[mlir][sparse][NFC] Switch InitOp to bufferization::AllocTensorOp

Now that we have an AllocTensorOp (previously InitTensorOp) in the bufferization dialect, the InitOp in the sparse dialect is no longer needed.

Differential Revision: https://reviews.llvm.org/D126180
This commit is contained in:
Matthias Springer
2022-06-01 23:59:58 +02:00
parent 902360d21d
commit 6232a8f3d6
34 changed files with 117 additions and 200 deletions

View File

@@ -119,6 +119,7 @@ struct SparseTensorConversionPass
target
.addLegalDialect<bufferization::BufferizationDialect, LLVM::LLVMDialect,
memref::MemRefDialect, scf::SCFDialect>();
target.addIllegalOp<bufferization::AllocTensorOp>();
// Translate strategy flags to strategy options.
SparseTensorConversionOptions options(
sparseToSparseConversionStrategy(sparseToSparse));