clang/test/CodeGen/AArch64/sme2-intrinsics/acle_sme2_mlsl.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // REQUIRES: aarch64-registered-target
   3
   4 // RUN: %clang_cc1 -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
   5 // RUN: %clang_cc1 -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   6 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s
   7 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -p mem2reg,instcombine,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   8 // RUN: %clang_cc1 -triple aarch64 -target-feature +bf16 -target-feature +sme -target-feature +sme2 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
   9 #include <arm_sme.h>
  10
  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,A5) A1##A3##A5
  14 #else
  15 #define SVE_ACLE_FUNC(A1,A2,A3,A4,A5) A1##A2##A3##A4##A5
  16 #endif
  17
  18 //
  19 // Multi, multi
  20 // CHECK-LABEL: @test_svmls2_f16(
  21 // CHECK-NEXT:  entry:
  22 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM_COERCE0:%.*]], <vscale x 8 x half> [[ZM_COERCE1:%.*]])
  23 // CHECK-NEXT:    ret void
  24 //
  25 // CPP-CHECK-LABEL: @_Z15test_svmls2_f16j13svfloat16x2_tS_(
  26 // CPP-CHECK-NEXT:  entry:
  27 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM_COERCE0:%.*]], <vscale x 8 x half> [[ZM_COERCE1:%.*]])
  28 // CPP-CHECK-NEXT:    ret void
  29 //
  30 void test_svmls2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16x2_t zm) __arm_streaming __arm_inout("za")
  31 {
  32    SVE_ACLE_FUNC(svmls_za32,_f16,_vg2x2,,)(slice_base, zn, zm);
  33 }
  34
  35 // CHECK-LABEL: @test_svmls2_bf16(
  36 // CHECK-NEXT:  entry:
  37 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE1:%.*]])
  38 // CHECK-NEXT:    ret void
  39 //
  40 // CPP-CHECK-LABEL: @_Z16test_svmls2_bf16j14svbfloat16x2_tS_(
  41 // CPP-CHECK-NEXT:  entry:
  42 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE1:%.*]])
  43 // CPP-CHECK-NEXT:    ret void
  44 //
  45 void test_svmls2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16x2_t zm) __arm_streaming __arm_inout("za")
  46 {
  47    SVE_ACLE_FUNC(svmls_za32,_bf16,_vg2x2,,)(slice_base, zn, zm);
  48 }
  49
  50 // CHECK-LABEL: @test_svmls2_u16(
  51 // CHECK-NEXT:  entry:
  52 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]])
  53 // CHECK-NEXT:    ret void
  54 //
  55 // CPP-CHECK-LABEL: @_Z15test_svmls2_u16j12svuint16x2_tS_(
  56 // CPP-CHECK-NEXT:  entry:
  57 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]])
  58 // CPP-CHECK-NEXT:    ret void
  59 //
  60 void test_svmls2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16x2_t zm) __arm_streaming __arm_inout("za")
  61 {
  62    SVE_ACLE_FUNC(svmls_za32,_u16,_vg2x2,,)(slice_base, zn, zm);
  63 }
  64
  65 // CHECK-LABEL: @test_svmls2_s16(
  66 // CHECK-NEXT:  entry:
  67 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]])
  68 // CHECK-NEXT:    ret void
  69 //
  70 // CPP-CHECK-LABEL: @_Z15test_svmls2_s16j11svint16x2_tS_(
  71 // CPP-CHECK-NEXT:  entry:
  72 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]])
  73 // CPP-CHECK-NEXT:    ret void
  74 //
  75 void test_svmls2_s16(uint32_t slice_base, svint16x2_t zn, svint16x2_t zm) __arm_streaming __arm_inout("za")
  76 {
  77    SVE_ACLE_FUNC(svmls_za32,_s16,_vg2x2,,)(slice_base, zn, zm);
  78 }
  79
  80 // CHECK-LABEL: @test_svmls4_f16(
  81 // CHECK-NEXT:  entry:
  82 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM_COERCE0:%.*]], <vscale x 8 x half> [[ZM_COERCE1:%.*]], <vscale x 8 x half> [[ZM_COERCE2:%.*]], <vscale x 8 x half> [[ZM_COERCE3:%.*]])
  83 // CHECK-NEXT:    ret void
  84 //
  85 // CPP-CHECK-LABEL: @_Z15test_svmls4_f16j13svfloat16x4_tS_(
  86 // CPP-CHECK-NEXT:  entry:
  87 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM_COERCE0:%.*]], <vscale x 8 x half> [[ZM_COERCE1:%.*]], <vscale x 8 x half> [[ZM_COERCE2:%.*]], <vscale x 8 x half> [[ZM_COERCE3:%.*]])
  88 // CPP-CHECK-NEXT:    ret void
  89 //
  90 void test_svmls4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16x4_t zm) __arm_streaming __arm_inout("za")
  91 {
  92    SVE_ACLE_FUNC(svmls_za32,_f16,_vg2x4,,)(slice_base, zn, zm);
  93 }
  94
  95 // CHECK-LABEL: @test_svmls4_bf16(
  96 // CHECK-NEXT:  entry:
  97 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE3:%.*]])
  98 // CHECK-NEXT:    ret void
  99 //
 100 // CPP-CHECK-LABEL: @_Z16test_svmls4_bf16j14svbfloat16x4_tS_(
 101 // CPP-CHECK-NEXT:  entry:
 102 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZM_COERCE3:%.*]])
 103 // CPP-CHECK-NEXT:    ret void
 104 //
 105 void test_svmls4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16x4_t zm) __arm_streaming __arm_inout("za")
 106 {
 107    SVE_ACLE_FUNC(svmls_za32,_bf16,_vg2x4,,)(slice_base, zn, zm);
 108 }
 109
 110 // CHECK-LABEL: @test_svmls4_u16(
 111 // CHECK-NEXT:  entry:
 112 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE2:%.*]], <vscale x 8 x i16> [[ZM_COERCE3:%.*]])
 113 // CHECK-NEXT:    ret void
 114 //
 115 // CPP-CHECK-LABEL: @_Z15test_svmls4_u16j12svuint16x4_tS_(
 116 // CPP-CHECK-NEXT:  entry:
 117 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE2:%.*]], <vscale x 8 x i16> [[ZM_COERCE3:%.*]])
 118 // CPP-CHECK-NEXT:    ret void
 119 //
 120 void test_svmls4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16x4_t zm) __arm_streaming __arm_inout("za")
 121 {
 122    SVE_ACLE_FUNC(svmls_za32,_u16,_vg2x4,,)(slice_base, zn, zm);
 123 }
 124
 125 // CHECK-LABEL: @test_svmls4_s16(
 126 // CHECK-NEXT:  entry:
 127 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE2:%.*]], <vscale x 8 x i16> [[ZM_COERCE3:%.*]])
 128 // CHECK-NEXT:    ret void
 129 //
 130 // CPP-CHECK-LABEL: @_Z15test_svmls4_s16j11svint16x4_tS_(
 131 // CPP-CHECK-NEXT:  entry:
 132 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM_COERCE0:%.*]], <vscale x 8 x i16> [[ZM_COERCE1:%.*]], <vscale x 8 x i16> [[ZM_COERCE2:%.*]], <vscale x 8 x i16> [[ZM_COERCE3:%.*]])
 133 // CPP-CHECK-NEXT:    ret void
 134 //
 135 void test_svmls4_s16(uint32_t slice_base, svint16x4_t zn, svint16x4_t zm) __arm_streaming __arm_inout("za")
 136 {
 137    SVE_ACLE_FUNC(svmls_za32,_s16,_vg2x4,,)(slice_base, zn, zm);
 138 }
 139
 140 //
 141 // Multi, single
 142 // CHECK-LABEL: @test_svmls_single1_f16(
 143 // CHECK-NEXT:  entry:
 144 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 145 // CHECK-NEXT:    ret void
 146 //
 147 // CPP-CHECK-LABEL: @_Z22test_svmls_single1_f16ju13__SVFloat16_tS_(
 148 // CPP-CHECK-NEXT:  entry:
 149 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 150 // CPP-CHECK-NEXT:    ret void
 151 //
 152 void test_svmls_single1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 153 {
 154    SVE_ACLE_FUNC(svmls_za32,_f16,_vg2x1,,)(slice_base, zn, zm);
 155 }
 156
 157 // CHECK-LABEL: @test_svmls_single1_bf16(
 158 // CHECK-NEXT:  entry:
 159 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 160 // CHECK-NEXT:    ret void
 161 //
 162 // CPP-CHECK-LABEL: @_Z23test_svmls_single1_bf16ju14__SVBfloat16_tS_(
 163 // CPP-CHECK-NEXT:  entry:
 164 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 165 // CPP-CHECK-NEXT:    ret void
 166 //
 167 void test_svmls_single1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 168 {
 169    SVE_ACLE_FUNC(svmls_za32,_bf16,_vg2x1,,)(slice_base, zn, zm);
 170 }
 171
 172 // CHECK-LABEL: @test_svmls_single1_u16(
 173 // CHECK-NEXT:  entry:
 174 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 175 // CHECK-NEXT:    ret void
 176 //
 177 // CPP-CHECK-LABEL: @_Z22test_svmls_single1_u16ju12__SVUint16_tS_(
 178 // CPP-CHECK-NEXT:  entry:
 179 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 180 // CPP-CHECK-NEXT:    ret void
 181 //
 182 void test_svmls_single1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 183 {
 184    SVE_ACLE_FUNC(svmls_za32,_u16,_vg2x1,,)(slice_base, zn, zm);
 185 }
 186
 187 // CHECK-LABEL: @test_svmls_single1_s16(
 188 // CHECK-NEXT:  entry:
 189 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 190 // CHECK-NEXT:    ret void
 191 //
 192 // CPP-CHECK-LABEL: @_Z22test_svmls_single1_s16ju11__SVInt16_tS_(
 193 // CPP-CHECK-NEXT:  entry:
 194 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 195 // CPP-CHECK-NEXT:    ret void
 196 //
 197 void test_svmls_single1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 198 {
 199    SVE_ACLE_FUNC(svmls_za32,_s16,_vg2x1,,)(slice_base, zn, zm);
 200 }
 201
 202 // CHECK-LABEL: @test_svmls_single2_f16(
 203 // CHECK-NEXT:  entry:
 204 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 205 // CHECK-NEXT:    ret void
 206 //
 207 // CPP-CHECK-LABEL: @_Z22test_svmls_single2_f16j13svfloat16x2_tu13__SVFloat16_t(
 208 // CPP-CHECK-NEXT:  entry:
 209 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 210 // CPP-CHECK-NEXT:    ret void
 211 //
 212 void test_svmls_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 213 {
 214    SVE_ACLE_FUNC(svmls,_single,_za32,_f16,_vg2x2)(slice_base, zn, zm);
 215 }
 216
 217 // CHECK-LABEL: @test_svmls_single2_bf16(
 218 // CHECK-NEXT:  entry:
 219 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 220 // CHECK-NEXT:    ret void
 221 //
 222 // CPP-CHECK-LABEL: @_Z23test_svmls_single2_bf16j14svbfloat16x2_tu14__SVBfloat16_t(
 223 // CPP-CHECK-NEXT:  entry:
 224 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 225 // CPP-CHECK-NEXT:    ret void
 226 //
 227 void test_svmls_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 228 {
 229    SVE_ACLE_FUNC(svmls,_single,_za32,_bf16,_vg2x2)(slice_base, zn, zm);
 230 }
 231
 232 // CHECK-LABEL: @test_svmls_single2_u16(
 233 // CHECK-NEXT:  entry:
 234 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 235 // CHECK-NEXT:    ret void
 236 //
 237 // CPP-CHECK-LABEL: @_Z22test_svmls_single2_u16j12svuint16x2_tu12__SVUint16_t(
 238 // CPP-CHECK-NEXT:  entry:
 239 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 240 // CPP-CHECK-NEXT:    ret void
 241 //
 242 void test_svmls_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 243 {
 244    SVE_ACLE_FUNC(svmls,_single,_za32,_u16,_vg2x2)(slice_base, zn, zm);
 245 }
 246
 247 // CHECK-LABEL: @test_svmls_single2_s16(
 248 // CHECK-NEXT:  entry:
 249 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 250 // CHECK-NEXT:    ret void
 251 //
 252 // CPP-CHECK-LABEL: @_Z22test_svmls_single2_s16j11svint16x2_tu11__SVInt16_t(
 253 // CPP-CHECK-NEXT:  entry:
 254 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 255 // CPP-CHECK-NEXT:    ret void
 256 //
 257 void test_svmls_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 258 {
 259    SVE_ACLE_FUNC(svmls,_single,_za32,_s16,_vg2x2)(slice_base, zn, zm);
 260 }
 261
 262 // CHECK-LABEL: @test_svmls_single4_f16(
 263 // CHECK-NEXT:  entry:
 264 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 265 // CHECK-NEXT:    ret void
 266 //
 267 // CPP-CHECK-LABEL: @_Z22test_svmls_single4_f16j13svfloat16x4_tu13__SVFloat16_t(
 268 // CPP-CHECK-NEXT:  entry:
 269 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM:%.*]])
 270 // CPP-CHECK-NEXT:    ret void
 271 //
 272 void test_svmls_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 273 {
 274    SVE_ACLE_FUNC(svmls,_single,_za32,_f16,_vg2x4)(slice_base, zn, zm);
 275 }
 276
 277 // CHECK-LABEL: @test_svmls_single4_bf16(
 278 // CHECK-NEXT:  entry:
 279 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 280 // CHECK-NEXT:    ret void
 281 //
 282 // CPP-CHECK-LABEL: @_Z23test_svmls_single4_bf16j14svbfloat16x4_tu14__SVBfloat16_t(
 283 // CPP-CHECK-NEXT:  entry:
 284 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.single.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]])
 285 // CPP-CHECK-NEXT:    ret void
 286 //
 287 void test_svmls_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 288 {
 289    SVE_ACLE_FUNC(svmls,_single,_za32,_bf16,_vg2x4)(slice_base, zn, zm);
 290 }
 291
 292 // CHECK-LABEL: @test_svmls_single4_u16(
 293 // CHECK-NEXT:  entry:
 294 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 295 // CHECK-NEXT:    ret void
 296 //
 297 // CPP-CHECK-LABEL: @_Z22test_svmls_single4_u16j12svuint16x4_tu12__SVUint16_t(
 298 // CPP-CHECK-NEXT:  entry:
 299 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 300 // CPP-CHECK-NEXT:    ret void
 301 //
 302 void test_svmls_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 303 {
 304    SVE_ACLE_FUNC(svmls,_single,_za32,_u16,_vg2x4)(slice_base, zn, zm);
 305 }
 306
 307 // CHECK-LABEL: @test_svmls_single4_s16(
 308 // CHECK-NEXT:  entry:
 309 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 310 // CHECK-NEXT:    ret void
 311 //
 312 // CPP-CHECK-LABEL: @_Z22test_svmls_single4_s16j11svint16x4_tu11__SVInt16_t(
 313 // CPP-CHECK-NEXT:  entry:
 314 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.single.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]])
 315 // CPP-CHECK-NEXT:    ret void
 316 //
 317 void test_svmls_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 318 {
 319    SVE_ACLE_FUNC(svmls,_single,_za32,_s16,_vg2x4)(slice_base, zn, zm);
 320 }
 321
 322 //
 323 // Multi, indexed
 324 //
 325
 326 // CHECK-LABEL: @test_svmls_lane1_f16(
 327 // CHECK-NEXT:  entry:
 328 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 329 // CHECK-NEXT:    ret void
 330 //
 331 // CPP-CHECK-LABEL: @_Z20test_svmls_lane1_f16ju13__SVFloat16_tS_(
 332 // CPP-CHECK-NEXT:  entry:
 333 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x1.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 334 // CPP-CHECK-NEXT:    ret void
 335 //
 336 void test_svmls_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 337 {
 338    SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x1,,)(slice_base, zn, zm, 7);
 339 }
 340
 341 // CHECK-LABEL: @test_svmls_lane1_bf16(
 342 // CHECK-NEXT:  entry:
 343 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 344 // CHECK-NEXT:    ret void
 345 //
 346 // CPP-CHECK-LABEL: @_Z21test_svmls_lane1_bf16ju14__SVBfloat16_tS_(
 347 // CPP-CHECK-NEXT:  entry:
 348 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x1.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 349 // CPP-CHECK-NEXT:    ret void
 350 //
 351 void test_svmls_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 352 {
 353    SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x1,,)(slice_base, zn, zm, 7);
 354 }
 355
 356 // CHECK-LABEL: @test_svmls_lane1_u16(
 357 // CHECK-NEXT:  entry:
 358 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 359 // CHECK-NEXT:    ret void
 360 //
 361 // CPP-CHECK-LABEL: @_Z20test_svmls_lane1_u16ju12__SVUint16_tS_(
 362 // CPP-CHECK-NEXT:  entry:
 363 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 364 // CPP-CHECK-NEXT:    ret void
 365 //
 366 void test_svmls_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 367 {
 368    SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x1,,)(slice_base, zn, zm, 7);
 369 }
 370
 371 // CHECK-LABEL: @test_svmls_lane1_s16(
 372 // CHECK-NEXT:  entry:
 373 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 374 // CHECK-NEXT:    ret void
 375 //
 376 // CPP-CHECK-LABEL: @_Z20test_svmls_lane1_s16ju11__SVInt16_tS_(
 377 // CPP-CHECK-NEXT:  entry:
 378 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x1.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 379 // CPP-CHECK-NEXT:    ret void
 380 //
 381 void test_svmls_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 382 {
 383    SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x1,,)(slice_base, zn, zm, 7);
 384 }
 385
 386 // CHECK-LABEL: @test_svmls_lane2_f16(
 387 // CHECK-NEXT:  entry:
 388 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 389 // CHECK-NEXT:    ret void
 390 //
 391 // CPP-CHECK-LABEL: @_Z20test_svmls_lane2_f16j13svfloat16x2_tu13__SVFloat16_t(
 392 // CPP-CHECK-NEXT:  entry:
 393 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x2.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 394 // CPP-CHECK-NEXT:    ret void
 395 //
 396 void test_svmls_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 397 {
 398    SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x2,,)(slice_base, zn, zm, 7);
 399 }
 400
 401 // CHECK-LABEL: @test_svmls_lane2_bf16(
 402 // CHECK-NEXT:  entry:
 403 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 404 // CHECK-NEXT:    ret void
 405 //
 406 // CPP-CHECK-LABEL: @_Z21test_svmls_lane2_bf16j14svbfloat16x2_tu14__SVBfloat16_t(
 407 // CPP-CHECK-NEXT:  entry:
 408 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x2.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 409 // CPP-CHECK-NEXT:    ret void
 410 //
 411 void test_svmls_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 412 {
 413    SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x2,,)(slice_base, zn, zm, 7);
 414 }
 415
 416 // CHECK-LABEL: @test_svmls_lane2_u16(
 417 // CHECK-NEXT:  entry:
 418 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 419 // CHECK-NEXT:    ret void
 420 //
 421 // CPP-CHECK-LABEL: @_Z20test_svmls_lane2_u16j12svuint16x2_tu12__SVUint16_t(
 422 // CPP-CHECK-NEXT:  entry:
 423 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 424 // CPP-CHECK-NEXT:    ret void
 425 //
 426 void test_svmls_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 427 {
 428    SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x2,,)(slice_base, zn, zm, 7);
 429 }
 430
 431 // CHECK-LABEL: @test_svmls_lane2_s16(
 432 // CHECK-NEXT:  entry:
 433 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 434 // CHECK-NEXT:    ret void
 435 //
 436 // CPP-CHECK-LABEL: @_Z20test_svmls_lane2_s16j11svint16x2_tu11__SVInt16_t(
 437 // CPP-CHECK-NEXT:  entry:
 438 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x2.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 439 // CPP-CHECK-NEXT:    ret void
 440 //
 441 void test_svmls_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 442 {
 443    SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x2,,)(slice_base, zn, zm, 7);
 444 }
 445
 446 // CHECK-LABEL: @test_svmls_lane4_f16(
 447 // CHECK-NEXT:  entry:
 448 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 449 // CHECK-NEXT:    ret void
 450 //
 451 // CPP-CHECK-LABEL: @_Z20test_svmls_lane4_f16j13svfloat16x4_tu13__SVFloat16_t(
 452 // CPP-CHECK-NEXT:  entry:
 453 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x4.nxv8f16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x half> [[ZN_COERCE0:%.*]], <vscale x 8 x half> [[ZN_COERCE1:%.*]], <vscale x 8 x half> [[ZN_COERCE2:%.*]], <vscale x 8 x half> [[ZN_COERCE3:%.*]], <vscale x 8 x half> [[ZM:%.*]], i32 7)
 454 // CPP-CHECK-NEXT:    ret void
 455 //
 456 void test_svmls_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_inout("za")
 457 {
 458    SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x4,,)(slice_base, zn, zm, 7);
 459 }
 460
 461 // CHECK-LABEL: @test_svmls_lane4_bf16(
 462 // CHECK-NEXT:  entry:
 463 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 464 // CHECK-NEXT:    ret void
 465 //
 466 // CPP-CHECK-LABEL: @_Z21test_svmls_lane4_bf16j14svbfloat16x4_tu14__SVBfloat16_t(
 467 // CPP-CHECK-NEXT:  entry:
 468 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.fmlsl.lane.vg2x4.nxv8bf16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE0:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE1:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE2:%.*]], <vscale x 8 x bfloat> [[ZN_COERCE3:%.*]], <vscale x 8 x bfloat> [[ZM:%.*]], i32 7)
 469 // CPP-CHECK-NEXT:    ret void
 470 //
 471 void test_svmls_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_inout("za")
 472 {
 473    SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x4,,)(slice_base, zn, zm, 7);
 474 }
 475
 476 // CHECK-LABEL: @test_svmls_lane4_u16(
 477 // CHECK-NEXT:  entry:
 478 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 479 // CHECK-NEXT:    ret void
 480 //
 481 // CPP-CHECK-LABEL: @_Z20test_svmls_lane4_u16j12svuint16x4_tu12__SVUint16_t(
 482 // CPP-CHECK-NEXT:  entry:
 483 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.umlsl.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 484 // CPP-CHECK-NEXT:    ret void
 485 //
 486 void test_svmls_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_inout("za")
 487 {
 488    SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x4,,)(slice_base, zn, zm, 7);
 489 }
 490
 491 // CHECK-LABEL: @test_svmls_lane4_s16(
 492 // CHECK-NEXT:  entry:
 493 // CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 494 // CHECK-NEXT:    ret void
 495 //
 496 // CPP-CHECK-LABEL: @_Z20test_svmls_lane4_s16j11svint16x4_tu11__SVInt16_t(
 497 // CPP-CHECK-NEXT:  entry:
 498 // CPP-CHECK-NEXT:    tail call void @llvm.aarch64.sme.smlsl.lane.vg2x4.nxv8i16(i32 [[SLICE_BASE:%.*]], <vscale x 8 x i16> [[ZN_COERCE0:%.*]], <vscale x 8 x i16> [[ZN_COERCE1:%.*]], <vscale x 8 x i16> [[ZN_COERCE2:%.*]], <vscale x 8 x i16> [[ZN_COERCE3:%.*]], <vscale x 8 x i16> [[ZM:%.*]], i32 7)
 499 // CPP-CHECK-NEXT:    ret void
 500 //
 501 void test_svmls_lane4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_inout("za")
 502 {
 503    SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x4,,)(slice_base, zn, zm, 7);
 504 }