[sparse] Make GenericOpSparsifier not crash on multi-output dense linalg.generic

[bkramer]

Summary

The actual transformation doesn't support multi-output GenericOps, but
if we encounter one without sparse annotations we can just leave it
alone.

[aartbik]

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

1857

how about,

if (op.getNumOutputs() != 1)

return failure();

so we don't enter the annotation analysis at all?

[bkramer]

Fail gracefully instead of asserting

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

 public:
   GenericOpSparsifier(MLIRContext *context, SparsificationOptions o)
       : OpRewritePattern<linalg::GenericOp>(context), options(o) {}
 
   LogicalResult matchAndRewrite(linalg::GenericOp op,
                                 PatternRewriter &rewriter) const override {
     // Detects sparse annotations and translate the per-dimension sparsity
     // information for all tensors to loop indices in the kernel.
-    assert(op.getNumOutputs() == 1);
     unsigned numTensors = op.getNumInputsAndOutputs();
     unsigned numLoops = op.iterator_types().getValue().size();
     Merger merger(numTensors, numLoops);
     if (!findSparseAnnotations(merger, op))
       return failure();
 
+    assert(op.getNumOutputs() == 1);
     // Builds the tensor expression for the Linalg operation in SSA form.
     Optional<unsigned> optExp = merger.buildTensorExpFromLinalg(op);
     if (!optExp.has_value())
       return failure();
 
     unsigned exp = optExp.value();
     OpOperand *sparseOut = nullptr;
     unsigned outerParNest = 0;