xref: /llvm-project/llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll (revision 9e9907f1cfa424366fba58d9520f9305b537cec9)
1; RUN: llc -mtriple=r600 -mcpu=redwood -verify-machineinstrs < %s
2
3declare float @llvm.r600.dot4(<4 x float>, <4 x float>) nounwind readnone
4
5define amdgpu_kernel void @test_dp4(ptr addrspace(1) %out, ptr addrspace(1) %a, ptr addrspace(1) %b) nounwind {
6  %src0 = load <4 x float>, ptr addrspace(1) %a, align 16
7  %src1 = load <4 x float>, ptr addrspace(1) %b, align 16
8  %dp4 = call float @llvm.r600.dot4(<4 x float> %src0, <4 x float> %src1) nounwind readnone
9  store float %dp4, ptr addrspace(1) %out, align 4
10  ret void
11}
12