int m = (end + start) >>> 1;
        return listProduct(nums, start, m).multiply(listProduct(nums, m, end));
    }
  }

  /**
   * Returns {@code n} choose {@code k}, also known as the binomial coefficient of {@code n} and
   * {@code k}, that is, {@code n! / (k! (n - k)!)}.
   *
   * <p><b>Warning:</b> the result can take as much as <i>O(k log n)</i> space.
   *

    1 * 2 * 3 * 4 * 5 * 6 * 7 * 8 * 9 * 10 * 11,
    1 * 2 * 3 * 4 * 5 * 6 * 7 * 8 * 9 * 10 * 11 * 12
  };

  /**
   * Returns {@code n} choose {@code k}, also known as the binomial coefficient of {@code n} and
   * {@code k}, or {@link Integer#MAX_VALUE} if the result does not fit in an {@code int}.
   *
   * @throws IllegalArgumentException if {@code n < 0}, {@code k < 0} or {@code k > n}
   */

    1L * 2 * 3 * 4 * 5 * 6 * 7 * 8 * 9 * 10 * 11 * 12 * 13 * 14 * 15 * 16 * 17 * 18 * 19 * 20
  };

  /**
   * Returns {@code n} choose {@code k}, also known as the binomial coefficient of {@code n} and
   * {@code k}, or {@link Long#MAX_VALUE} if the result does not fit in a {@code long}.
   *
   * @throws IllegalArgumentException if {@code n < 0}, {@code k < 0}, or {@code k > n}
   */

        `tf.contrib.data.sloppy_interleave()`.
*   Add `train_and_evaluate` for simple distributed `Estimator` training.
*   Add `tf.spectral.dct` for computing the DCT-II.
*   Add Mel-Frequency Cepstral Coefficient support to `tf.contrib.signal` (with
    GPU and gradient support).
*   Add a self-check on `import tensorflow` for Windows DLL issues.
*   Add NCHW support to `tf.depth_to_space` on GPU.