Vendor import of llvm release_40 branch r292009:vendor/llvm/llvm-release_40-r292009

https://llvm.org/svn/llvm-project/llvm/branches/release_40@292009

1 files changed, 16 insertions, 0 deletions

diff --git a/test/CodeGen/AMDGPU/fp16_to_fp64.ll b/test/CodeGen/AMDGPU/fp16_to_fp64.ll
new file mode 100644
index 000000000000..8b05d7b88a10
--- /dev/null
+++ b/test/CodeGen/AMDGPU/fp16_to_fp64.ll
@@ -0,0 +1,16 @@
+; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+
+declare double @llvm.convert.from.fp16.f64(i16) nounwind readnone
+
+; FUNC-LABEL: {{^}}test_convert_fp16_to_fp64:
+; GCN: buffer_load_ushort [[VAL:v[0-9]+]]
+; GCN: v_cvt_f32_f16_e32 [[RESULT32:v[0-9]+]], [[VAL]]
+; GCN: v_cvt_f64_f32_e32 [[RESULT:v\[[0-9]+:[0-9]+\]]], [[RESULT32]]
+; GCN: buffer_store_dwordx2 [[RESULT]]
+define void @test_convert_fp16_to_fp64(double addrspace(1)* noalias %out, i16 addrspace(1)* noalias %in) nounwind {
+  %val = load i16, i16 addrspace(1)* %in, align 2
+  %cvt = call double @llvm.convert.from.fp16.f64(i16 %val) nounwind readnone
+  store double %cvt, double addrspace(1)* %out, align 4
+  ret void
+}