1 ; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
3 define <8 x i8> @v_dup8(i8 %A) nounwind {
6 %tmp1 = insertelement <8 x i8> zeroinitializer, i8 %A, i32 0
7 %tmp2 = insertelement <8 x i8> %tmp1, i8 %A, i32 1
8 %tmp3 = insertelement <8 x i8> %tmp2, i8 %A, i32 2
9 %tmp4 = insertelement <8 x i8> %tmp3, i8 %A, i32 3
10 %tmp5 = insertelement <8 x i8> %tmp4, i8 %A, i32 4
11 %tmp6 = insertelement <8 x i8> %tmp5, i8 %A, i32 5
12 %tmp7 = insertelement <8 x i8> %tmp6, i8 %A, i32 6
13 %tmp8 = insertelement <8 x i8> %tmp7, i8 %A, i32 7
17 define <4 x i16> @v_dup16(i16 %A) nounwind {
20 %tmp1 = insertelement <4 x i16> zeroinitializer, i16 %A, i32 0
21 %tmp2 = insertelement <4 x i16> %tmp1, i16 %A, i32 1
22 %tmp3 = insertelement <4 x i16> %tmp2, i16 %A, i32 2
23 %tmp4 = insertelement <4 x i16> %tmp3, i16 %A, i32 3
27 define <2 x i32> @v_dup32(i32 %A) nounwind {
30 %tmp1 = insertelement <2 x i32> zeroinitializer, i32 %A, i32 0
31 %tmp2 = insertelement <2 x i32> %tmp1, i32 %A, i32 1
35 define <2 x float> @v_dupfloat(float %A) nounwind {
38 %tmp1 = insertelement <2 x float> zeroinitializer, float %A, i32 0
39 %tmp2 = insertelement <2 x float> %tmp1, float %A, i32 1
43 define <16 x i8> @v_dupQ8(i8 %A) nounwind {
46 %tmp1 = insertelement <16 x i8> zeroinitializer, i8 %A, i32 0
47 %tmp2 = insertelement <16 x i8> %tmp1, i8 %A, i32 1
48 %tmp3 = insertelement <16 x i8> %tmp2, i8 %A, i32 2
49 %tmp4 = insertelement <16 x i8> %tmp3, i8 %A, i32 3
50 %tmp5 = insertelement <16 x i8> %tmp4, i8 %A, i32 4
51 %tmp6 = insertelement <16 x i8> %tmp5, i8 %A, i32 5
52 %tmp7 = insertelement <16 x i8> %tmp6, i8 %A, i32 6
53 %tmp8 = insertelement <16 x i8> %tmp7, i8 %A, i32 7
54 %tmp9 = insertelement <16 x i8> %tmp8, i8 %A, i32 8
55 %tmp10 = insertelement <16 x i8> %tmp9, i8 %A, i32 9
56 %tmp11 = insertelement <16 x i8> %tmp10, i8 %A, i32 10
57 %tmp12 = insertelement <16 x i8> %tmp11, i8 %A, i32 11
58 %tmp13 = insertelement <16 x i8> %tmp12, i8 %A, i32 12
59 %tmp14 = insertelement <16 x i8> %tmp13, i8 %A, i32 13
60 %tmp15 = insertelement <16 x i8> %tmp14, i8 %A, i32 14
61 %tmp16 = insertelement <16 x i8> %tmp15, i8 %A, i32 15
65 define <8 x i16> @v_dupQ16(i16 %A) nounwind {
68 %tmp1 = insertelement <8 x i16> zeroinitializer, i16 %A, i32 0
69 %tmp2 = insertelement <8 x i16> %tmp1, i16 %A, i32 1
70 %tmp3 = insertelement <8 x i16> %tmp2, i16 %A, i32 2
71 %tmp4 = insertelement <8 x i16> %tmp3, i16 %A, i32 3
72 %tmp5 = insertelement <8 x i16> %tmp4, i16 %A, i32 4
73 %tmp6 = insertelement <8 x i16> %tmp5, i16 %A, i32 5
74 %tmp7 = insertelement <8 x i16> %tmp6, i16 %A, i32 6
75 %tmp8 = insertelement <8 x i16> %tmp7, i16 %A, i32 7
79 define <4 x i32> @v_dupQ32(i32 %A) nounwind {
82 %tmp1 = insertelement <4 x i32> zeroinitializer, i32 %A, i32 0
83 %tmp2 = insertelement <4 x i32> %tmp1, i32 %A, i32 1
84 %tmp3 = insertelement <4 x i32> %tmp2, i32 %A, i32 2
85 %tmp4 = insertelement <4 x i32> %tmp3, i32 %A, i32 3
89 define <4 x float> @v_dupQfloat(float %A) nounwind {
92 %tmp1 = insertelement <4 x float> zeroinitializer, float %A, i32 0
93 %tmp2 = insertelement <4 x float> %tmp1, float %A, i32 1
94 %tmp3 = insertelement <4 x float> %tmp2, float %A, i32 2
95 %tmp4 = insertelement <4 x float> %tmp3, float %A, i32 3
99 ; Check to make sure it works with shuffles, too.
101 define <8 x i8> @v_shuffledup8(i8 %A) nounwind {
102 ;CHECK: v_shuffledup8:
104 %tmp1 = insertelement <8 x i8> undef, i8 %A, i32 0
105 %tmp2 = shufflevector <8 x i8> %tmp1, <8 x i8> undef, <8 x i32> zeroinitializer
109 define <4 x i16> @v_shuffledup16(i16 %A) nounwind {
110 ;CHECK: v_shuffledup16:
112 %tmp1 = insertelement <4 x i16> undef, i16 %A, i32 0
113 %tmp2 = shufflevector <4 x i16> %tmp1, <4 x i16> undef, <4 x i32> zeroinitializer
117 define <2 x i32> @v_shuffledup32(i32 %A) nounwind {
118 ;CHECK: v_shuffledup32:
120 %tmp1 = insertelement <2 x i32> undef, i32 %A, i32 0
121 %tmp2 = shufflevector <2 x i32> %tmp1, <2 x i32> undef, <2 x i32> zeroinitializer
125 define <2 x float> @v_shuffledupfloat(float %A) nounwind {
126 ;CHECK: v_shuffledupfloat:
128 %tmp1 = insertelement <2 x float> undef, float %A, i32 0
129 %tmp2 = shufflevector <2 x float> %tmp1, <2 x float> undef, <2 x i32> zeroinitializer
130 ret <2 x float> %tmp2
133 define <16 x i8> @v_shuffledupQ8(i8 %A) nounwind {
134 ;CHECK: v_shuffledupQ8:
136 %tmp1 = insertelement <16 x i8> undef, i8 %A, i32 0
137 %tmp2 = shufflevector <16 x i8> %tmp1, <16 x i8> undef, <16 x i32> zeroinitializer
141 define <8 x i16> @v_shuffledupQ16(i16 %A) nounwind {
142 ;CHECK: v_shuffledupQ16:
144 %tmp1 = insertelement <8 x i16> undef, i16 %A, i32 0
145 %tmp2 = shufflevector <8 x i16> %tmp1, <8 x i16> undef, <8 x i32> zeroinitializer
149 define <4 x i32> @v_shuffledupQ32(i32 %A) nounwind {
150 ;CHECK: v_shuffledupQ32:
152 %tmp1 = insertelement <4 x i32> undef, i32 %A, i32 0
153 %tmp2 = shufflevector <4 x i32> %tmp1, <4 x i32> undef, <4 x i32> zeroinitializer
157 define <4 x float> @v_shuffledupQfloat(float %A) nounwind {
158 ;CHECK: v_shuffledupQfloat:
160 %tmp1 = insertelement <4 x float> undef, float %A, i32 0
161 %tmp2 = shufflevector <4 x float> %tmp1, <4 x float> undef, <4 x i32> zeroinitializer
162 ret <4 x float> %tmp2
165 define <8 x i8> @vduplane8(<8 x i8>* %A) nounwind {
168 %tmp1 = load <8 x i8>* %A
169 %tmp2 = shufflevector <8 x i8> %tmp1, <8 x i8> undef, <8 x i32> < i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1 >
173 define <4 x i16> @vduplane16(<4 x i16>* %A) nounwind {
176 %tmp1 = load <4 x i16>* %A
177 %tmp2 = shufflevector <4 x i16> %tmp1, <4 x i16> undef, <4 x i32> < i32 1, i32 1, i32 1, i32 1 >
181 define <2 x i32> @vduplane32(<2 x i32>* %A) nounwind {
184 %tmp1 = load <2 x i32>* %A
185 %tmp2 = shufflevector <2 x i32> %tmp1, <2 x i32> undef, <2 x i32> < i32 1, i32 1 >
189 define <2 x float> @vduplanefloat(<2 x float>* %A) nounwind {
190 ;CHECK: vduplanefloat:
192 %tmp1 = load <2 x float>* %A
193 %tmp2 = shufflevector <2 x float> %tmp1, <2 x float> undef, <2 x i32> < i32 1, i32 1 >
194 ret <2 x float> %tmp2
197 define <16 x i8> @vduplaneQ8(<8 x i8>* %A) nounwind {
200 %tmp1 = load <8 x i8>* %A
201 %tmp2 = shufflevector <8 x i8> %tmp1, <8 x i8> undef, <16 x i32> < i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1 >
205 define <8 x i16> @vduplaneQ16(<4 x i16>* %A) nounwind {
208 %tmp1 = load <4 x i16>* %A
209 %tmp2 = shufflevector <4 x i16> %tmp1, <4 x i16> undef, <8 x i32> < i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1 >
213 define <4 x i32> @vduplaneQ32(<2 x i32>* %A) nounwind {
216 %tmp1 = load <2 x i32>* %A
217 %tmp2 = shufflevector <2 x i32> %tmp1, <2 x i32> undef, <4 x i32> < i32 1, i32 1, i32 1, i32 1 >
221 define <4 x float> @vduplaneQfloat(<2 x float>* %A) nounwind {
222 ;CHECK: vduplaneQfloat:
224 %tmp1 = load <2 x float>* %A
225 %tmp2 = shufflevector <2 x float> %tmp1, <2 x float> undef, <4 x i32> < i32 1, i32 1, i32 1, i32 1 >
226 ret <4 x float> %tmp2
229 define <2 x i64> @foo(<2 x i64> %arg0_int64x1_t) nounwind readnone {
231 %0 = shufflevector <2 x i64> %arg0_int64x1_t, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
235 define <2 x i64> @bar(<2 x i64> %arg0_int64x1_t) nounwind readnone {
237 %0 = shufflevector <2 x i64> %arg0_int64x1_t, <2 x i64> undef, <2 x i32> <i32 0, i32 0>
241 define <2 x double> @baz(<2 x double> %arg0_int64x1_t) nounwind readnone {
243 %0 = shufflevector <2 x double> %arg0_int64x1_t, <2 x double> undef, <2 x i32> <i32 1, i32 1>
247 define <2 x double> @qux(<2 x double> %arg0_int64x1_t) nounwind readnone {
249 %0 = shufflevector <2 x double> %arg0_int64x1_t, <2 x double> undef, <2 x i32> <i32 0, i32 0>
254 ;CHECK: redundantVdup:
258 define void @redundantVdup(<8 x i8>* %ptr) nounwind {
259 %1 = insertelement <8 x i8> undef, i8 -128, i32 0
260 %2 = shufflevector <8 x i8> %1, <8 x i8> undef, <8 x i32> zeroinitializer
261 store <8 x i8> %2, <8 x i8>* %ptr, align 8