[PATCH] D88569: [DAGCombiner] Call SimplifyDemandedBits to simplify EXTRACT_VECTOR_ELT

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: llvm/test/CodeGen/AMDGPU/sdiv64.ll:502
 ; GCN-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x9
-; GCN-NEXT:    s_load_dword s1, s[0:1], 0xe
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[0:1], 0xd
 ; GCN-NEXT:    s_mov_b32 s3, 0xf000
----------------
Is this a regression? It looks like we're lost track that we only need 1 element


================
Comment at: llvm/test/CodeGen/ARM/func-argpassing-endian.ll:106
+; CHECK-LE-NEXT:    vldr d16, .LCPI6_0
+; CHECK-LE-NEXT:    vldr d17, .LCPI6_1
 ; CHECK-LE-NEXT:    vmov r0, r1, d16
----------------
regression? we're no longer doing a single multiple load


================
Comment at: llvm/test/CodeGen/ARM/vdup.ll:59-68
+; CHECK-NEXT:    vmov.i32 d17, #0x0
+; CHECK-NEXT:    vdup.8 d16, r0
+; CHECK-NEXT:    vmov.8 d17[0], r0
+; CHECK-NEXT:    vmov.8 d17[1], r0
+; CHECK-NEXT:    vmov.8 d17[2], r0
+; CHECK-NEXT:    vmov.8 d17[3], r0
+; CHECK-NEXT:    vmov.8 d17[4], r0
----------------
foad wrote:
> Regression in lots of cases in this file.
poor duplicate/splat detection?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D88569/new/

https://reviews.llvm.org/D88569