fixed multiple definition issue of bfp16/fp32 conversion function when building ckProfiler (#51)

* fixed bfloat16 issues

* refactor type_convert

Co-authored-by: Chao Liu <chao.liu2@amd.com>

[ROCm/composable_kernel commit: 0a66c54e95]

host/host_tensor/src/host_tensor.cpp

@@ -61,3 +61,13 @@ void ostream_HostTensorDescriptor(const HostTensorDescriptor& desc, std::ostream
     LogRange(os, desc.GetStrides(), ", ");
     os << "}" << std::endl;
 }
+
+float bf16_to_f32_(ushort src_val)
+{
+    union
+    {
+        uint32_t int32;
+        float fp32;
+    } u = {uint32_t(src_val) << 16};
+    return u.fp32;
+}