mirror of https://github.com/llvm/circt.git
0f8f616fec
Operation-based canonicalization is needed because the unrolled `affine.parallel` with `scf.execute_region`s in it is semantically incorrect - all the `scf.execute_region`s would be inlined if a generic `canonicalization` pass were to be applied. |
||
|---|---|---|
| .. | ||
| affine-parallel-unroll.mlir | ||
| canonicalization.mlir | ||
| clk-insertion.mlir | ||
| compile-control.mlir | ||
| emit-comb-component.mlir | ||
| emit-custom-primitives.mlir | ||
| emit-invoke-with-ref.mlir | ||
| emit-invoke.mlir | ||
| emit-seq-memory.mlir | ||
| emit-static.mlir | ||
| emit.mlir | ||
| errors.mlir | ||
| gicm.mlir | ||
| go-insertion.mlir | ||
| remove-comb-groups.mlir | ||
| remove-groups.mlir | ||
| round-trip.mlir | ||