[VectorCombine] foldShuffleChainsToReduce - add support for partial vector reductions

[RKSimon]

define i16 @test_reduce_v16i16_v8i16(<16 x i16> %a0) {
; CHECK-LABEL: @test_reduce_v16i16_v8i16(
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <16 x i16> [[A0:%.*]], <16 x i16> poison, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
; CHECK-NEXT:    [[TMP2:%.*]] = tail call <16 x i16> @llvm.smax.v16i16(<16 x i16> [[A0]], <16 x i16> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <16 x i16> [[TMP2]], <16 x i16> poison, <16 x i32> <i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
; CHECK-NEXT:    [[TMP4:%.*]] = tail call <16 x i16> @llvm.smax.v16i16(<16 x i16> [[TMP2]], <16 x i16> [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <16 x i16> [[TMP4]], <16 x i16> poison, <16 x i32> <i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
; CHECK-NEXT:    [[TMP6:%.*]] = tail call <16 x i16> @llvm.smax.v16i16(<16 x i16> [[TMP4]], <16 x i16> [[TMP5]])
; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <16 x i16> [[TMP6]], i64 0
; CHECK-NEXT:    ret i16 [[TMP10]]
;
  %1  = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %2  = icmp sgt <16 x i16> %a0, %1
  %3  = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %1
  %4  = shufflevector <16 x i16> %3, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5  = icmp sgt <16 x i16> %3, %4
  %6  = select <16 x i1> %5, <16 x i16> %3, <16 x i16> %4
  %7  = shufflevector <16 x i16> %6, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %8  = icmp sgt <16 x i16> %6, %7
  %9  = select <16 x i1> %8, <16 x i16> %6, <16 x i16> %7
  %10 = extractelement <16 x i16> %9, i32 0
  ret i16 %10
}

we should be able to recognise that this is a partial reduction of just the lower subvector:

define i16 @test_reduce_v16i16_v8i16(<16 x i16> %a0) {
; CHECK-LABEL: @test_reduce_v16i16_v8i16(
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <16 x i16> [[A0:%.*]], <16 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6 i32 7>
; CHECK-NEXT:    [[TMP2:%.*]] = tail call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> [[TMP1]])
; CHECK-NEXT:    ret i16 [[TMP2]]
;
  %1  = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %2  = icmp sgt <16 x i16> %a0, %1
  %3  = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %1
  %4  = shufflevector <16 x i16> %3, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5  = icmp sgt <16 x i16> %3, %4
  %6  = select <16 x i1> %5, <16 x i16> %3, <16 x i16> %4
  %7  = shufflevector <16 x i16> %6, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %8  = icmp sgt <16 x i16> %6, %7
  %9  = select <16 x i1> %8, <16 x i16> %6, <16 x i16> %7
  %10 = extractelement <16 x i16> %9, i32 0
  ret i16 %10
}