clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
   2 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   3 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   4 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s
   5 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -S -disable-O0-optnone -Werror -Wall -emit-llvm -o - -x c++ %s | opt -S -passes=mem2reg,tailcallelim | FileCheck %s -check-prefix=CPP-CHECK
   6
   7 // REQUIRES: aarch64-registered-target
   8
   9 #include <arm_sve.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) A1##A3
  14 #else
  15 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
  16 #endif
  17
  18 // CHECK-LABEL: @test_svcvt_bf16_f32_x(
  19 // CHECK-NEXT:  entry:
  20 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  21 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  22 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  23 //
  24 // CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_xu10__SVBool_tu13__SVFloat32_t(
  25 // CPP-CHECK-NEXT:  entry:
  26 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  27 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  28 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  29 //
  30 svbfloat16_t test_svcvt_bf16_f32_x(svbool_t pg, svfloat32_t op) {
  31   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _x, )(pg, op);
  32 }
  33
  34 // CHECK-LABEL: @test_svcvt_bf16_f32_z(
  35 // CHECK-NEXT:  entry:
  36 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  37 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  38 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  39 //
  40 // CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_zu10__SVBool_tu13__SVFloat32_t(
  41 // CPP-CHECK-NEXT:  entry:
  42 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  43 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  44 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  45 //
  46 svbfloat16_t test_svcvt_bf16_f32_z(svbool_t pg, svfloat32_t op) {
  47   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _z, )(pg, op);
  48 }
  49
  50 // CHECK-LABEL: @test_svcvt_bf16_f32_m(
  51 // CHECK-NEXT:  entry:
  52 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  53 // CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  54 // CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  55 //
  56 // CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_mu14__SVBfloat16_tu10__SVBool_tu13__SVFloat32_t(
  57 // CPP-CHECK-NEXT:  entry:
  58 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
  59 // CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
  60 // CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
  61 //
  62 svbfloat16_t test_svcvt_bf16_f32_m(svbfloat16_t inactive, svbool_t pg, svfloat32_t op) {
  63   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _m, )(inactive, pg, op);
  64 }