Run DCE after a LoopFlatten test to reduce spurious output [nfc]
[llvm-project.git] / clang / test / CodeGen / aarch64-sve2-intrinsics / acle_sve2_histcnt.c
blob15c6f0f1aa7c1ffdf2ceabec34cc0d20f848e048
1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
2 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
4 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
7 // REQUIRES: aarch64-registered-target
9 #include <arm_sve.h>
11 #ifdef SVE_OVERLOADED_FORMS
12 // A simple used,unused... macro, long enough to represent any SVE builtin.
13 #define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
14 #else
15 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
16 #endif
18 // CHECK-LABEL: @test_svhistcnt_s32_z(
19 // CHECK-NEXT: entry:
20 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
21 // CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
22 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP1]]
24 // CPP-CHECK-LABEL: @_Z20test_svhistcnt_s32_zu10__SVBool_tu11__SVInt32_tS0_(
25 // CPP-CHECK-NEXT: entry:
26 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
27 // CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
28 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP1]]
30 svuint32_t test_svhistcnt_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
32 return SVE_ACLE_FUNC(svhistcnt,_s32,_z,)(pg, op1, op2);
35 // CHECK-LABEL: @test_svhistcnt_s64_z(
36 // CHECK-NEXT: entry:
37 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
38 // CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
39 // CHECK-NEXT: ret <vscale x 2 x i64> [[TMP1]]
41 // CPP-CHECK-LABEL: @_Z20test_svhistcnt_s64_zu10__SVBool_tu11__SVInt64_tS0_(
42 // CPP-CHECK-NEXT: entry:
43 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
44 // CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
45 // CPP-CHECK-NEXT: ret <vscale x 2 x i64> [[TMP1]]
47 svuint64_t test_svhistcnt_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
49 return SVE_ACLE_FUNC(svhistcnt,_s64,_z,)(pg, op1, op2);
52 // CHECK-LABEL: @test_svhistcnt_u32_z(
53 // CHECK-NEXT: entry:
54 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
55 // CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
56 // CHECK-NEXT: ret <vscale x 4 x i32> [[TMP1]]
58 // CPP-CHECK-LABEL: @_Z20test_svhistcnt_u32_zu10__SVBool_tu12__SVUint32_tS0_(
59 // CPP-CHECK-NEXT: entry:
60 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
61 // CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
62 // CPP-CHECK-NEXT: ret <vscale x 4 x i32> [[TMP1]]
64 svuint32_t test_svhistcnt_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
66 return SVE_ACLE_FUNC(svhistcnt,_u32,_z,)(pg, op1, op2);
69 // CHECK-LABEL: @test_svhistcnt_u64_z(
70 // CHECK-NEXT: entry:
71 // CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
72 // CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
73 // CHECK-NEXT: ret <vscale x 2 x i64> [[TMP1]]
75 // CPP-CHECK-LABEL: @_Z20test_svhistcnt_u64_zu10__SVBool_tu12__SVUint64_tS0_(
76 // CPP-CHECK-NEXT: entry:
77 // CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
78 // CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
79 // CPP-CHECK-NEXT: ret <vscale x 2 x i64> [[TMP1]]
81 svuint64_t test_svhistcnt_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
83 return SVE_ACLE_FUNC(svhistcnt,_u64,_z,)(pg, op1, op2);