[mlir][sparse] Only insert non-zero values to the result of the concatenate operation.

[bixia]

Summary

Modify the integration test to check number_of_entries and use it to limit for
outputing sparse tensor values.

[bixia]

Fix a comment.

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

static void genSparseCOOIterationLoop(
   SmallVector<Type> noTypes;
   auto whileOp = rewriter.create<scf::WhileOp>(loc, noTypes, noArgs);
   Block *before = rewriter.createBlock(&whileOp.getBefore(), {}, noTypes);
   rewriter.setInsertionPointToEnd(before);
   Value cond = genGetNextCall(rewriter, loc, iter, srcIdx, elemPtr);
   rewriter.create<scf::ConditionOp>(loc, cond, before->getArguments());
   Block *after = rewriter.createBlock(&whileOp.getAfter(), {}, noTypes);
   rewriter.setInsertionPointToStart(after);
+
+  bool hasDenseDim = llvm::any_of(
+      enc.getDimLevelType(), [](DimLevelType dlt) { return isDenseDLT(dlt); });
+  if (hasDenseDim) {
+    Value elemV = rewriter.create<memref::LoadOp>(loc, elemPtr);
+    Value isZero = genIsNonzero(rewriter, loc, elemV);
+    scf::IfOp ifOp = rewriter.create<scf::IfOp>(loc, isZero, /*else*/ false);
+    rewriter.setInsertionPointToStart(&ifOp.getThenRegion().front());
+  }
   // Callback here to build loop body.
   bodyBuilder(rewriter, loc, srcIdx, elemPtr);
+
+  // Exit the scope from the IfOp.
+  if (hasDenseDim)
+    rewriter.setInsertionPointToEnd(after);
+
   rewriter.create<scf::YieldOp>(loc);
   // Finish generating loop.
   rewriter.setInsertionPointAfter(whileOp);
 
   // Free memory for iterator.
   genDelIteratorCall(rewriter, loc, elemTp, iter);
 }
 

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

module {
   func.func @dump_mat_9x4(%A: tensor<9x4xf64, #MAT_C_C>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %c = sparse_tensor.convert %A : tensor<9x4xf64, #MAT_C_C> to tensor<9x4xf64>
     %v = vector.transfer_read %c[%c0, %c0], %du: tensor<9x4xf64>, vector<9x4xf64>
     vector.print %v : vector<9x4xf64>
 
+    %n = sparse_tensor.number_of_entries %A : tensor<9x4xf64, #MAT_C_C>
+    vector.print %n : index
+
     %1 = sparse_tensor.values %A : tensor<9x4xf64, #MAT_C_C> to memref<?xf64>
-    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<36xf64>
-    vector.print %2 : vector<36xf64>
+    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<18xf64>
+    vector.print %2 : vector<18xf64>
 
     return
   }
 
   func.func @dump_mat_perm_9x4(%A: tensor<9x4xf64, #MAT_C_C_P>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %c = sparse_tensor.convert %A : tensor<9x4xf64, #MAT_C_C_P> to tensor<9x4xf64>
     %v = vector.transfer_read %c[%c0, %c0], %du: tensor<9x4xf64>, vector<9x4xf64>
     vector.print %v : vector<9x4xf64>
 
+    %n = sparse_tensor.number_of_entries %A : tensor<9x4xf64, #MAT_C_C_P>
+    vector.print %n : index
+
     %1 = sparse_tensor.values %A : tensor<9x4xf64, #MAT_C_C_P> to memref<?xf64>
-    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<36xf64>
-    vector.print %2 : vector<36xf64>
+    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<18xf64>
+    vector.print %2 : vector<18xf64>
 
     return
   }
 
   func.func @dump_mat_dense_9x4(%A: tensor<9x4xf64>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %v = vector.transfer_read %A[%c0, %c0], %du: tensor<9x4xf64>, vector<9x4xf64>
     vector.print %v : vector<9x4xf64>
 
     return
   }
 
   func.func @dump_mat_4x9(%A: tensor<4x9xf64, #MAT_C_C>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %c = sparse_tensor.convert %A : tensor<4x9xf64, #MAT_C_C> to tensor<4x9xf64>
     %v = vector.transfer_read %c[%c0, %c0], %du: tensor<4x9xf64>, vector<4x9xf64>
     vector.print %v : vector<4x9xf64>
 
+    %n = sparse_tensor.number_of_entries %A : tensor<4x9xf64, #MAT_C_C>
+    vector.print %n : index
+
     %1 = sparse_tensor.values %A : tensor<4x9xf64, #MAT_C_C> to memref<?xf64>
-    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<36xf64>
-    vector.print %2 : vector<36xf64>
+    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<18xf64>
+    vector.print %2 : vector<18xf64>
 
     return
   }
 
   func.func @dump_mat_dyn(%A: tensor<?x?xf64, #MAT_C_C>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %c = sparse_tensor.convert %A : tensor<?x?xf64, #MAT_C_C> to tensor<?x?xf64>
     %v = vector.transfer_read %c[%c0, %c0], %du: tensor<?x?xf64>, vector<4x9xf64>
     vector.print %v : vector<4x9xf64>
 
+    %n = sparse_tensor.number_of_entries %A : tensor<?x?xf64, #MAT_C_C>
+    vector.print %n : index
+
     %1 = sparse_tensor.values %A : tensor<?x?xf64, #MAT_C_C> to memref<?xf64>
-    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<36xf64>
-    vector.print %2 : vector<36xf64>
+    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<18xf64>
+    vector.print %2 : vector<18xf64>
 
     return
   }
 
   func.func @dump_mat_perm_4x9(%A: tensor<4x9xf64, #MAT_C_C_P>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
     %c = sparse_tensor.convert %A : tensor<4x9xf64, #MAT_C_C_P> to tensor<4x9xf64>
     %v = vector.transfer_read %c[%c0, %c0], %du: tensor<4x9xf64>, vector<4x9xf64>
     vector.print %v : vector<4x9xf64>
 
+    %n = sparse_tensor.number_of_entries %A : tensor<4x9xf64, #MAT_C_C_P>
+    vector.print %n : index
+
     %1 = sparse_tensor.values %A : tensor<4x9xf64, #MAT_C_C_P> to memref<?xf64>
-    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<36xf64>
-    vector.print %2 : vector<36xf64>
+    %2 = vector.transfer_read %1[%c0], %du: memref<?xf64>, vector<18xf64>
+    vector.print %2 : vector<18xf64>
 
     return
   }
 
   func.func @dump_mat_dense_4x9(%A: tensor<4x9xf64>) {
     %c0 = arith.constant 0 : index
     %du = arith.constant -1.0 : f64
 
func.func @entry() {
     %sm42ccp = sparse_tensor.convert %m42 : tensor<4x2xf64> to tensor<4x2xf64, #MAT_C_C_P>
     %sm43cdp = sparse_tensor.convert %m43 : tensor<4x3xf64> to tensor<4x3xf64, #MAT_C_D_P>
     %sm44dcp = sparse_tensor.convert %m44 : tensor<4x4xf64> to tensor<4x4xf64, #MAT_D_C_P>
 
     %sm43cd_dyn = sparse_tensor.convert %m43 : tensor<4x3xf64> to tensor<?x?xf64, #MAT_C_D>
     %sm44dc_dyn = sparse_tensor.convert %m44 : tensor<4x4xf64> to tensor<?x?xf64, #MAT_D_C>
 
     // CHECK:    ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 3, 2, 1, 0, 1, 1, 0, 0.5, 0, 0, 1, 5, 2, 0, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 3, 2, 1, 1, 1, 0.5, 1, 5, 2, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5 )
     %0 = call @concat_sparse_sparse(%sm24cc, %sm34cd, %sm44dc)
                : (tensor<2x4xf64, #MAT_C_C>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64, #MAT_C_C>
     call @dump_mat_9x4(%0) : (tensor<9x4xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
     %1 = call @concat_sparse_dense(%sm24cc, %sm34cd, %sm44dc)
                : (tensor<2x4xf64, #MAT_C_C>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64>
     call @dump_mat_dense_9x4(%1) : (tensor<9x4xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 3, 2, 1, 0, 1, 1, 0, 0.5, 0, 0, 1, 5, 2, 0, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 3, 2, 1, 1, 1, 0.5, 1, 5, 2, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5 )
     %2 = call @concat_mix_sparse(%m24, %sm34cd, %sm44dc)
                : (tensor<2x4xf64>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64, #MAT_C_C>
     call @dump_mat_9x4(%2) : (tensor<9x4xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
     %3 = call @concat_mix_dense(%m24, %sm34cd, %sm44dc)
                : (tensor<2x4xf64>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64>
     call @dump_mat_dense_9x4(%3) : (tensor<9x4xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 1, 0, 1, 1, 1, 2, 0, 0.5, 5, 3.5, 5, 0.5, 3, 1, 0, 2, 1.5, 2, 1, 0, 0, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 1, 1, 1, 1, 2, 0.5, 5, 3.5, 5, 0.5, 3, 1, 2, 1.5, 2, 1, 1 )
     %4 = call @concat_sparse_sparse_perm(%sm24ccp, %sm34cd, %sm44dc)
                : (tensor<2x4xf64, #MAT_C_C_P>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64, #MAT_C_C_P>
     call @dump_mat_perm_9x4(%4) : (tensor<9x4xf64, #MAT_C_C_P>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
     %5 = call @concat_sparse_dense_perm(%sm24ccp, %sm34cdp, %sm44dc)
                : (tensor<2x4xf64, #MAT_C_C_P>, tensor<3x4xf64, #MAT_C_D_P>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64>
     call @dump_mat_dense_9x4(%5) : (tensor<9x4xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 3, 2, 1, 0, 1, 1, 0, 0.5, 0, 0, 1, 5, 2, 0, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 3, 2, 1, 1, 1, 0.5, 1, 5, 2, 1.5, 1, 3.5, 1, 5, 2, 1, 0.5 )
     %6 = call @concat_mix_sparse_perm(%m24, %sm34cdp, %sm44dc)
                : (tensor<2x4xf64>, tensor<3x4xf64, #MAT_C_D_P>, tensor<4x4xf64, #MAT_D_C>) -> tensor<9x4xf64, #MAT_C_C>
     call @dump_mat_9x4(%6) : (tensor<9x4xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 3, 0 ), ( 0, 2, 0, 0 ), ( 1, 0, 1, 1 ), ( 0, 0.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 0, 0, 1.5, 1 ), ( 0, 3.5, 0, 0 ), ( 1, 5, 2, 0 ), ( 1, 0.5, 0, 0 ) )
     %7 = call @concat_mix_dense_perm(%m24, %sm34cd, %sm44dcp)
                : (tensor<2x4xf64>, tensor<3x4xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C_P>) -> tensor<9x4xf64>
     call @dump_mat_dense_9x4(%7) : (tensor<9x4xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 1, 0, 1, 1.5, 1, 3.1, 1, 0, 0.5, 3.5, 2, 0, 0, 1, 1, 5, 2, 5, 2, 0, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 1, 1, 1.5, 1, 3.1, 1, 0.5, 3.5, 2, 1, 1, 5, 2, 5, 2, 1, 0.5 )
     %8 = call @concat_sparse_sparse_dim1(%sm42cc, %sm43cd, %sm44dc)
                : (tensor<4x2xf64, #MAT_C_C>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64, #MAT_C_C>
     call @dump_mat_4x9(%8) : (tensor<4x9xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
     %9 = call @concat_sparse_dense_dim1(%sm42cc, %sm43cd, %sm44dc)
                : (tensor<4x2xf64, #MAT_C_C>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64>
     call @dump_mat_dense_4x9(%9) : (tensor<4x9xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 1, 0, 1, 1.5, 1, 3.1, 1, 0, 0.5, 3.5, 2, 0, 0, 1, 1, 5, 2, 5, 2, 0, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 1, 1, 1.5, 1, 3.1, 1, 0.5, 3.5, 2, 1, 1, 5, 2, 5, 2, 1, 0.5 )
     %10 = call @concat_mix_sparse_dim1(%m42, %sm43cd, %sm44dc)
                : (tensor<4x2xf64>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64, #MAT_C_C>
     call @dump_mat_4x9(%10) : (tensor<4x9xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
     %11 = call @concat_mix_dense_dim1(%m42, %sm43cd, %sm44dc)
                : (tensor<4x2xf64>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64>
     call @dump_mat_dense_4x9(%11) : (tensor<4x9xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 3.1, 2, 1, 1, 0, 5, 0, 0, 0, 2, 1, 0.5, 1, 0, 1, 1, 3.5, 5, 0.5, 1.5, 2, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 3.1, 2, 1, 1, 5, 2, 1, 0.5, 1, 1, 1, 3.5, 5, 0.5, 1.5, 2, 1 )
     %12 = call @concat_sparse_sparse_perm_dim1(%sm42ccp, %sm43cd, %sm44dc)
                : (tensor<4x2xf64, #MAT_C_C_P>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64, #MAT_C_C_P>
     call @dump_mat_perm_4x9(%12) : (tensor<4x9xf64, #MAT_C_C_P>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
     %13 = call @concat_sparse_dense_perm_dim1(%sm42ccp, %sm43cdp, %sm44dc)
                : (tensor<4x2xf64, #MAT_C_C_P>, tensor<4x3xf64, #MAT_C_D_P>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64>
     call @dump_mat_dense_4x9(%13) : (tensor<4x9xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 1, 0, 1, 1.5, 1, 3.1, 1, 0, 0.5, 3.5, 2, 0, 0, 1, 1, 5, 2, 5, 2, 0, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 1, 1, 1.5, 1, 3.1, 1, 0.5, 3.5, 2, 1, 1, 5, 2, 5, 2, 1, 0.5 )
     %14 = call @concat_mix_sparse_perm_dim1(%m42, %sm43cdp, %sm44dc)
                : (tensor<4x2xf64>, tensor<4x3xf64, #MAT_C_D_P>, tensor<4x4xf64, #MAT_D_C>) -> tensor<4x9xf64, #MAT_C_C>
     call @dump_mat_4x9(%14) : (tensor<4x9xf64, #MAT_C_C>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
     %15 = call @concat_mix_dense_perm_dim1(%m42, %sm43cd, %sm44dcp)
                : (tensor<4x2xf64>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C_P>) -> tensor<4x9xf64>
     call @dump_mat_dense_4x9(%15) : (tensor<4x9xf64>) -> ()
 
     // CHECK-NEXT: ( ( 1, 0, 1, 0, 1, 0, 0, 1.5, 1 ), ( 3.1, 0, 1, 0, 0.5, 0, 3.5, 0, 0 ), ( 0, 2, 0, 0, 1, 1, 5, 2, 0 ), ( 0, 0, 5, 2, 0, 1, 0.5, 0, 0 ) )
-    // CHECK-NEXT: ( 1, 1, 0, 1, 1.5, 1, 3.1, 1, 0, 0.5, 3.5, 2, 0, 0, 1, 1, 5, 2, 5, 2, 0, 1, 0.5, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 )
+    // CHECK-NEXT: 18
+    // CHECK-NEXT: ( 1, 1, 1, 1.5, 1, 3.1, 1, 0.5, 3.5, 2, 1, 1, 5, 2, 5, 2, 1, 0.5 )
     %16 = call @concat_mix_sparse_dyn(%m42, %sm43cd, %sm44dc)
                : (tensor<4x2xf64>, tensor<4x3xf64, #MAT_C_D>, tensor<4x4xf64, #MAT_D_C>) -> tensor<?x?xf64, #MAT_C_C>
     call @dump_mat_dyn(%16) : (tensor<?x?xf64, #MAT_C_C>) -> ()
 
     // Release resources.
     bufferization.dealloc_tensor %sm24cc  : tensor<2x4xf64, #MAT_C_C>
     bufferization.dealloc_tensor %sm34cd  : tensor<3x4xf64, #MAT_C_D>
     bufferization.dealloc_tensor %sm42cc  : tensor<4x2xf64, #MAT_C_C>