TensorFlow: more features, performance improvements, and doc fixes.

Changes:
- Add Split/Concat() methods to TensorUtil (meant for convenience, not
  speed) by Chris.

- Changes to linear algebra ops interface by Rasmus

- Tests for tensorboard by Daniel

- Fix bug in histogram calculation by Cassandra

- Added tool for backwards compatibility of OpDefs.  Tool
  Checks in history of opdefs and their changes, checks for
  backwards-incompatible changes.  All done by @josh11b

- Fix some protobuf example proto docs by Oliver

- Add derivative of MatrixDeterminant by @yaroslavvb

- Add a priority queue queue by @ebrevdo

- Doc and typo fixes by Aurelien and @dave-andersen

- Speed improvements to ConvBackwardFilter by @andydavis

- Improve speed of Alexnet on TitanX by @zheng-xq

- Add some host memory annotations to some GPU kernels by Yuan.

- Add support for doubles in histogram summary by @jmchen-g

Base CL: 108158338

1 files changed, 45 insertions, 67 deletions

diff --git a/tensorflow/g3doc/api_docs/python/constant_op.md b/tensorflow/g3doc/api_docs/python/constant_op.md
index 0e3abf6676..13aa64aef7 100644
--- a/tensorflow/g3doc/api_docs/python/constant_op.md
+++ b/tensorflow/g3doc/api_docs/python/constant_op.md
@@ -1,41 +1,19 @@
 <!-- This file is machine generated: DO NOT EDIT! -->
 
-# Constants, Sequences, and Random Values <a class="md-anchor" id="AUTOGENERATED-constants--sequences--and-random-values"></a>
+# Constants, Sequences, and Random Values
 
 Note: Functions taking `Tensor` arguments can also take anything accepted by
 [`tf.convert_to_tensor`](../../api_docs/python/framework.md#convert_to_tensor).
 
-<!-- TOC-BEGIN This section is generated by neural network: DO NOT EDIT! -->
-## Contents
-### [Constants, Sequences, and Random Values](#AUTOGENERATED-constants--sequences--and-random-values)
-* [Constant Value Tensors](#AUTOGENERATED-constant-value-tensors)
-  * [`tf.zeros(shape, dtype=tf.float32, name=None)`](#zeros)
-  * [`tf.zeros_like(tensor, dtype=None, name=None)`](#zeros_like)
-  * [`tf.ones(shape, dtype=tf.float32, name=None)`](#ones)
-  * [`tf.ones_like(tensor, dtype=None, name=None)`](#ones_like)
-  * [`tf.fill(dims, value, name=None)`](#fill)
-  * [`tf.constant(value, dtype=None, shape=None, name='Const')`](#constant)
-* [Sequences](#AUTOGENERATED-sequences)
-  * [`tf.linspace(start, stop, num, name=None)`](#linspace)
-  * [`tf.range(start, limit=None, delta=1, name='range')`](#range)
-* [Random Tensors](#AUTOGENERATED-random-tensors)
-  * [Examples:](#AUTOGENERATED-examples-)
-  * [`tf.random_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)`](#random_normal)
-  * [`tf.truncated_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)`](#truncated_normal)
-  * [`tf.random_uniform(shape, minval=0.0, maxval=1.0, dtype=tf.float32, seed=None, name=None)`](#random_uniform)
-  * [`tf.random_shuffle(value, seed=None, name=None)`](#random_shuffle)
-  * [`tf.set_random_seed(seed)`](#set_random_seed)
-
-
-<!-- TOC-END This section was generated by neural network, THANKS FOR READING! -->
-
-## Constant Value Tensors <a class="md-anchor" id="AUTOGENERATED-constant-value-tensors"></a>
+[TOC]
+
+## Constant Value Tensors
 
 TensorFlow provides several operations that you can use to generate constants.
 
 - - -
 
-### `tf.zeros(shape, dtype=tf.float32, name=None)` <a class="md-anchor" id="zeros"></a>
+### `tf.zeros(shape, dtype=tf.float32, name=None)` {#zeros}
 
 Creates a tensor with all elements set to zero.
 
@@ -48,21 +26,21 @@ For example:
 tf.zeros([3, 4], int32) ==> [[0, 0, 0, 0], [0, 0, 0, 0], [0, 0, 0, 0]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`shape`</b>: Either a list of integers, or a 1-D `Tensor` of type `int32`.
 *  <b>`dtype`</b>: The type of an element in the resulting `Tensor`.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor` with all elements set to zero.
 
 
 - - -
 
-### `tf.zeros_like(tensor, dtype=None, name=None)` <a class="md-anchor" id="zeros_like"></a>
+### `tf.zeros_like(tensor, dtype=None, name=None)` {#zeros_like}
 
 Creates a tensor with all elements set to zero.
 
@@ -77,7 +55,7 @@ For example:
 tf.zeros_like(tensor) ==> [[0, 0, 0], [0, 0, 0]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`tensor`</b>: A `Tensor`.
@@ -86,7 +64,7 @@ tf.zeros_like(tensor) ==> [[0, 0, 0], [0, 0, 0]]
 
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor` with all elements set to zero.
 
@@ -94,7 +72,7 @@ tf.zeros_like(tensor) ==> [[0, 0, 0], [0, 0, 0]]
 
 - - -
 
-### `tf.ones(shape, dtype=tf.float32, name=None)` <a class="md-anchor" id="ones"></a>
+### `tf.ones(shape, dtype=tf.float32, name=None)` {#ones}
 
 Creates a tensor with all elements set to 1.
 
@@ -107,21 +85,21 @@ For example:
 tf.ones([2, 3], int32) ==> [[1, 1, 1], [1, 1, 1]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`shape`</b>: Either a list of integers, or a 1-D `Tensor` of type `int32`.
 *  <b>`dtype`</b>: The type of an element in the resulting `Tensor`.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor` with all elements set to 1.
 
 
 - - -
 
-### `tf.ones_like(tensor, dtype=None, name=None)` <a class="md-anchor" id="ones_like"></a>
+### `tf.ones_like(tensor, dtype=None, name=None)` {#ones_like}
 
 Creates a tensor with all elements set to 1.
 
@@ -136,7 +114,7 @@ For example:
 tf.ones_like(tensor) ==> [[1, 1, 1], [1, 1, 1]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`tensor`</b>: A `Tensor`.
@@ -145,7 +123,7 @@ tf.ones_like(tensor) ==> [[1, 1, 1], [1, 1, 1]]
 
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor` with all elements set to 1.
 
@@ -153,7 +131,7 @@ tf.ones_like(tensor) ==> [[1, 1, 1], [1, 1, 1]]
 
 - - -
 
-### `tf.fill(dims, value, name=None)` <a class="md-anchor" id="fill"></a>
+### `tf.fill(dims, value, name=None)` {#fill}
 
 Creates a tensor filled with a scalar value.
 
@@ -168,7 +146,7 @@ fill(dims, 9) ==> [[9, 9, 9]
                    [9, 9, 9]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`dims`</b>: A `Tensor` of type `int32`.
@@ -176,7 +154,7 @@ fill(dims, 9) ==> [[9, 9, 9]
 *  <b>`value`</b>: A `Tensor`. 0-D (scalar). Value to fill the returned tensor.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor`. Has the same type as `value`.
 
@@ -184,7 +162,7 @@ fill(dims, 9) ==> [[9, 9, 9]
 
 - - -
 
-### `tf.constant(value, dtype=None, shape=None, name='Const')` <a class="md-anchor" id="constant"></a>
+### `tf.constant(value, dtype=None, shape=None, name='Const')` {#constant}
 
 Creates a constant tensor.
 
@@ -217,7 +195,7 @@ Creates a constant tensor.
                                               [-1. -1. -1.]]
  ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`value`</b>: A constant value (or list) of output type `dtype`.
@@ -231,17 +209,17 @@ Creates a constant tensor.
 
 *  <b>`name`</b>: Optional name for the tensor.
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A Constant Tensor.
 
 
 
-## Sequences <a class="md-anchor" id="AUTOGENERATED-sequences"></a>
+## Sequences
 
 - - -
 
-### `tf.linspace(start, stop, num, name=None)` <a class="md-anchor" id="linspace"></a>
+### `tf.linspace(start, stop, num, name=None)` {#linspace}
 
 Generates values in an interval.
 
@@ -255,7 +233,7 @@ For example:
 tf.linspace(10.0, 12.0, 3, name="linspace") => [ 10.0  11.0  12.0]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`start`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
@@ -265,7 +243,7 @@ tf.linspace(10.0, 12.0, 3, name="linspace") => [ 10.0  11.0  12.0]
 *  <b>`num`</b>: A `Tensor` of type `int32`. Number of values to generate.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A `Tensor`. Has the same type as `start`. 1-D. The generated values.
 
@@ -273,7 +251,7 @@ tf.linspace(10.0, 12.0, 3, name="linspace") => [ 10.0  11.0  12.0]
 
 - - -
 
-### `tf.range(start, limit=None, delta=1, name='range')` <a class="md-anchor" id="range"></a>
+### `tf.range(start, limit=None, delta=1, name='range')` {#range}
 
 Creates a sequence of integers.
 
@@ -295,7 +273,7 @@ tf.range(start, limit, delta) ==> [3, 6, 9, 12, 15]
 tf.range(limit) ==> [0, 1, 2, 3, 4]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`start`</b>: A 0-D (scalar) of type `int32`. First entry in sequence.
@@ -306,13 +284,13 @@ tf.range(limit) ==> [0, 1, 2, 3, 4]
     Number that increments `start`.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   An 1-D `int32` `Tensor`.
 
 
 
-## Random Tensors <a class="md-anchor" id="AUTOGENERATED-random-tensors"></a>
+## Random Tensors
 
 TensorFlow has several ops that create random tensors with different
 distributions.  The random ops are stateful, and create new random values each
@@ -328,7 +306,7 @@ See [`set_random_seed`](../../api_docs/python/constant_op.md#set_random_seed)
 for details on the interaction between operation-level and graph-level random
 seeds.
 
-### Examples: <a class="md-anchor" id="AUTOGENERATED-examples-"></a>
+### Examples:
 
 ```python
 # Create a tensor of shape [2, 3] consisting of random normal values, with mean
@@ -368,11 +346,11 @@ print sess.run(var)
 
 - - -
 
-### `tf.random_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)` <a class="md-anchor" id="random_normal"></a>
+### `tf.random_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)` {#random_normal}
 
 Outputs random values from a normal distribution.
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`shape`</b>: A 1-D integer Tensor or Python array. The shape of the output tensor.
@@ -387,14 +365,14 @@ Outputs random values from a normal distribution.
     for behavior.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A tensor of the specified shape filled with random normal values.
 
 
 - - -
 
-### `tf.truncated_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)` <a class="md-anchor" id="truncated_normal"></a>
+### `tf.truncated_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)` {#truncated_normal}
 
 Outputs random values from a truncated normal distribution.
 
@@ -402,7 +380,7 @@ The generated values follow a normal distribution with specified mean and
 standard deviation, except that values whose magnitude is more than 2 standard
 deviations from the mean are dropped and re-picked.
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`shape`</b>: A 1-D integer Tensor or Python array. The shape of the output tensor.
@@ -417,14 +395,14 @@ deviations from the mean are dropped and re-picked.
     for behavior.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A tensor of the specified shape filled with random truncated normal values.
 
 
 - - -
 
-### `tf.random_uniform(shape, minval=0.0, maxval=1.0, dtype=tf.float32, seed=None, name=None)` <a class="md-anchor" id="random_uniform"></a>
+### `tf.random_uniform(shape, minval=0.0, maxval=1.0, dtype=tf.float32, seed=None, name=None)` {#random_uniform}
 
 Outputs random values from a uniform distribution.
 
@@ -432,7 +410,7 @@ The generated values follow a uniform distribution in the range
 `[minval, maxval)`. The lower bound `minval` is included in the range, while
 the upper bound `maxval` is excluded.
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`shape`</b>: A 1-D integer Tensor or Python array. The shape of the output tensor.
@@ -447,14 +425,14 @@ the upper bound `maxval` is excluded.
     for behavior.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A tensor of the specified shape filled with random uniform values.
 
 
 - - -
 
-### `tf.random_shuffle(value, seed=None, name=None)` <a class="md-anchor" id="random_shuffle"></a>
+### `tf.random_shuffle(value, seed=None, name=None)` {#random_shuffle}
 
 Randomly shuffles a tensor along its first dimension.
 
@@ -468,7 +446,7 @@ to one and only one `output[i]`. For example, a mapping that might occur for a
  [5, 6]]        [3, 4]]
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`value`</b>: A Tensor to be shuffled.
@@ -478,7 +456,7 @@ to one and only one `output[i]`. For example, a mapping that might occur for a
     for behavior.
 *  <b>`name`</b>: A name for the operation (optional).
 
-##### Returns: <a class="md-anchor" id="AUTOGENERATED-returns-"></a>
+##### Returns:
 
   A tensor of same shape and type as `value`, shuffled along its first
   dimension.
@@ -486,7 +464,7 @@ to one and only one `output[i]`. For example, a mapping that might occur for a
 
 - - -
 
-### `tf.set_random_seed(seed)` <a class="md-anchor" id="set_random_seed"></a>
+### `tf.set_random_seed(seed)` {#set_random_seed}
 
 Sets the graph-level random seed.
 
@@ -579,7 +557,7 @@ with tf.Session() as sess2:
   print sess2.run(b)  # generates 'B2'
 ```
 
-##### Args: <a class="md-anchor" id="AUTOGENERATED-args-"></a>
+##### Args:
 
 
 *  <b>`seed`</b>: integer.