* method is overflow resilient.
   *
   * @since 14.0
   */
  public static int mean(int x, int y) {
    // Efficient method for computing the arithmetic mean.
    // The alternative (x + y) / 2 fails for large values.
    // The alternative (x + y) >>> 1 fails for negative values.
    return (x & y) + ((x ^ y) >> 1);
  }

  /**

#### `lru_cache` Technical Details

`@lru_cache` modifies the function it decorates to return the same value that was returned the first time, instead of computing it again, executing the code of the function every time.

     * is an ArrayList of SIDs represents the local groups that the account is
     * a member of.
     * <p/>
     * This method is designed to assist with computing access control for a
     * given user when the target object's ACL has local groups. Local groups
     * are not listed in a user's group membership (e.g. as represented by the

        PartialTensorShape first_shape;
        TF_RETURN_IF_ERROR(unwrap(tensors_[0].get())->Shape(&first_shape));
        return errors::Unimplemented(absl::StrCat(
            "Computing the shape of a ParallelTensor when the components do "
            "not all have the same rank is not supported. One tensor had "
            "shape ",
            first_shape.DebugString(), " and another had shape ",

For example:

* **Audio** or **image processing**.
* **Computer vision**: an image is composed of millions of pixels, each pixel has 3 values / colors, processing that normally requires computing something on those pixels, all at the same time.
* **Machine Learning**: it normally requires lots of "matrix" and "vector" multiplications. Think of a huge spreadsheet with numbers and multiplying all of them together at the same time.

//
// First we filter the tape to just the subset of operations we want to
// differentiate. In the process of doing so we count how many times each Tensor
// is used as an input to an op (so we know when we're done computing gradients
// for that Tensor). We also count, for each tape entry, how many of its output
// Tensors need gradients to be computed (Tensors which are not used do not need
// any gradients to be computed).
//

    cache.put("a", "b");
    cache.asMap().computeIfAbsent("a", k -> "c");
    assertTrue(notifications.toString(), notifications.isEmpty());
  }

  public void testCopyEntry_computing() {
    final CountDownLatch startSignal = new CountDownLatch(1);
    final CountDownLatch computingSignal = new CountDownLatch(1);
    final CountDownLatch doneSignal = new CountDownLatch(2);

    map.put("quux", "quux");
    assertFalse(map.values() instanceof Set);
    assertTrue(map.values().removeAll(ImmutableSet.of("bar")));
    assertEquals(1, map.size());
  }

  public void testCopyEntry_computing() {
    final CountDownLatch startSignal = new CountDownLatch(1);
    final CountDownLatch computingSignal = new CountDownLatch(1);
    final CountDownLatch doneSignal = new CountDownLatch(2);

   * method is resilient to overflow.
   *
   * @since 14.0
   */
  public static long mean(long x, long y) {
    // Efficient method for computing the arithmetic mean.
    // The alternative (x + y) / 2 fails for large values.
    // The alternative (x + y) >>> 1 fails for negative values.
    return (x & y) + ((x ^ y) >> 1);
  }

  /*

* Add a Deployment example for kubectl expose. ([#23222](https://github.com/kubernetes/kubernetes/pull/23222), [@madhusudancs](https://github.com/madhusudancs))
* Use versioned object when computing patch ([#23145](https://github.com/kubernetes/kubernetes/pull/23145), [@liggitt](https://github.com/liggitt))