// a copy, returns the destination device of the copy. Note that the returned
// device name is not always the device holding the tensor handle's memory. If
// you want the latter, use TFE_TensorHandleBackingDeviceName. This function
// will block till the operation that produces `h` has completed.
TF_CAPI_EXPORT extern const char* TFE_TensorHandleDeviceName(
    TFE_TensorHandle* h, TF_Status* status);

  //
  // Like GradientTape::RecordOperation, this method takes the operation type
  // `op_type` (e.g. "Add"), the operation's inputs (`input_tensors`,
  // `input_tensor_id`, and `input_dtypes`; the latter two are somewhat
  // redundant but taken as arguments to avoid repeatedly fetching these values
  // between calls to ShouldRecord and Accumulator), and its outputs
  // (`output_tensors`).
  //

        // but need special handling for gradients (gradient of copy-on is not
        // just copy-off but includes a sum) and consideration of performance.
        //
        // TODO(allenl): There may be smarter ways to do this copy in some
        // cases, i.e. with a collective broadcast. We'll need to be careful
        // about things that are taken as inputs on the host or on their