Missed optimization: (vector.reduce.add (splat %0, 4)) ==> shl i32 %0, 2

[GINN-Imp]

The following reduced IR is derived from https://github.com/berkeley-abc/abc/blob/b28e042afdf0a869f3ececbe054a60e02fcd2ffc/src/bool/bdc/bdcSpfd.c#L83

Godbolt: https://godbolt.org/z/z9frEof9E
alive2 proof: https://alive2.llvm.org/ce/z/ULwEjv

define range(i32 0, 7) i32 @Bdc_SpfdAdjCost(i32 %0) local_unnamed_addr #0 {
  %2 = insertelement <4 x i32> poison, i32 %0, i64 0
  %3 = shufflevector <4 x i32> %2, <4 x i32> poison, <4 x i32> zeroinitializer
  %4 = tail call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %3)
  ret i32 %4
}

expected:

define range(i32 0, 7) i32 @tgt(i32 %0) local_unnamed_addr #0 {
  %2 = shl i32 %0, 2
  ret i32 %2
}