[PATCH] D127595: [DAG] visitINSERT_VECTOR_ELT - attempt to reconstruct BUILD_VECTOR before other fold interfere

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-insert-vector-elt.ll:232
+; VBITS_GE_256-NEXT:    mov x8, #4617315517961601024
+; VBITS_GE_256-NEXT:    fmov d0, x8
 ; VBITS_GE_256-NEXT:    ret
----------------
dmgreen wrote:
> RKSimon wrote:
> > Minor aarch64 optimization - insertelement v1f64 folds to a move but buildvector doesn't?
> Should a single element buildvector be canonicalized to a scalar_to_reg?
I tested with folding to scalar_to_vector and this test had the same problem.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D127595/new/

https://reviews.llvm.org/D127595