ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());
    ASSERT_TRUE(absl::StrContains(device_name, "GPU:0")) << device_name;

    // .backing_device of shape is CPU since the tensor is backed by CPU
    backing_device_name =
        TFE_TensorHandleBackingDeviceName(retvals[0], status.get());
    ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

            static_cast<TF_Code>(tensorflow::error::FAILED_PRECONDITION));
  EXPECT_EQ(TF_ABORTED, static_cast<TF_Code>(tensorflow::error::ABORTED));
  EXPECT_EQ(TF_OUT_OF_RANGE,
            static_cast<TF_Code>(tensorflow::error::OUT_OF_RANGE));
  EXPECT_EQ(TF_UNIMPLEMENTED,
            static_cast<TF_Code>(tensorflow::error::UNIMPLEMENTED));
  EXPECT_EQ(TF_INTERNAL, static_cast<TF_Code>(tensorflow::error::INTERNAL));

                                       bool remote_func_outputs = false,
                                       bool has_packed_input = false) {
  return TestRemoteExecuteSilentCopies(async, remote, /*func=*/true,
                                       heavy_load_on_streaming_rpc,
                                       remote_func_outputs, has_packed_input);
}

TEST(CAPI, RemoteExecuteSilentCopiesAsyncFunc) {

Status ImmediateExecutionTensorHandle::SummarizeValue(
    std::string& summary) const {
  Status status;
  AbstractTensorPtr resolved(
      // TODO(allenl): Resolve should be const, and the caches that get updated
      // marked mutable.
      const_cast<ImmediateExecutionTensorHandle*>(this)->Resolve(&status));
  if (!status.ok()) {
    return status;
  }
  summary = resolved->SummarizeValue();
  return absl::OkStatus();
}

  TFE_TensorHandle* packed_handle =
      TFE_CreatePackedTensorHandle(ctx, handles.data(), &num_replicas, status);
  ASSERT_EQ(TF_GetCode(status), TF_OK) << TF_Message(status);
  EXPECT_EQ(TFE_TensorHandleDataType(packed_handle), TF_RESOURCE);
  EXPECT_EQ(TFE_TensorHandleNumDims(packed_handle, status), 0);
  EXPECT_EQ(TFE_TensorHandleNumElements(packed_handle, status), 1);

  if (*requested_placement == '\0') {
    TF_SetStatus(
        status, TF_INTERNAL,
        "Ops must be placed on the parallel device explicitly, or their inputs "
        "first un-packed. Got an un-placed op with an input placed on the "
        "parallel device.");
    return;
  }
  TFE_Context* context = TFE_OpGetContext(original_op, status);
  if (TF_GetCode(status) != TF_OK) return;

  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
  TF_GraphCopyFunction(host_graph_, grad_func, nullptr, s_);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

  // Check that functions are added but not linked
  GraphDef gdef;
  GetGraphDef(host_graph_, &gdef);
  std::vector<string> func_names = GetFuncNames(gdef);
  ASSERT_EQ(2, func_names.size());
  ASSERT_EQ("FooFunc", func_names[0]);