xref: /llvm-project/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-deinterleave-load.ll (revision bc74a1edbe5e6a3603e65efe06116fa72747acab)
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -mtriple=riscv32 -mattr=+v,+zvfh,+zvfbfmin | FileCheck %s
3; RUN: llc < %s -mtriple=riscv64 -mattr=+v,+zvfh,+zvfbfmin | FileCheck %s
4; RUN: llc < %s -mtriple=riscv32 -mattr=+v,+zvfhmin,+zvfbfmin | FileCheck %s
5; RUN: llc < %s -mtriple=riscv64 -mattr=+v,+zvfhmin,+zvfbfmin | FileCheck %s
6
7; Integers
8
9define {<16 x i1>, <16 x i1>} @vector_deinterleave_load_v16i1_v32i1(ptr %p) {
10; CHECK-LABEL: vector_deinterleave_load_v16i1_v32i1:
11; CHECK:       # %bb.0:
12; CHECK-NEXT:    li a1, 32
13; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
14; CHECK-NEXT:    vmv.v.i v10, 0
15; CHECK-NEXT:    vid.v v9
16; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
17; CHECK-NEXT:    vlm.v v8, (a0)
18; CHECK-NEXT:    li a0, -256
19; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
20; CHECK-NEXT:    vadd.vv v11, v9, v9
21; CHECK-NEXT:    vsetvli zero, zero, e16, m2, ta, ma
22; CHECK-NEXT:    vmv.s.x v9, a0
23; CHECK-NEXT:    vsetvli zero, zero, e8, m1, ta, ma
24; CHECK-NEXT:    vadd.vi v12, v11, -16
25; CHECK-NEXT:    vsetivli zero, 2, e8, mf4, ta, ma
26; CHECK-NEXT:    vslidedown.vi v0, v8, 2
27; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
28; CHECK-NEXT:    vadd.vi v11, v11, -15
29; CHECK-NEXT:    vmerge.vim v13, v10, 1, v0
30; CHECK-NEXT:    vmv1r.v v0, v8
31; CHECK-NEXT:    vmerge.vim v8, v10, 1, v0
32; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
33; CHECK-NEXT:    vnsrl.wi v10, v8, 0
34; CHECK-NEXT:    vnsrl.wi v8, v8, 8
35; CHECK-NEXT:    vmv1r.v v0, v9
36; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, mu
37; CHECK-NEXT:    vrgather.vv v10, v13, v12, v0.t
38; CHECK-NEXT:    vrgather.vv v8, v13, v11, v0.t
39; CHECK-NEXT:    vmsne.vi v0, v10, 0
40; CHECK-NEXT:    vmsne.vi v8, v8, 0
41; CHECK-NEXT:    ret
42  %vec = load <32 x i1>, ptr %p
43  %deinterleaved.results = call {<16 x i1>, <16 x i1>} @llvm.vector.deinterleave2.v32i1(<32 x i1> %vec)
44  %t0 = extractvalue { <16 x i1>, <16 x i1> } %deinterleaved.results, 0
45  %t1 = extractvalue { <16 x i1>, <16 x i1> } %deinterleaved.results, 1
46  %res0 = insertvalue { <16 x i1>, <16 x i1> } undef, <16 x i1> %t0, 0
47  %res1 = insertvalue { <16 x i1>, <16 x i1> } %res0, <16 x i1> %t1, 1
48  ret {<16 x i1>, <16 x i1>} %res1
49}
50
51define {<16 x i8>, <16 x i8>} @vector_deinterleave_load_v16i8_v32i8(ptr %p) {
52; CHECK-LABEL: vector_deinterleave_load_v16i8_v32i8:
53; CHECK:       # %bb.0:
54; CHECK-NEXT:    vsetivli zero, 16, e8, m1, ta, ma
55; CHECK-NEXT:    vlseg2e8.v v8, (a0)
56; CHECK-NEXT:    ret
57  %vec = load <32 x i8>, ptr %p
58  %deinterleaved.results = call {<16 x i8>, <16 x i8>} @llvm.vector.deinterleave2.v32i8(<32 x i8> %vec)
59  %t0 = extractvalue { <16 x i8>, <16 x i8> } %deinterleaved.results, 0
60  %t1 = extractvalue { <16 x i8>, <16 x i8> } %deinterleaved.results, 1
61  %res0 = insertvalue { <16 x i8>, <16 x i8> } undef, <16 x i8> %t0, 0
62  %res1 = insertvalue { <16 x i8>, <16 x i8> } %res0, <16 x i8> %t1, 1
63  ret {<16 x i8>, <16 x i8>} %res1
64}
65
66; Shouldn't be lowered to vlseg because it's unaligned
67define {<8 x i16>, <8 x i16>} @vector_deinterleave_load_v8i16_v16i16_align1(ptr %p) {
68; CHECK-LABEL: vector_deinterleave_load_v8i16_v16i16_align1:
69; CHECK:       # %bb.0:
70; CHECK-NEXT:    li a1, 32
71; CHECK-NEXT:    vsetvli zero, a1, e8, m2, ta, ma
72; CHECK-NEXT:    vle8.v v10, (a0)
73; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
74; CHECK-NEXT:    vnsrl.wi v8, v10, 0
75; CHECK-NEXT:    vnsrl.wi v9, v10, 16
76; CHECK-NEXT:    ret
77  %vec = load <16 x i16>, ptr %p, align 1
78  %deinterleaved.results = call {<8 x i16>, <8 x i16>} @llvm.vector.deinterleave2.v16i16(<16 x i16> %vec)
79  %t0 = extractvalue { <8 x i16>, <8 x i16> } %deinterleaved.results, 0
80  %t1 = extractvalue { <8 x i16>, <8 x i16> } %deinterleaved.results, 1
81  %res0 = insertvalue { <8 x i16>, <8 x i16> } undef, <8 x i16> %t0, 0
82  %res1 = insertvalue { <8 x i16>, <8 x i16> } %res0, <8 x i16> %t1, 1
83  ret {<8 x i16>, <8 x i16>} %res1
84}
85
86define {<8 x i16>, <8 x i16>} @vector_deinterleave_load_v8i16_v16i16(ptr %p) {
87; CHECK-LABEL: vector_deinterleave_load_v8i16_v16i16:
88; CHECK:       # %bb.0:
89; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
90; CHECK-NEXT:    vlseg2e16.v v8, (a0)
91; CHECK-NEXT:    ret
92  %vec = load <16 x i16>, ptr %p
93  %deinterleaved.results = call {<8 x i16>, <8 x i16>} @llvm.vector.deinterleave2.v16i16(<16 x i16> %vec)
94  %t0 = extractvalue { <8 x i16>, <8 x i16> } %deinterleaved.results, 0
95  %t1 = extractvalue { <8 x i16>, <8 x i16> } %deinterleaved.results, 1
96  %res0 = insertvalue { <8 x i16>, <8 x i16> } undef, <8 x i16> %t0, 0
97  %res1 = insertvalue { <8 x i16>, <8 x i16> } %res0, <8 x i16> %t1, 1
98  ret {<8 x i16>, <8 x i16>} %res1
99}
100
101define {<4 x i32>, <4 x i32>} @vector_deinterleave_load_v4i32_vv8i32(ptr %p) {
102; CHECK-LABEL: vector_deinterleave_load_v4i32_vv8i32:
103; CHECK:       # %bb.0:
104; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
105; CHECK-NEXT:    vlseg2e32.v v8, (a0)
106; CHECK-NEXT:    ret
107  %vec = load <8 x i32>, ptr %p
108  %deinterleaved.results = call {<4 x i32>, <4 x i32>} @llvm.vector.deinterleave2.v8i32(<8 x i32> %vec)
109  %t0 = extractvalue { <4 x i32>, <4 x i32> } %deinterleaved.results, 0
110  %t1 = extractvalue { <4 x i32>, <4 x i32> } %deinterleaved.results, 1
111  %res0 = insertvalue { <4 x i32>, <4 x i32> } undef, <4 x i32> %t0, 0
112  %res1 = insertvalue { <4 x i32>, <4 x i32> } %res0, <4 x i32> %t1, 1
113  ret {<4 x i32>, <4 x i32>} %res1
114}
115
116define {<2 x i64>, <2 x i64>} @vector_deinterleave_load_v2i64_v4i64(ptr %p) {
117; CHECK-LABEL: vector_deinterleave_load_v2i64_v4i64:
118; CHECK:       # %bb.0:
119; CHECK-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
120; CHECK-NEXT:    vlseg2e64.v v8, (a0)
121; CHECK-NEXT:    ret
122  %vec = load <4 x i64>, ptr %p
123  %deinterleaved.results = call {<2 x i64>, <2 x i64>} @llvm.vector.deinterleave2.v4i64(<4 x i64> %vec)
124  %t0 = extractvalue { <2 x i64>, <2 x i64> } %deinterleaved.results, 0
125  %t1 = extractvalue { <2 x i64>, <2 x i64> } %deinterleaved.results, 1
126  %res0 = insertvalue { <2 x i64>, <2 x i64> } undef, <2 x i64> %t0, 0
127  %res1 = insertvalue { <2 x i64>, <2 x i64> } %res0, <2 x i64> %t1, 1
128  ret {<2 x i64>, <2 x i64>} %res1
129}
130
131; Floats
132
133define {<2 x bfloat>, <2 x bfloat>} @vector_deinterleave_load_v2bf16_v4bf16(ptr %p) {
134; CHECK-LABEL: vector_deinterleave_load_v2bf16_v4bf16:
135; CHECK:       # %bb.0:
136; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
137; CHECK-NEXT:    vlseg2e16.v v8, (a0)
138; CHECK-NEXT:    ret
139  %vec = load <4 x bfloat>, ptr %p
140  %deinterleaved.results = call {<2 x bfloat>, <2 x bfloat>} @llvm.vector.deinterleave2.v4bf16(<4 x bfloat> %vec)
141  %t0 = extractvalue { <2 x bfloat>, <2 x bfloat> } %deinterleaved.results, 0
142  %t1 = extractvalue { <2 x bfloat>, <2 x bfloat> } %deinterleaved.results, 1
143  %res0 = insertvalue { <2 x bfloat>, <2 x bfloat> } undef, <2 x bfloat> %t0, 0
144  %res1 = insertvalue { <2 x bfloat>, <2 x bfloat> } %res0, <2 x bfloat> %t1, 1
145  ret {<2 x bfloat>, <2 x bfloat>} %res1
146}
147
148define {<4 x bfloat>, <4 x bfloat>} @vector_deinterleave_load_v4bf16_v8bf16(ptr %p) {
149; CHECK-LABEL: vector_deinterleave_load_v4bf16_v8bf16:
150; CHECK:       # %bb.0:
151; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
152; CHECK-NEXT:    vlseg2e16.v v8, (a0)
153; CHECK-NEXT:    ret
154  %vec = load <8 x bfloat>, ptr %p
155  %deinterleaved.results = call {<4 x bfloat>, <4 x bfloat>} @llvm.vector.deinterleave2.v8bf16(<8 x bfloat> %vec)
156  %t0 = extractvalue { <4 x bfloat>, <4 x bfloat> } %deinterleaved.results, 0
157  %t1 = extractvalue { <4 x bfloat>, <4 x bfloat> } %deinterleaved.results, 1
158  %res0 = insertvalue { <4 x bfloat>, <4 x bfloat> } undef, <4 x bfloat> %t0, 0
159  %res1 = insertvalue { <4 x bfloat>, <4 x bfloat> } %res0, <4 x bfloat> %t1, 1
160  ret {<4 x bfloat>, <4 x bfloat>} %res1
161}
162
163define {<2 x half>, <2 x half>} @vector_deinterleave_load_v2f16_v4f16(ptr %p) {
164; CHECK-LABEL: vector_deinterleave_load_v2f16_v4f16:
165; CHECK:       # %bb.0:
166; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
167; CHECK-NEXT:    vlseg2e16.v v8, (a0)
168; CHECK-NEXT:    ret
169  %vec = load <4 x half>, ptr %p
170  %deinterleaved.results = call {<2 x half>, <2 x half>} @llvm.vector.deinterleave2.v4f16(<4 x half> %vec)
171  %t0 = extractvalue { <2 x half>, <2 x half> } %deinterleaved.results, 0
172  %t1 = extractvalue { <2 x half>, <2 x half> } %deinterleaved.results, 1
173  %res0 = insertvalue { <2 x half>, <2 x half> } undef, <2 x half> %t0, 0
174  %res1 = insertvalue { <2 x half>, <2 x half> } %res0, <2 x half> %t1, 1
175  ret {<2 x half>, <2 x half>} %res1
176}
177
178define {<4 x half>, <4 x half>} @vector_deinterleave_load_v4f16_v8f16(ptr %p) {
179; CHECK-LABEL: vector_deinterleave_load_v4f16_v8f16:
180; CHECK:       # %bb.0:
181; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
182; CHECK-NEXT:    vlseg2e16.v v8, (a0)
183; CHECK-NEXT:    ret
184  %vec = load <8 x half>, ptr %p
185  %deinterleaved.results = call {<4 x half>, <4 x half>} @llvm.vector.deinterleave2.v8f16(<8 x half> %vec)
186  %t0 = extractvalue { <4 x half>, <4 x half> } %deinterleaved.results, 0
187  %t1 = extractvalue { <4 x half>, <4 x half> } %deinterleaved.results, 1
188  %res0 = insertvalue { <4 x half>, <4 x half> } undef, <4 x half> %t0, 0
189  %res1 = insertvalue { <4 x half>, <4 x half> } %res0, <4 x half> %t1, 1
190  ret {<4 x half>, <4 x half>} %res1
191}
192
193define {<2 x float>, <2 x float>} @vector_deinterleave_load_v2f32_v4f32(ptr %p) {
194; CHECK-LABEL: vector_deinterleave_load_v2f32_v4f32:
195; CHECK:       # %bb.0:
196; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
197; CHECK-NEXT:    vlseg2e32.v v8, (a0)
198; CHECK-NEXT:    ret
199  %vec = load <4 x float>, ptr %p
200  %deinterleaved.results = call {<2 x float>, <2 x float>} @llvm.vector.deinterleave2.v4f32(<4 x float> %vec)
201  %t0 = extractvalue { <2 x float>, <2 x float> } %deinterleaved.results, 0
202  %t1 = extractvalue { <2 x float>, <2 x float> } %deinterleaved.results, 1
203  %res0 = insertvalue { <2 x float>, <2 x float> } undef, <2 x float> %t0, 0
204  %res1 = insertvalue { <2 x float>, <2 x float> } %res0, <2 x float> %t1, 1
205  ret {<2 x float>, <2 x float>} %res1
206}
207
208define {<8 x bfloat>, <8 x bfloat>} @vector_deinterleave_load_v8bf16_v16bf16(ptr %p) {
209; CHECK-LABEL: vector_deinterleave_load_v8bf16_v16bf16:
210; CHECK:       # %bb.0:
211; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
212; CHECK-NEXT:    vlseg2e16.v v8, (a0)
213; CHECK-NEXT:    ret
214  %vec = load <16 x bfloat>, ptr %p
215  %deinterleaved.results = call {<8 x bfloat>, <8 x bfloat>} @llvm.vector.deinterleave2.v16bf16(<16 x bfloat> %vec)
216  %t0 = extractvalue { <8 x bfloat>, <8 x bfloat> } %deinterleaved.results, 0
217  %t1 = extractvalue { <8 x bfloat>, <8 x bfloat> } %deinterleaved.results, 1
218  %res0 = insertvalue { <8 x bfloat>, <8 x bfloat> } undef, <8 x bfloat> %t0, 0
219  %res1 = insertvalue { <8 x bfloat>, <8 x bfloat> } %res0, <8 x bfloat> %t1, 1
220  ret {<8 x bfloat>, <8 x bfloat>} %res1
221}
222
223define {<8 x half>, <8 x half>} @vector_deinterleave_load_v8f16_v16f16(ptr %p) {
224; CHECK-LABEL: vector_deinterleave_load_v8f16_v16f16:
225; CHECK:       # %bb.0:
226; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
227; CHECK-NEXT:    vlseg2e16.v v8, (a0)
228; CHECK-NEXT:    ret
229  %vec = load <16 x half>, ptr %p
230  %deinterleaved.results = call {<8 x half>, <8 x half>} @llvm.vector.deinterleave2.v16f16(<16 x half> %vec)
231  %t0 = extractvalue { <8 x half>, <8 x half> } %deinterleaved.results, 0
232  %t1 = extractvalue { <8 x half>, <8 x half> } %deinterleaved.results, 1
233  %res0 = insertvalue { <8 x half>, <8 x half> } undef, <8 x half> %t0, 0
234  %res1 = insertvalue { <8 x half>, <8 x half> } %res0, <8 x half> %t1, 1
235  ret {<8 x half>, <8 x half>} %res1
236}
237
238define {<4 x float>, <4 x float>} @vector_deinterleave_load_v4f32_v8f32(ptr %p) {
239; CHECK-LABEL: vector_deinterleave_load_v4f32_v8f32:
240; CHECK:       # %bb.0:
241; CHECK-NEXT:    vsetivli zero, 4, e32, m1, ta, ma
242; CHECK-NEXT:    vlseg2e32.v v8, (a0)
243; CHECK-NEXT:    ret
244  %vec = load <8 x float>, ptr %p
245  %deinterleaved.results = call {<4 x float>, <4 x float>} @llvm.vector.deinterleave2.v8f32(<8 x float> %vec)
246  %t0 = extractvalue { <4 x float>, <4 x float> } %deinterleaved.results, 0
247  %t1 = extractvalue { <4 x float>, <4 x float> } %deinterleaved.results, 1
248  %res0 = insertvalue { <4 x float>, <4 x float> } undef, <4 x float> %t0, 0
249  %res1 = insertvalue { <4 x float>, <4 x float> } %res0, <4 x float> %t1, 1
250  ret {<4 x float>, <4 x float>} %res1
251}
252
253define {<2 x double>, <2 x double>} @vector_deinterleave_load_v2f64_v4f64(ptr %p) {
254; CHECK-LABEL: vector_deinterleave_load_v2f64_v4f64:
255; CHECK:       # %bb.0:
256; CHECK-NEXT:    vsetivli zero, 2, e64, m1, ta, ma
257; CHECK-NEXT:    vlseg2e64.v v8, (a0)
258; CHECK-NEXT:    ret
259  %vec = load <4 x double>, ptr %p
260  %deinterleaved.results = call {<2 x double>, <2 x double>} @llvm.vector.deinterleave2.v4f64(<4 x double> %vec)
261  %t0 = extractvalue { <2 x double>, <2 x double> } %deinterleaved.results, 0
262  %t1 = extractvalue { <2 x double>, <2 x double> } %deinterleaved.results, 1
263  %res0 = insertvalue { <2 x double>, <2 x double> } undef, <2 x double> %t0, 0
264  %res1 = insertvalue { <2 x double>, <2 x double> } %res0, <2 x double> %t1, 1
265  ret {<2 x double>, <2 x double>} %res1
266}
267
268define { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @vector_deinterleave_load_factor4(ptr %p) {
269; CHECK-LABEL: vector_deinterleave_load_factor4:
270; CHECK:       # %bb.0:
271; CHECK-NEXT:    vsetivli zero, 8, e8, mf2, ta, ma
272; CHECK-NEXT:    vlseg4e8.v v8, (a0)
273; CHECK-NEXT:    ret
274  %vec = load <32 x i8>, ptr %p
275  %d0 = call {<16 x i8>, <16 x i8>} @llvm.vector.deinterleave2.v32i8(<32 x i8> %vec)
276  %d0.0 = extractvalue { <16 x i8>, <16 x i8> } %d0, 0
277  %d0.1 = extractvalue { <16 x i8>, <16 x i8> } %d0, 1
278  %d1 = call {<8 x i8>, <8 x i8>} @llvm.vector.deinterleave2.v16i8(<16 x i8> %d0.0)
279  %t0 = extractvalue { <8 x i8>, <8 x i8> } %d1, 0
280  %t2 = extractvalue { <8 x i8>, <8 x i8> } %d1, 1
281  %d2 = call {<8 x i8>, <8 x i8>} @llvm.vector.deinterleave2.v16i8(<16 x i8> %d0.1)
282  %t1 = extractvalue { <8 x i8>, <8 x i8> } %d2, 0
283  %t3 = extractvalue { <8 x i8>, <8 x i8> } %d2, 1
284
285  %res0 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } undef, <8 x i8> %t0, 0
286  %res1 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res0, <8 x i8> %t1, 1
287  %res2 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res1, <8 x i8> %t2, 2
288  %res3 = insertvalue { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res2, <8 x i8> %t3, 3
289  ret { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } %res3
290}
291
292define {<2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>} @vector_deinterleave_load_factor8(ptr %ptr) {
293; CHECK-LABEL: vector_deinterleave_load_factor8:
294; CHECK:       # %bb.0:
295; CHECK-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
296; CHECK-NEXT:    vlseg8e32.v v8, (a0)
297; CHECK-NEXT:    ret
298  %vec = load <16 x i32>, ptr %ptr
299  %d0 = call { <8 x i32>, <8 x i32> } @llvm.vector.deinterleave2.v16i32(<16 x i32> %vec)
300  %d0.0 = extractvalue { <8 x i32>, <8 x i32> } %d0, 0
301  %d0.1 = extractvalue { <8 x i32>, <8 x i32> } %d0, 1
302  %d1 = call { <4 x i32>, <4 x i32> } @llvm.vector.deinterleave2.v8i32(<8 x i32> %d0.0)
303  %d1.0 = extractvalue { <4 x i32>, <4 x i32> } %d1, 0
304  %d1.1 = extractvalue { <4 x i32>, <4 x i32> } %d1, 1
305  %d2 = call { <4 x i32>, <4 x i32> } @llvm.vector.deinterleave2.v8i32(<8 x i32> %d0.1)
306  %d2.0 = extractvalue { <4 x i32>, <4 x i32> } %d2, 0
307  %d2.1 = extractvalue { <4 x i32>, <4 x i32> } %d2, 1
308
309  %d3 = call { <2 x i32>, <2 x i32> } @llvm.vector.deinterleave2.v4i32(<4 x i32> %d1.0)
310  %t0 = extractvalue { <2 x i32>, <2 x i32> } %d3, 0
311  %t4 = extractvalue { <2 x i32>, <2 x i32> } %d3, 1
312  %d4 = call { <2 x i32>, <2 x i32> } @llvm.vector.deinterleave2.v4i32(<4 x i32> %d1.1)
313  %t2 = extractvalue { <2 x i32>, <2 x i32> } %d4, 0
314  %t6 = extractvalue { <2 x i32>, <2 x i32> } %d4, 1
315  %d5 = call { <2 x i32>, <2 x i32> } @llvm.vector.deinterleave2.v4i32(<4 x i32> %d2.0)
316  %t1 = extractvalue { <2 x i32>, <2 x i32> } %d5, 0
317  %t5 = extractvalue { <2 x i32>, <2 x i32> } %d5, 1
318  %d6 = call { <2 x i32>, <2 x i32> } @llvm.vector.deinterleave2.v4i32(<4 x i32> %d2.1)
319  %t3 = extractvalue { <2 x i32>, <2 x i32> } %d6, 0
320  %t7 = extractvalue { <2 x i32>, <2 x i32> } %d6, 1
321
322  %res0 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } undef, <2 x i32> %t0, 0
323  %res1 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res0, <2 x i32> %t1, 1
324  %res2 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res1, <2 x i32> %t2, 2
325  %res3 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res2, <2 x i32> %t3, 3
326  %res4 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res3, <2 x i32> %t4, 4
327  %res5 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res4, <2 x i32> %t5, 5
328  %res6 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res5, <2 x i32> %t6, 6
329  %res7 = insertvalue { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res6, <2 x i32> %t7, 7
330  ret { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> } %res7
331}
332