// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// REQUIRES: aarch64-registered-target
// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +bf16 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +bf16 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
// RUN: %clang_cc1 -D__SVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +bf16 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
// RUN: %clang_cc1 -D__SVE_OVERLOADED_FORMS -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +bf16 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
// RUN: %clang_cc1 -triple aarch64 -target-feature +sme -target-feature +sme2 -target-feature +bf16 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
#include <arm_sme.h>
#ifdef SVE_OVERLOADED_FORMS
// A simple used,unused... macro, long enough to represent any SVE builtin.
#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
#else
#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
#endif
// CHECK-LABEL: @test_cvt_f16_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.x2.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret <vscale x 8 x half> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z15test_cvt_f16_x213svfloat32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.x2.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x half> [[TMP0]]
//
svfloat16_t test_cvt_f16_x2(svfloat32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f16,_f32_x2,,)(zn);
}
// CHECK-LABEL: @test_cvt_bf16_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.bfcvt.x2(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret <vscale x 8 x bfloat> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z16test_cvt_bf16_x213svfloat32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.bfcvt.x2(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x bfloat> [[TMP0]]
//
svbfloat16_t test_cvt_bf16_x2(svfloat32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_bf16,_f32_x2,,)(zn);
}
// x2
// CHECK-LABEL: @test_svcvt_f32_u32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ucvtf.x2.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_f32_u32_x212svuint32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ucvtf.x2.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
svfloat32x2_t test_svcvt_f32_u32_x2(svuint32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f32,_u32_x2,,)(zn);
}
// CHECK-LABEL: @test_svcvt_f32_s32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.scvtf.x2.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_f32_s32_x211svint32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.scvtf.x2.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
svfloat32x2_t test_svcvt_f32_s32_x2(svint32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f32,_s32_x2,,)(zn);
}
// CHECK-LABEL: @test_svcvt_u32_f32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzu.x2.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_u32_f32_x213svfloat32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzu.x2.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
svuint32x2_t test_svcvt_u32_f32_x2(svfloat32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_u32,_f32_x2,,)(zn);
}
// CHECK-LABEL: @test_svcvt_s32_f32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzs.x2.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_s32_f32_x213svfloat32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzs.x2.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
svint32x2_t test_svcvt_s32_f32_x2(svfloat32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_s32,_f32_x2,,)(zn);
}
// x4
// CHECK-LABEL: @test_svcvt_f32_u32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ucvtf.x4.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_f32_u32_x412svuint32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ucvtf.x4.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
svfloat32x4_t test_svcvt_f32_u32_x4(svuint32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f32,_u32_x4,,)(zn);
}
// CHECK-LABEL: @test_svcvt_f32_s32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.scvtf.x4.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_f32_s32_x411svint32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.scvtf.x4.nxv4f32.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
svfloat32x4_t test_svcvt_f32_s32_x4(svint32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f32,_s32_x4,,)(zn);
}
// CHECK-LABEL: @test_svcvt_u32_f32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzu.x4.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]], <vscale x 4 x float> [[ZN_COERCE2:%.*]], <vscale x 4 x float> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_u32_f32_x413svfloat32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzu.x4.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]], <vscale x 4 x float> [[ZN_COERCE2:%.*]], <vscale x 4 x float> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
svuint32x4_t test_svcvt_u32_f32_x4(svfloat32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_u32,_f32_x4,,)(zn);
}
// CHECK-LABEL: @test_svcvt_s32_f32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzs.x4.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]], <vscale x 4 x float> [[ZN_COERCE2:%.*]], <vscale x 4 x float> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z21test_svcvt_s32_f32_x413svfloat32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.fcvtzs.x4.nxv4i32.nxv4f32(<vscale x 4 x float> [[ZN_COERCE0:%.*]], <vscale x 4 x float> [[ZN_COERCE1:%.*]], <vscale x 4 x float> [[ZN_COERCE2:%.*]], <vscale x 4 x float> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } [[TMP0]]
//
svint32x4_t test_svcvt_s32_f32_x4(svfloat32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_s32,_f32_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_s16_s32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvt.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_s16_s32_x211svint32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvt.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svint16_t test_qcvt_s16_s32_x2(svint32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_s16,_s32_x2,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u16_u32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqcvt.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_u16_u32_x212svuint32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqcvt.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svuint16_t test_qcvt_u16_u32_x2(svuint32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u16,_u32_x2,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u16_s32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvtu.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_u16_s32_x211svint32x2_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvtu.x2.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svuint16_t test_qcvt_u16_s32_x2(svint32x2_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u16,_s32_x2,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u8_u32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqcvt.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z19test_qcvt_u8_u32_x412svuint32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqcvt.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
svuint8_t test_qcvt_u8_u32_x4(svuint32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u8,_u32_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u16_u64_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqcvt.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_u16_u64_x412svuint64x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqcvt.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svuint16_t test_qcvt_u16_u64_x4(svuint64x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u16,_u64_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_s8_s32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqcvt.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z19test_qcvt_s8_s32_x411svint32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqcvt.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
svint8_t test_qcvt_s8_s32_x4(svint32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_s8,_s32_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_s16_s64_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvt.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_s16_s64_x411svint64x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvt.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svint16_t test_qcvt_s16_s64_x4(svint64x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_s16,_s64_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u8_s32_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqcvtu.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z19test_qcvt_u8_s32_x411svint32x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqcvtu.x4.nxv4i32(<vscale x 4 x i32> [[ZN_COERCE0:%.*]], <vscale x 4 x i32> [[ZN_COERCE1:%.*]], <vscale x 4 x i32> [[ZN_COERCE2:%.*]], <vscale x 4 x i32> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 16 x i8> [[TMP0]]
//
svuint8_t test_qcvt_u8_s32_x4(svint32x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u8,_s32_x4,,)(zn);
}
// CHECK-LABEL: @test_qcvt_u16_s64_x4(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvtu.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z20test_qcvt_u16_s64_x411svint64x4_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqcvtu.x4.nxv2i64(<vscale x 2 x i64> [[ZN_COERCE0:%.*]], <vscale x 2 x i64> [[ZN_COERCE1:%.*]], <vscale x 2 x i64> [[ZN_COERCE2:%.*]], <vscale x 2 x i64> [[ZN_COERCE3:%.*]])
// CPP-CHECK-NEXT: ret <vscale x 8 x i16> [[TMP0]]
//
svuint16_t test_qcvt_u16_s64_x4(svint64x4_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svqcvt_u16,_s64_x4,,)(zn);
}
// CHECK-LABEL: @test_cvt_f32_x2(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.fcvt.widen.x2.nxv4f32(<vscale x 8 x half> [[ZN:%.*]])
// CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z15test_cvt_f32_x2u13__SVFloat16_t(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.fcvt.widen.x2.nxv4f32(<vscale x 8 x half> [[ZN:%.*]])
// CPP-CHECK-NEXT: ret { <vscale x 4 x float>, <vscale x 4 x float> } [[TMP0]]
//
__attribute__((target("sme-f16f16"))) svfloat32x2_t test_cvt_f32_x2(svfloat16_t zn) __arm_streaming {
return SVE_ACLE_FUNC(svcvt_f32,_f16_x2,,)(zn);
}