
     call @closeTensor(%tensor) : (!Tensor) -> ()
 
     //
     // Verify that the results are as expected.
     //
     // CHECK: 8
     // CHECK: 16
     // CHECK: 7
     // CHECK: 3
     // CHECK: 3
     // CHECK: 3
     // CHECK: 3
     // CHECK: 3
     // CHECK: 5
     // CHECK: 3
     //
     // CHECK:      ( 0, 0, 0, 0, 0, 0, 0, 0 )
     // CHECK-NEXT: 1
     // CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 0, 2 )
     // CHECK-NEXT: 1.3
     // CHECK-NEXT: ( 0, 0, 0, 0, 0, 0, 4, 0 )
     // CHECK-NEXT: 1.5
     // CHECK-NEXT: ( 0, 0, 0, 1, 0, 0, 0, 1 )
     // CHECK-NEXT: 1.22
     // CHECK-NEXT: ( 0, 0, 0, 1, 0, 0, 0, 2 )
     // CHECK-NEXT: 1.23
     // CHECK-NEXT: ( 1, 0, 1, 0, 1, 1, 1, 0 )
     // CHECK-NEXT: 2.111
     // CHECK-NEXT: ( 1, 0, 1, 0, 1, 1, 1, 2 )
     // CHECK-NEXT: 2.113
     // CHECK-NEXT: ( 1, 1, 1, 0, 1, 1, 1, 0 )
     // CHECK-NEXT: 2.11
     // CHECK-NEXT: ( 1, 1, 1, 0, 1, 1, 1, 1 )
     // CHECK-NEXT: 2.1
     // CHECK-NEXT: ( 1, 1, 1, 1, 1, 1, 1, 1 )
     // CHECK-NEXT: 2
     // CHECK-NEXT: ( 2, 2, 2, 2, 0, 0, 1, 2 )
     // CHECK-NEXT: 3.112
     // CHECK-NEXT: ( 2, 2, 2, 2, 0, 1, 0, 2 )
     // CHECK-NEXT: 3.121
     // CHECK-NEXT: ( 2, 2, 2, 2, 0, 1, 1, 2 )
     // CHECK-NEXT: 3.122
     // CHECK-NEXT: ( 2, 2, 2, 2, 0, 2, 2, 2 )
     // CHECK-NEXT: 3.1
     // CHECK-NEXT: ( 2, 2, 2, 2, 2, 2, 2, 2 )
     // CHECK-NEXT: 3
     // CHECK-NEXT: ( 6, 0, 0, 0, 0, 0, 0, 0 )
     // CHECK-NEXT: 7
     //
 
     //
     // Free.
     //
     memref.dealloc %idata : memref<?xindex>
     memref.dealloc %ddata : memref<?xf64>
 
     return
   }
 }

mlir/test/Integration/Dialect/SparseTensor/CPU/lit.local.cfg

 import sys
 
 # No JIT on win32.
 if sys.platform == 'win32':
     config.unsupported = True

mlir/test/Integration/Dialect/SparseTensor/CPU/matrix-market-example.mlir

 // RUN: mlir-opt %s \
 // RUN:  -convert-scf-to-std -convert-vector-to-scf \
 // RUN:  -convert-linalg-to-llvm -lower-affine -convert-scf-to-std -convert-vector-to-llvm -convert-std-to-llvm | \
 // RUN: TENSOR0="%mlir_integration_test_dir/data/test.mtx" \
 // RUN: mlir-cpu-runner \
 // RUN:  -e entry -entry-point-result=void  \
 // RUN:  -shared-libs=%mlir_integration_test_dir/libmlir_c_runner_utils%shlibext | \
 // RUN: FileCheck %s
 
 //
 // Use descriptive names for opaque pointers.
 //
 !Filename = type !llvm.ptr<i8>
 !Tensor   = type !llvm.ptr<i8>
 
 module {
   //
   // Example of using the sparse runtime support library to read a sparse matrix
   // in the Matrix Market Exchange Format (https://math.nist.gov/MatrixMarket).
   //
   func private @getTensorFilename(index) -> (!Filename)
   func private @openTensor(!Filename, memref<?xindex>) -> (!Tensor)
   func private @readTensorItem(!Tensor, memref<?xindex>, memref<?xf64>) -> ()
   func private @closeTensor(!Tensor) -> ()
 
   func @entry() {
     %d0  = constant 0.0 : f64
     %c0  = constant 0 : index
     %c1  = constant 1 : index
     %c2  = constant 2 : index
     %c3  = constant 3 : index
     %c4  = constant 4 : index
     %c5  = constant 5 : index
 
     //
     // Setup memrefs to get meta data, indices, and values.
     //
     %idata = memref.alloc(%c4) : memref<?xindex>
     %ddata = memref.alloc(%c1) : memref<?xf64>
 
     //
     // Obtain the sparse matrix filename through this test helper.
     //
     %fileName = call @getTensorFilename(%c0) : (index) -> (!Filename)
 
     //
     // Read a sparse matrix. The call yields a pointer to an opaque
     // memory-resident sparse tensor object that is only understood by
     // other methods in the sparse runtime support library. This call also
     // provides the rank (always 2 for the Matrix Market), number of
     // nonzero elements (nnz), and the size (m x n) through a memref array.
     //
     %tensor = call @openTensor(%fileName, %idata) : (!Filename, memref<?xindex>) -> (!Tensor)
     %rank = memref.load %idata[%c0] : memref<?xindex>
     %nnz  = memref.load %idata[%c1] : memref<?xindex>
     %m    = memref.load %idata[%c2] : memref<?xindex>
     %n    = memref.load %idata[%c3] : memref<?xindex>
 
     //
     // At this point, code should prepare a proper sparse storage scheme for
     // an m x n matrix with nnz nonzero elements. For simplicity, here we
     // simply intialize a dense m x n matrix to all zeroes.
     //
     %a = memref.alloc(%m, %n) : memref<?x?xf64>
     scf.for %ii = %c0 to %m step %c1 {
       scf.for %jj = %c0 to %n step %c1 {
         memref.store %d0, %a[%ii, %jj] : memref<?x?xf64>
       }
     }
 
     //
     // Now we are ready to read in nnz nonzero elements of the sparse matrix
     // and insert these into a sparse storage scheme. In this example, we
     // simply insert them in the dense matrix.
     //
     scf.for %k = %c0 to %nnz step %c1 {
       call @readTensorItem(%tensor, %idata, %ddata) : (!Tensor, memref<?xindex>, memref<?xf64>) -> ()
       %i = memref.load %idata[%c0] : memref<?xindex>
       %j = memref.load %idata[%c1] : memref<?xindex>
       %d = memref.load %ddata[%c0] : memref<?xf64>
       memref.store %d, %a[%i, %j] : memref<?x?xf64>
     }
 
     //
     // Since at this point we have copied the sparse matrix to our own
     // storage scheme, make sure to close the matrix to release its
     // memory resources.
     //
     call @closeTensor(%tensor) : (!Tensor) -> ()
 
     //
     // Verify that the results are as expected.
     //
     %A = vector.transfer_read %a[%c0, %c0], %d0 : memref<?x?xf64>, vector<5x5xf64>
     vector.print %rank : index
     vector.print %nnz  : index
     vector.print %m    : index
     vector.print %n    : index
     vector.print %A    : vector<5x5xf64>
     //
     // CHECK: 2
     // CHECK: 9
     // CHECK: 5
     // CHECK: 5
     //
     // CHECK:      ( ( 1, 0, 0, 1.4, 0 ),
     // CHECK-SAME:   ( 0, 2, 0, 0, 2.5 ),
     // CHECK-SAME:   ( 0, 0, 3, 0, 0 ),
     // CHECK-SAME:   ( 4.1, 0, 0, 4, 0 ),
     // CHECK-SAME:   ( 0, 5.2, 0, 0, 5 ) )
 
     //
     // Free.
     //
     memref.dealloc %idata : memref<?xindex>
     memref.dealloc %ddata : memref<?xf64>
     memref.dealloc %a     : memref<?x?xf64>
 
     return
   }
 }

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

 module {
   //
   // The kernel expressed as an annotated Linalg op. The kernel multiplies
   // a sparse matrix A with a dense vector b into a dense vector x.
   //
   func @kernel_matvec(%argA: !SparseTensor,
                       %argb: tensor<?xi32>,
                       %argx: tensor<?xi32>) -> tensor<?xi32> {
-    %arga = linalg.sparse_tensor %argA : !SparseTensor to tensor<?x?xi32>
+    %arga = sparse_tensor.fromPtr %argA : !SparseTensor to tensor<?x?xi32>
     %0 = linalg.generic #matvec
       ins(%arga, %argb: tensor<?x?xi32>, tensor<?xi32>)
       outs(%argx: tensor<?xi32>) {
       ^bb(%a: i32, %b: i32, %x: i32):
         %0 = muli %a, %b : i32
         %1 = addi %x, %0 : i32
         linalg.yield %1 : i32
     } -> tensor<?xi32>

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

module {
   //
   // The kernel expressed as an annotated Linalg op. The kernel
   // computes a sampled matrix matrix multiplication.
   //
   func @sampled_dense_dense(%argS: !SparseTensor,
                             %arga: tensor<?x?xf32>,
                             %argb: tensor<?x?xf32>,
                             %argx: tensor<?x?xf32>) -> tensor<?x?xf32> {
-    %args = linalg.sparse_tensor %argS : !SparseTensor to tensor<?x?xf32>
+    %args = sparse_tensor.fromPtr %argS : !SparseTensor to tensor<?x?xf32>
     %0 = linalg.generic #trait_sampled_dense_dense
       ins(%args, %arga, %argb: tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>)
       outs(%argx: tensor<?x?xf32>) {
         ^bb(%s: f32, %a: f32, %b: f32, %x: f32):
           %0 = mulf %a, %b : f32
           %1 = mulf %s, %0 : f32
           %2 = addf %x, %1 : f32
           linalg.yield %2 : f32

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

 //
 module {
   //
   // The kernel expressed as an annotated Linalg op. The kernel
   // sum reduces a matrix to a single scalar.
   //
   func @kernel_sum_reduce(%argA: !SparseTensor,
                           %argx: tensor<f64>) -> tensor<f64> {
-    %arga = linalg.sparse_tensor %argA : !SparseTensor to tensor<?x?xf64>
+    %arga = sparse_tensor.fromPtr %argA : !SparseTensor to tensor<?x?xf64>
     %0 = linalg.generic #trait_sum_reduce
       ins(%arga: tensor<?x?xf64>)
       outs(%argx: tensor<f64>) {
       ^bb(%a: f64, %x: f64):
         %0 = addf %x, %a : f64
         linalg.yield %0 : f64
     } -> tensor<f64>
     return %0 : tensor<f64>

mlir/test/lib/Transforms/CMakeLists.txt

add_mlir_library(MLIRTestTransforms
   MLIRNVVMIR
   MLIRNVVMToLLVMIRTranslation
   MLIRPass
   MLIRROCDLIR
   MLIRROCDLToLLVMIRTranslation
   MLIRSCF
   MLIRSCFTransforms
   MLIRStandardOpsTransforms
+  MLIRSparseTensor
+  MLIRSparseTensorTransforms
   MLIRTargetLLVMIRExport
   MLIRTestDialect
   MLIRTransformUtils
   MLIRVector
   MLIRVectorToSCF
   )
 
 include_directories(${CMAKE_CURRENT_SOURCE_DIR}/../Dialect/Test)
 include_directories(${CMAKE_CURRENT_BINARY_DIR}/../Dialect/Test)

mlir/test/lib/Transforms/TestSparsification.cpp

 //===- TestSparsification.cpp - Test sparsification of tensors ------------===//
 //
 // 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/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/Linalg/Transforms/Transforms.h"
+#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
+#include "mlir/Dialect/SparseTensor/Transforms/Transforms.h"
 #include "mlir/Dialect/Vector/VectorOps.h"
 #include "mlir/Pass/Pass.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 
 using namespace mlir;
 
 namespace {
 
Option<bool> fastOutput{*this, "fast-output",
                           llvm::cl::init(false)};
 
   Option<bool> lower{*this, "lower", llvm::cl::desc("Lower sparse primitives"),
                      llvm::cl::init(false)};
 
   /// Registers all dialects required by testing.
   void getDependentDialects(DialectRegistry &registry) const override {
     registry.insert<memref::MemRefDialect, scf::SCFDialect,
-                    vector::VectorDialect, LLVM::LLVMDialect>();
+                    sparse_tensor::SparseTensorDialect, vector::VectorDialect,
+                    LLVM::LLVMDialect>();
   }
 
   /// Returns parallelization strategy given on command line.
   linalg::SparseParallelizationStrategy parallelOption() {
     switch (parallelization) {
     default:
       return linalg::SparseParallelizationStrategy::kNone;
     case 1:
void runOnOperation() override {
     // Apply rewriting.
     linalg::populateSparsificationPatterns(patterns, options);
     vector::populateVectorToVectorCanonicalizationPatterns(patterns);
     (void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
     // Lower sparse primitives to calls into runtime support library.
     if (lower) {
       RewritePatternSet conversionPatterns(ctx);
       ConversionTarget target(*ctx);
-      target.addIllegalOp<linalg::SparseTensorFromPointerOp,
-                          linalg::SparseTensorToPointersMemRefOp,
-                          linalg::SparseTensorToIndicesMemRefOp,
-                          linalg::SparseTensorToValuesMemRefOp>();
+      target.addIllegalOp<
+          sparse_tensor::FromPointerOp, sparse_tensor::ToPointersOp,
+          sparse_tensor::ToIndicesOp, sparse_tensor::ToValuesOp>();
       target.addLegalOp<CallOp>();
-      linalg::populateSparsificationConversionPatterns(conversionPatterns);
+      sparse_tensor::populateSparsificationConversionPatterns(
+          conversionPatterns);
       if (failed(applyPartialConversion(getOperation(), target,
                                         std::move(conversionPatterns))))
         signalPassFailure();
     }
   }
 };
 
 } // end anonymous namespace

mlir/test/mlir-opt/commandline.mlir

 // CHECK-NEXT: omp
 // CHECK-NEXT: pdl
 // CHECK-NEXT: pdl_interp
 // CHECK-NEXT: quant
 // CHECK-NEXT: rocdl
 // CHECK-NEXT: scf
 // CHECK-NEXT: sdbm
 // CHECK-NEXT: shape
+// CHECK-NEXT: sparse
 // CHECK-NEXT: spv
 // CHECK-NEXT: std
 // CHECK-NEXT: tensor
 // CHECK-NEXT: test
 // CHECK-NEXT: tosa
 // CHECK-NEXT: vector
 // CHECK-NEXT: x86vector