// can correspond to the data contained in another dimension in on-host
// representation. The dimensions are indexed using the standard TensorFlow
// major-to-minor order (slowest varying dimension first),
// not the XLA's minor-to-major order.
// On-device dimensions can be padded. TFE_TensorDebugInfoOnDeviceDim returns
// the number of elements in a dimension after padding.

  // otherwise be the name of this custom device. Ops are placed onto a custom
  // device if any of their inputs are on that custom device, but custom devices
  // are free to set a bad status in order to require explicit placement.
  void (*execute)(const TFE_Op* op, int* num_outputs,
                  TFE_TensorHandle** outputs, TF_Status* s, void* device_info);

  // Method to delete a device.

  // from a backward gradient function. Fills `out_grads` corresponding to
  // `output_tensors`. `out_grads` must not be null.
  //
  // Executes the backward function in order to trace its gradient, which will
  // waste computation if executing eagerly (when graph building the unneeded
  // computation is pruned). Temporarily sets `backward_tape` so that

@@ -47,29 +47,7 @@
 #endif
 
 #if _STRING_ARCH_unaligned
-/* If we can do unaligned memory accesses we must know the endianess.  */
-# include <endian.h>
 # include <bits/types.h>
-
-# if __BYTE_ORDER == __LITTLE_ENDIAN
-#  define __STRING2_SMALL_GET16(src, idx) \
-     (((const unsigned char *) (const char *) (src))[idx + 1] << 8	      \
-      | ((const unsigned char *) (const char *) (src))[idx])

//
// DeviceThread itself is thread-safe, in that StartExecute will block if there
// is a pending execution. Since StartExecute is equivalent to grabbing a lock,
// multiple DeviceThreads should always be accessed in the same order to avoid
// deadlocks.
class DeviceThread {
 public:
  // Starts a background thread waiting for `StartExecute`.
  explicit DeviceThread(const std::string& device, const bool is_async,

  // parallel.
  //
  // Conceptually this is a thread pool with one thread per device. It requires
  // less synchronization than a thread pool would for this task, since Execute
  // acquires each thread in order (and so only one Execute will schedule
  // blocking collective operations at a time), and avoids some dynamic
  // allocation/scheduling.
  //
  // TODO(allenl): Keep a map from outer thread to list of inner threads rather