Run DCE after a LoopFlatten test to reduce spurious output [nfc]
[llvm-project.git] / clang / test / CodeGen / aarch64-sve-intrinsics / acle_sve_trn2.c
blob74a7e73eab77e784cd38f303eab977e0016b4a25
1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
2 // REQUIRES: aarch64-registered-target
3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
4 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
6 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
7 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -S -disable-O0-optnone -Werror -o /dev/null %s
8 #include <arm_sve.h>
10 #ifdef SVE_OVERLOADED_FORMS
11 // A simple used,unused... macro, long enough to represent any SVE builtin.
12 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
13 #else
14 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
15 #endif
17 // CHECK-LABEL: @test_svtrn2_s8(
18 // CHECK-NEXT: entry:
19 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
20 // CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
22 // CPP-CHECK-LABEL: @_Z14test_svtrn2_s8u10__SVInt8_tS_(
23 // CPP-CHECK-NEXT: entry:
24 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
25 // CPP-CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
27 svint8_t test_svtrn2_s8(svint8_t op1, svint8_t op2)
29 return SVE_ACLE_FUNC(svtrn2,_s8,,)(op1, op2);
32 // CHECK-LABEL: @test_svtrn2_s16(
33 // CHECK-NEXT: entry:
34 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
35 // CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
37 // CPP-CHECK-LABEL: @_Z15test_svtrn2_s16u11__SVInt16_tS_(
38 // CPP-CHECK-NEXT: entry:
39 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
40 // CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
42 svint16_t test_svtrn2_s16(svint16_t op1, svint16_t op2)
44 return SVE_ACLE_FUNC(svtrn2,_s16,,)(op1, op2);
47 // CHECK-LABEL: @test_svtrn2_s32(
48 // CHECK-NEXT: entry:
49 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
50 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
52 // CPP-CHECK-LABEL: @_Z15test_svtrn2_s32u11__SVInt32_tS_(
53 // CPP-CHECK-NEXT: entry:
54 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
55 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
57 svint32_t test_svtrn2_s32(svint32_t op1, svint32_t op2)
59 return SVE_ACLE_FUNC(svtrn2,_s32,,)(op1, op2);
62 // CHECK-LABEL: @test_svtrn2_s64(
63 // CHECK-NEXT: entry:
64 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
65 // CHECK-NEXT: ret <vscale x 2 x i64> [[TMP0]]
67 // CPP-CHECK-LABEL: @_Z15test_svtrn2_s64u11__SVInt64_tS_(
68 // CPP-CHECK-NEXT: entry:
69 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
70 // CPP-CHECK-NEXT: ret <vscale x 2 x i64> [[TMP0]]
72 svint64_t test_svtrn2_s64(svint64_t op1, svint64_t op2)
74 return SVE_ACLE_FUNC(svtrn2,_s64,,)(op1, op2);
77 // CHECK-LABEL: @test_svtrn2_u8(
78 // CHECK-NEXT: entry:
79 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
80 // CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
82 // CPP-CHECK-LABEL: @_Z14test_svtrn2_u8u11__SVUint8_tS_(
83 // CPP-CHECK-NEXT: entry:
84 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
85 // CPP-CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
87 svuint8_t test_svtrn2_u8(svuint8_t op1, svuint8_t op2)
89 return SVE_ACLE_FUNC(svtrn2,_u8,,)(op1, op2);
92 // CHECK-LABEL: @test_svtrn2_u16(
93 // CHECK-NEXT: entry:
94 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
95 // CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
97 // CPP-CHECK-LABEL: @_Z15test_svtrn2_u16u12__SVUint16_tS_(
98 // CPP-CHECK-NEXT: entry:
99 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
100 // CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
102 svuint16_t test_svtrn2_u16(svuint16_t op1, svuint16_t op2)
104 return SVE_ACLE_FUNC(svtrn2,_u16,,)(op1, op2);
107 // CHECK-LABEL: @test_svtrn2_u32(
108 // CHECK-NEXT: entry:
109 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
110 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
112 // CPP-CHECK-LABEL: @_Z15test_svtrn2_u32u12__SVUint32_tS_(
113 // CPP-CHECK-NEXT: entry:
114 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
115 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP0]]
117 svuint32_t test_svtrn2_u32(svuint32_t op1, svuint32_t op2)
119 return SVE_ACLE_FUNC(svtrn2,_u32,,)(op1, op2);
122 // CHECK-LABEL: @test_svtrn2_u64(
123 // CHECK-NEXT: entry:
124 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
125 // CHECK-NEXT: ret <vscale x 2 x i64> [[TMP0]]
127 // CPP-CHECK-LABEL: @_Z15test_svtrn2_u64u12__SVUint64_tS_(
128 // CPP-CHECK-NEXT: entry:
129 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
130 // CPP-CHECK-NEXT: ret <vscale x 2 x i64> [[TMP0]]
132 svuint64_t test_svtrn2_u64(svuint64_t op1, svuint64_t op2)
134 return SVE_ACLE_FUNC(svtrn2,_u64,,)(op1, op2);
137 // CHECK-LABEL: @test_svtrn2_f16(
138 // CHECK-NEXT: entry:
139 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.trn2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
140 // CHECK-NEXT: ret <vscale x 8 x half> [[TMP0]]
142 // CPP-CHECK-LABEL: @_Z15test_svtrn2_f16u13__SVFloat16_tS_(
143 // CPP-CHECK-NEXT: entry:
144 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.trn2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
145 // CPP-CHECK-NEXT: ret <vscale x 8 x half> [[TMP0]]
147 svfloat16_t test_svtrn2_f16(svfloat16_t op1, svfloat16_t op2)
149 return SVE_ACLE_FUNC(svtrn2,_f16,,)(op1, op2);
152 // CHECK-LABEL: @test_svtrn2_f32(
153 // CHECK-NEXT: entry:
154 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.trn2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
155 // CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
157 // CPP-CHECK-LABEL: @_Z15test_svtrn2_f32u13__SVFloat32_tS_(
158 // CPP-CHECK-NEXT: entry:
159 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x float> @llvm.aarch64.sve.trn2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
160 // CPP-CHECK-NEXT: ret <vscale x 4 x float> [[TMP0]]
162 svfloat32_t test_svtrn2_f32(svfloat32_t op1, svfloat32_t op2)
164 return SVE_ACLE_FUNC(svtrn2,_f32,,)(op1, op2);
167 // CHECK-LABEL: @test_svtrn2_f64(
168 // CHECK-NEXT: entry:
169 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.trn2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
170 // CHECK-NEXT: ret <vscale x 2 x double> [[TMP0]]
172 // CPP-CHECK-LABEL: @_Z15test_svtrn2_f64u13__SVFloat64_tS_(
173 // CPP-CHECK-NEXT: entry:
174 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x double> @llvm.aarch64.sve.trn2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
175 // CPP-CHECK-NEXT: ret <vscale x 2 x double> [[TMP0]]
177 svfloat64_t test_svtrn2_f64(svfloat64_t op1, svfloat64_t op2)
179 return SVE_ACLE_FUNC(svtrn2,_f64,,)(op1, op2);
182 // CHECK-LABEL: @test_svtrn2_b8(
183 // CHECK-NEXT: entry:
184 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
185 // CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
187 // CPP-CHECK-LABEL: @_Z14test_svtrn2_b8u10__SVBool_tS_(
188 // CPP-CHECK-NEXT: entry:
189 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
190 // CPP-CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
192 svbool_t test_svtrn2_b8(svbool_t op1, svbool_t op2)
194 return svtrn2_b8(op1, op2);
197 // CHECK-LABEL: @test_svtrn2_b16(
198 // CHECK-NEXT: entry:
199 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b16(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
200 // CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
202 // CPP-CHECK-LABEL: @_Z15test_svtrn2_b16u10__SVBool_tS_(
203 // CPP-CHECK-NEXT: entry:
204 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b16(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
205 // CPP-CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
207 svbool_t test_svtrn2_b16(svbool_t op1, svbool_t op2)
209 return svtrn2_b16(op1, op2);
212 // CHECK-LABEL: @test_svtrn2_b32(
213 // CHECK-NEXT: entry:
214 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b32(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
215 // CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
217 // CPP-CHECK-LABEL: @_Z15test_svtrn2_b32u10__SVBool_tS_(
218 // CPP-CHECK-NEXT: entry:
219 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b32(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
220 // CPP-CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
222 svbool_t test_svtrn2_b32(svbool_t op1, svbool_t op2)
224 return svtrn2_b32(op1, op2);
227 // CHECK-LABEL: @test_svtrn2_b64(
228 // CHECK-NEXT: entry:
229 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b64(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
230 // CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
232 // CPP-CHECK-LABEL: @_Z15test_svtrn2_b64u10__SVBool_tS_(
233 // CPP-CHECK-NEXT: entry:
234 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.b64(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
235 // CPP-CHECK-NEXT: ret <vscale x 16 x i1> [[TMP0]]
237 svbool_t test_svtrn2_b64(svbool_t op1, svbool_t op2)
239 return svtrn2_b64(op1, op2);