// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// REQUIRES: aarch64-registered-target

// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme2 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme2 -S -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 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sme2 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sme2 -S -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-none-linux-gnu -target-feature +sme2 -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,A5) A1##A3##A5
#else
#define SVE_ACLE_FUNC(A1,A2,A3,A4,A5) A1##A2##A3##A4##A5
#endif

//
// Multi, multi
// CHECK-LABEL: @test_svmla2_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZM]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla2_f16j13svfloat16x2_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16x2_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_f16,_vg2x2,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla2_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZM]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z16test_svmla2_bf16j14svbfloat16x2_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16x2_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_bf16,_vg2x2,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla2_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla2_u16j12svuint16x2_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16x2_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_u16,_vg2x2,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla2_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla2_s16j11svint16x2_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla2_s16(uint32_t slice_base, svint16x2_t zn, svint16x2_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_s16,_vg2x2,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla4_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 8)
// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 16)
// CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x half> [[TMP7]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla4_f16j13svfloat16x4_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 16)
// CPP-CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZM]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x half> [[TMP7]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16x4_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_f16,_vg2x4,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla4_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 8)
// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 16)
// CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x bfloat> [[TMP7]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z16test_svmla4_bf16j14svbfloat16x4_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 16)
// CPP-CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZM]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x bfloat> [[TMP7]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16x4_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_bf16,_vg2x4,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla4_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 8)
// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 16)
// CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i16> [[TMP7]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla4_u16j12svuint16x4_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 16)
// CPP-CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i16> [[TMP7]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16x4_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_u16,_vg2x4,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla4_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM:%.*]], i64 0)
// CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 8)
// CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 16)
// CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i16> [[TMP7]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z15test_svmla4_s16j11svint16x4_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP5:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 8)
// CPP-CHECK-NEXT:    [[TMP6:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 16)
// CPP-CHECK-NEXT:    [[TMP7:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZM]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i16> [[TMP7]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla4_s16(uint32_t slice_base, svint16x4_t zn, svint16x4_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_s16,_vg2x4,,)(slice_base, zn, zm);
}

//
// Multi, single
// CHECK-LABEL: @test_svmla_single1_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single1_f16ju13__SVFloat16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_f16,_vg2x1,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single1_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z23test_svmla_single1_bf16ju14__SVBfloat16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_bf16,_vg2x1,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single1_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single1_u16ju12__SVUint16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_u16,_vg2x1,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single1_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single1_s16ju11__SVInt16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_za32,_s16,_vg2x1,,)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single2_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single2_f16j13svfloat16x2_tu13__SVFloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_f16,_vg2x2)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single2_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z23test_svmla_single2_bf16j14svbfloat16x2_tu14__SVBfloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_bf16,_vg2x2)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single2_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single2_u16j12svuint16x2_tu12__SVUint16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_u16,_vg2x2)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single2_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single2_s16j11svint16x2_tu11__SVInt16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_s16,_vg2x2)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single4_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single4_f16j13svfloat16x4_tu13__SVFloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_f16,_vg2x4)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single4_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z23test_svmla_single4_bf16j14svbfloat16x4_tu14__SVBfloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.single.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_bf16,_vg2x4)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single4_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single4_u16j12svuint16x4_tu12__SVUint16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_u16,_vg2x4)(slice_base, zn, zm);
}

// CHECK-LABEL: @test_svmla_single4_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]])
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z22test_svmla_single4_s16j11svint16x4_tu11__SVInt16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]])
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla,_single,_za32,_s16,_vg2x4)(slice_base, zn, zm);
}

//
// Multi, indexed
//

// CHECK-LABEL: @test_svmla_lane1_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane1_f16ju13__SVFloat16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x1,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane1_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z21test_svmla_lane1_bf16ju14__SVBfloat16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x1,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane1_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane1_u16ju12__SVUint16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x1,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane1_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane1_s16ju11__SVInt16_tS_(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x1,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane2_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane2_f16j13svfloat16x2_tu13__SVFloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv16f16(<vscale x 16 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x2,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane2_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z21test_svmla_lane2_bf16j14svbfloat16x2_tu14__SVBfloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x2,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane2_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane2_u16j12svuint16x2_tu12__SVUint16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x2,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane2_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane2_s16j11svint16x2_tu11__SVInt16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv16i16(<vscale x 16 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x2,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane4_f16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane4_f16j13svfloat16x4_tu13__SVFloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x half> @llvm.vector.extract.nxv8f16.nxv32f16(<vscale x 32 x half> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x4,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane4_bf16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z21test_svmla_lane4_bf16j14svbfloat16x4_tu14__SVBfloat16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x bfloat> @llvm.vector.extract.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlal.lane.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x4,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane4_u16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane4_u16j12svuint16x4_tu12__SVUint16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlal.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x4,,)(slice_base, zn, zm, 7);
}

// CHECK-LABEL: @test_svmla_lane4_s16(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CHECK-NEXT:    ret void
//
// CPP-CHECK-LABEL: @_Z20test_svmla_lane4_s16j11svint16x4_tu11__SVInt16_t(
// CPP-CHECK-NEXT:  entry:
// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN:%.*]], i64 0)
// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 8)
// CPP-CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 16)
// CPP-CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 8 x i16> @llvm.vector.extract.nxv8i16.nxv32i16(<vscale x 32 x i16> [[ZN]], i64 24)
// CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlal.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
// CPP-CHECK-NEXT:    ret void
//
void test_svmla_lane4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
{
   SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x4,,)(slice_base, zn, zm, 7);
}