// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 6
// REQUIRES: amdgpu-registered-target
// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -target-cpu gfx1100 -emit-llvm -fcuda-is-device -o - %s | FileCheck %s

#define __device__ __attribute__((device))

typedef _Float16 __attribute__((ext_vector_type(2))) half2;

// CHECK-LABEL: define dso_local void @_Z29test_amdgcn_fdot2_f16_f16_hipPDF16_Dv2_DF16_S0_DF16_(
// CHECK-SAME: ptr noundef [[OUT:%.*]], <2 x half> noundef [[A:%.*]], <2 x half> noundef [[B:%.*]], half noundef [[C:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT:  [[ENTRY:.*:]]
// CHECK-NEXT:    [[OUT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
// CHECK-NEXT:    [[A_ADDR:%.*]] = alloca <2 x half>, align 4, addrspace(5)
// CHECK-NEXT:    [[B_ADDR:%.*]] = alloca <2 x half>, align 4, addrspace(5)
// CHECK-NEXT:    [[C_ADDR:%.*]] = alloca half, align 2, addrspace(5)
// CHECK-NEXT:    [[OUT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[OUT_ADDR]] to ptr
// CHECK-NEXT:    [[A_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A_ADDR]] to ptr
// CHECK-NEXT:    [[B_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[B_ADDR]] to ptr
// CHECK-NEXT:    [[C_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[C_ADDR]] to ptr
// CHECK-NEXT:    store ptr [[OUT]], ptr [[OUT_ADDR_ASCAST]], align 8
// CHECK-NEXT:    store <2 x half> [[A]], ptr [[A_ADDR_ASCAST]], align 4
// CHECK-NEXT:    store <2 x half> [[B]], ptr [[B_ADDR_ASCAST]], align 4
// CHECK-NEXT:    store half [[C]], ptr [[C_ADDR_ASCAST]], align 2
// CHECK-NEXT:    [[TMP0:%.*]] = load <2 x half>, ptr [[A_ADDR_ASCAST]], align 4
// CHECK-NEXT:    [[TMP1:%.*]] = load <2 x half>, ptr [[B_ADDR_ASCAST]], align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load half, ptr [[C_ADDR_ASCAST]], align 2
// CHECK-NEXT:    [[TMP3:%.*]] = call contract half @llvm.amdgcn.fdot2.f16.f16(<2 x half> [[TMP0]], <2 x half> [[TMP1]], half [[TMP2]])
// CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[OUT_ADDR_ASCAST]], align 8
// CHECK-NEXT:    store half [[TMP3]], ptr [[TMP4]], align 2
// CHECK-NEXT:    ret void
//
__device__ void test_amdgcn_fdot2_f16_f16_hip(_Float16* out, half2 a, half2 b, _Float16 c) {
  *out = __builtin_amdgcn_fdot2_f16_f16(a, b, c);
}