• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X32
3; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X64
4
5define <16 x i16> @phaddw1(<16 x i16> %x, <16 x i16> %y) {
6; X32-LABEL: phaddw1:
7; X32:       # %bb.0:
8; X32-NEXT:    vphaddw %ymm1, %ymm0, %ymm0
9; X32-NEXT:    retl
10;
11; X64-LABEL: phaddw1:
12; X64:       # %bb.0:
13; X64-NEXT:    vphaddw %ymm1, %ymm0, %ymm0
14; X64-NEXT:    retq
15  %a = shufflevector <16 x i16> %x, <16 x i16> %y, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 16, i32 18, i32 20, i32 22, i32 8, i32 10, i32 12, i32 14, i32 24, i32 26, i32 28, i32 30>
16  %b = shufflevector <16 x i16> %x, <16 x i16> %y, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 17, i32 19, i32 21, i32 23, i32 9, i32 11, i32 13, i32 15, i32 25, i32 27, i32 29, i32 31>
17  %r = add <16 x i16> %a, %b
18  ret <16 x i16> %r
19}
20
21define <16 x i16> @phaddw2(<16 x i16> %x, <16 x i16> %y) {
22; X32-LABEL: phaddw2:
23; X32:       # %bb.0:
24; X32-NEXT:    vphaddw %ymm1, %ymm0, %ymm0
25; X32-NEXT:    retl
26;
27; X64-LABEL: phaddw2:
28; X64:       # %bb.0:
29; X64-NEXT:    vphaddw %ymm1, %ymm0, %ymm0
30; X64-NEXT:    retq
31  %a = shufflevector <16 x i16> %x, <16 x i16> %y, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 17, i32 19, i32 21, i32 23, i32 9, i32 11, i32 13, i32 15, i32 25, i32 27, i32 29, i32 31>
32  %b = shufflevector <16 x i16> %y, <16 x i16> %x, <16 x i32> <i32 16, i32 18, i32 20, i32 22, i32 0, i32 2, i32 4, i32 6, i32 24, i32 26, i32 28, i32 30, i32 8, i32 10, i32 12, i32 14>
33  %r = add <16 x i16> %a, %b
34  ret <16 x i16> %r
35}
36
37define <8 x i32> @phaddd1(<8 x i32> %x, <8 x i32> %y) {
38; X32-LABEL: phaddd1:
39; X32:       # %bb.0:
40; X32-NEXT:    vphaddd %ymm1, %ymm0, %ymm0
41; X32-NEXT:    retl
42;
43; X64-LABEL: phaddd1:
44; X64:       # %bb.0:
45; X64-NEXT:    vphaddd %ymm1, %ymm0, %ymm0
46; X64-NEXT:    retq
47  %a = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 0, i32 2, i32 8, i32 10, i32 4, i32 6, i32 12, i32 14>
48  %b = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 1, i32 3, i32 9, i32 11, i32 5, i32 7, i32 13, i32 15>
49  %r = add <8 x i32> %a, %b
50  ret <8 x i32> %r
51}
52
53define <8 x i32> @phaddd2(<8 x i32> %x, <8 x i32> %y) {
54; X32-LABEL: phaddd2:
55; X32:       # %bb.0:
56; X32-NEXT:    vphaddd %ymm1, %ymm0, %ymm0
57; X32-NEXT:    retl
58;
59; X64-LABEL: phaddd2:
60; X64:       # %bb.0:
61; X64-NEXT:    vphaddd %ymm1, %ymm0, %ymm0
62; X64-NEXT:    retq
63  %a = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 1, i32 2, i32 9, i32 10, i32 5, i32 6, i32 13, i32 14>
64  %b = shufflevector <8 x i32> %y, <8 x i32> %x, <8 x i32> <i32 8, i32 11, i32 0, i32 3, i32 12, i32 15, i32 4, i32 7>
65  %r = add <8 x i32> %a, %b
66  ret <8 x i32> %r
67}
68
69define <8 x i32> @phaddd3(<8 x i32> %x) {
70; X32-LABEL: phaddd3:
71; X32:       # %bb.0:
72; X32-NEXT:    vphaddd %ymm0, %ymm0, %ymm0
73; X32-NEXT:    retl
74;
75; X64-LABEL: phaddd3:
76; X64:       # %bb.0:
77; X64-NEXT:    vphaddd %ymm0, %ymm0, %ymm0
78; X64-NEXT:    retq
79  %a = shufflevector <8 x i32> %x, <8 x i32> undef, <8 x i32> <i32 undef, i32 2, i32 8, i32 10, i32 4, i32 6, i32 undef, i32 14>
80  %b = shufflevector <8 x i32> %x, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 9, i32 undef, i32 5, i32 7, i32 13, i32 15>
81  %r = add <8 x i32> %a, %b
82  ret <8 x i32> %r
83}
84
85define <16 x i16> @phsubw1(<16 x i16> %x, <16 x i16> %y) {
86; X32-LABEL: phsubw1:
87; X32:       # %bb.0:
88; X32-NEXT:    vphsubw %ymm1, %ymm0, %ymm0
89; X32-NEXT:    retl
90;
91; X64-LABEL: phsubw1:
92; X64:       # %bb.0:
93; X64-NEXT:    vphsubw %ymm1, %ymm0, %ymm0
94; X64-NEXT:    retq
95  %a = shufflevector <16 x i16> %x, <16 x i16> %y, <16 x i32> <i32 0, i32 2, i32 4, i32 6, i32 16, i32 18, i32 20, i32 22, i32 8, i32 10, i32 12, i32 14, i32 24, i32 26, i32 28, i32 30>
96  %b = shufflevector <16 x i16> %x, <16 x i16> %y, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 17, i32 19, i32 21, i32 23, i32 9, i32 11, i32 13, i32 15, i32 25, i32 27, i32 29, i32 31>
97  %r = sub <16 x i16> %a, %b
98  ret <16 x i16> %r
99}
100
101define <8 x i32> @phsubd1(<8 x i32> %x, <8 x i32> %y) {
102; X32-LABEL: phsubd1:
103; X32:       # %bb.0:
104; X32-NEXT:    vphsubd %ymm1, %ymm0, %ymm0
105; X32-NEXT:    retl
106;
107; X64-LABEL: phsubd1:
108; X64:       # %bb.0:
109; X64-NEXT:    vphsubd %ymm1, %ymm0, %ymm0
110; X64-NEXT:    retq
111  %a = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 0, i32 2, i32 8, i32 10, i32 4, i32 6, i32 12, i32 14>
112  %b = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 1, i32 3, i32 9, i32 11, i32 5, i32 7, i32 13, i32 15>
113  %r = sub <8 x i32> %a, %b
114  ret <8 x i32> %r
115}
116
117define <8 x i32> @phsubd2(<8 x i32> %x, <8 x i32> %y) {
118; X32-LABEL: phsubd2:
119; X32:       # %bb.0:
120; X32-NEXT:    vphsubd %ymm1, %ymm0, %ymm0
121; X32-NEXT:    retl
122;
123; X64-LABEL: phsubd2:
124; X64:       # %bb.0:
125; X64-NEXT:    vphsubd %ymm1, %ymm0, %ymm0
126; X64-NEXT:    retq
127  %a = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 0, i32 undef, i32 8, i32 undef, i32 4, i32 6, i32 12, i32 14>
128  %b = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 1, i32 undef, i32 9, i32 11, i32 5, i32 7, i32 undef, i32 15>
129  %r = sub <8 x i32> %a, %b
130  ret <8 x i32> %r
131}
132