Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/clean_op_maker

cmake/external/mkldnn.cmake

@@ -53,7 +53,7 @@ ExternalProject_Add(
     ${EXTERNAL_PROJECT_LOG_ARGS}
     DEPENDS             ${MKLDNN_DEPENDS}
     GIT_REPOSITORY      "https://github.com/01org/mkl-dnn.git"
-    GIT_TAG             "v0.11"
+    GIT_TAG             "v0.14"
     PREFIX              ${MKLDNN_SOURCES_DIR}
     UPDATE_COMMAND      ""
     CMAKE_ARGS          -DCMAKE_INSTALL_PREFIX=${MKLDNN_INSTALL_DIR}

doc/fluid/build_and_install/paddleci.png

@@ -0,0 +1 @@
+../../v2/build_and_install/paddleci.png

doc/fluid/design/motivation/refactorization.md

@@ -125,12 +125,12 @@ Compile Time -> IR -> Runtime
 
 ## Operator/OpWithKernel/OpKernel
 
-![class_diagram](http://api.paddlepaddle.org/graphviz?dot=https://gist.githubusercontent.com/reyoung/53df507f6749762675dff3e7ce53372f/raw/49caf1fb70820fb4a6c217634317c9306f361f36/op_op_with_kern_class_diagram.dot)
+![class_diagram](https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/fluid/images/op_op_with_kern_class_diagram.dot)
 
 ---
 
 ## Operator
-![class_diagram](http://api.paddlepaddle.org/graphviz?dot=https://gist.githubusercontent.com/reyoung/53df507f6749762675dff3e7ce53372f/raw/dd598e8f1976f5759f58af5e5ef94738a6b2e661/op.dot)
+![class_diagram](https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/fluid/images/op.dot)
 
 * `Operator` is the fundamental building block of the user interface.
     * Operator stores input/output variable names and attributes.
@@ -141,7 +141,7 @@ Compile Time -> IR -> Runtime
 
 ## OpWithKernel/Kernel
 
-![class_diagram](http://api.paddlepaddle.org/graphviz?dot=https://gist.githubusercontent.com/reyoung/53df507f6749762675dff3e7ce53372f/raw/9d7f4eba185cf41c8e2fbfb40ae21890dbddcd39/op_with_kernel.dot)
+![class_diagram](https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/fluid/images/op_with_kernel.dot)
 
 * `OpWithKernel` inherits `Operator`.
 * `OpWithKernel` contains a Kernel map.

doc/fluid/images/op.dot

@@ -0,0 +1,4 @@
+digraph sample { 
+  graph [rankdir=TD]; node [shape=record];
+  op [label="{Operator| InferShape()=0\lRun()=0\l | map<string, string[]> inputs_\lmap<string, string[]> outputs_ \l AttributeMap attrs_\l}"]; 
+}

doc/fluid/images/op_op_with_kern_class_diagram.dot

@@ -0,0 +1,38 @@
+digraph sample { 
+  graph [rankdir=TD]; node [shape=record];
+  op [label="{Operator| InferShape()=0\lRun()=0\l | map<string, string[]> inputs_\lmap<string, string[]> outputs_ \l AttributeMap attrs_\l}"]; 
+  op_with_kern [label="{OpWithKernel | InferShape()=0\lRun()\l | map<OpKernelKey,OpKernel>kernels_ }"]
+  op_kernel [label="{OpKernel | Compute()=0}"]
+  op_kernel_key [label="{OpKernelKey| Place place\n...}"]
+
+  op -> op_with_kern [dir=back, arrowtail=onormal]
+  op_with_kern -> op_kernel [arrowhead=vee, label="contains many"]
+
+  {
+    rank=same;
+    op_with_kern
+    op_kernel
+  }
+
+  op_kernel -> op_kernel_key [style=invis]
+
+  {
+    rank=same;
+    op_kernel
+    op_kernel_key
+  }
+
+  op_with_kern -> op_kernel_key [arrowhead=vee, label ="\nas map key"]
+
+  mul_op [label="MulOp"]
+  op_with_kern -> mul_op [dir=back, arrowtail=onormal]
+  mul_kernel [label="template <typename Place>\lclass MulOpKernel\l"]
+  op_kernel -> mul_kernel [dir=back, arrowtail=onormal]
+  mul_op -> mul_kernel [arrowhead=vee, label="register many"]
+  
+  {
+    rank=same;
+    mul_op;
+    mul_kernel;
+  }
+}

doc/fluid/images/op_with_kernel.dot

@@ -0,0 +1,26 @@
+digraph sample { 
+  graph [rankdir=TD]; node [shape=record];
+  op [label="{Operator}"]; 
+  op_with_kern [label="{OpWithKernel | InferShape()=0\lRun()\l | map<OpKernelKey,OpKernel>kernels_ }"]
+  op_kernel [label="{OpKernel | Compute()=0}"]
+  op_kernel_key [label="{OpKernelKey| Place place\n...}"]
+
+  op -> op_with_kern [dir=back, arrowtail=onormal]
+  op_with_kern -> op_kernel [arrowhead=vee, label="contains many"]
+
+  {
+    rank=same;
+    op_with_kern
+    op_kernel
+  }
+
+  op_kernel -> op_kernel_key [style=invis]
+
+  {
+    rank=same;
+    op_kernel
+    op_kernel_key
+  }
+
+  op_with_kern -> op_kernel_key [arrowhead=vee, label ="\nas map key"]
+}

doc/v2/design/mkl/mkldnn.md

@@ -5,7 +5,7 @@
 充分展现英特尔平台的优势，有效提升PaddlePaddle在英特尔架构上的性能。
 
 <div align="center">
-<img src="image/overview.png"><br/>
+<img src="https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/v2/images/overview.png"><br/>
 Figure 1. PaddlePaddle on IA
 </div>
 
@@ -42,16 +42,43 @@ Figure 1. PaddlePaddle on IA
 
 MKL，MKLML以及MKL-DNN三者关系如下表：
 
-| Name        |  Open Source     | License     | Descriptions  |
-| :---------- | :--------------- | :---------- | :------------ |
-|   MKL       |     No           | Proprietary | Accelerate math processing routines |
-|   MKLML     |     No           | Proprietary | Small package of MKL, especially for Machine Learning |
-|   MKL-DNN   |     Yes          | Apache 2.0  | Accelerate primitives processing routines especially for Deep Neural Networks  |
+<table>
+<thead>
+<tr>
+<th>Name</th>
+<th>Open Source</th>
+<th>License</th>
+<th>Descriptions</th>
+</tr>
+</thead>
+<tbody>
+<tr>
+<td>MKL</td>
+<td>No</td>
+<td>Proprietary</td>
+<td>Accelerate math processing routines</td>
+</tr>
+<tr>
+<td>MKLML</td>
+<td>No</td>
+<td>Proprietary</td>
+<td>Small package of MKL, especially for Machine Learning</td>
+</tr>
+
+<tr>
+<td>MKL-DNN</td>
+<td>Yes</td>
+<td>Apache 2.0</td>
+<td>Accelerate primitives processing routines especially for Deep Neural Networks</td>
+</tr>
+
+</tbody>
+</table>
 
 MKLML可以与MKL-DNN共同使用，以此达到最好的性能。
 
 <div align="center">
-<img src="image/engine.png"><br/>
+<img src="https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/v2/images/engine.png"><br/>
 Figure 2. PaddlePaddle with MKL Engines
 </div>
 
@@ -103,7 +130,7 @@ MKL-DNN的库目前只有动态库`libmkldnn.so`。
 所以我们定义了一个`MKLDNNMatrix`用于管理MKL-DNN数据的不同格式以及相互之间的转换。
 
 <div align="center">
-<img src="image/matrix.png"><br/>
+<img src="https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/v2/images/matrix.png"><br/>
 Figure 3. MKLDNNMatrix
 </div>
 
@@ -113,7 +140,7 @@ Figure 3. MKLDNNMatrix
 子类只需要使用定义好的接口，实现具体的函数功能即可。
 
 <div align="center">
-<img src="image/layers.png"><br/>
+<img src="https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/v2/images/layers.png"><br/>
 Figure 4. MKLDNNLayer
 </div>
 
@@ -150,7 +177,7 @@ Figure 4. MKLDNNLayer
 所以整体上，在实现每个子类的时候就不需要关心分支的事情了。
 
 <div align="center">
-<img src="image/gradients.png"><br/>
+<img src="https://raw.githubusercontent.com/PaddlePaddle/Paddle/develop/doc/v2/images/gradients.png"><br/>
 Figure 5. Merge Gradients
 </div>
 

doc/v2/images/glossary_rnn.dot

@@ -0,0 +1,42 @@
+digraph G{
+	subgraph cluster_timestep0 {
+		label="recurrent timestep i-1"
+		bgcolor=lightgray
+		node [style=filled,color=white]
+		fc0_0 [label="fc 0"]
+		fc0_1 [label="fc 1"]
+		fc0_2 [label="fc 2"]
+
+		fc0_0 -> fc0_1
+		fc0_1 -> fc0_2
+	}
+
+	subgraph cluster_timestep1 {
+		label="recurrent timestep i"
+		node [style=filled];
+		fc1_0 [label="fc 0"]
+		fc1_1 [label="fc 1"]
+		fc1_2 [label="fc 2"]
+		color=blue
+
+		fc1_0 -> fc1_1
+		fc1_1 -> fc1_2
+	}
+
+	subgraph cluster_timestep2 {
+		label="recurrent timestep i+1"
+		bgcolor=lightgray
+		node [style=filled,color=white]
+		fc2_0 [label="fc 0"]
+		fc2_1 [label="fc 1"]
+		fc2_2 [label="fc 2"]
+
+		fc2_0 -> fc2_1
+		fc2_1 -> fc2_2
+	}
+	
+	
+	fc0_1 -> fc1_1 [style="dotted" constraint=false]
+	fc1_1 -> fc2_1 [style="dotted" constraint=false]
+
+}

doc/v2/images/glossary_rnn_with_memory.dot

@@ -0,0 +1,48 @@
+digraph G{
+	subgraph cluster_timestep0 {
+		label="recurrent timestep i-1"
+		bgcolor=lightgray
+		node [style=filled,color=white]
+		fc0_0 [label="fc 0"]
+		fc0_1 [label="fc 1"]
+		fc0_2 [label="fc 2"]
+		m0 [label="memory"]
+		fc0_0 -> fc0_1
+		fc0_1 -> fc0_2
+		fc0_1 -> m0
+		m0 -> fc0_1
+	}
+
+	subgraph cluster_timestep1 {
+		label="recurrent timestep i"
+		node [style=filled];
+		fc1_0 [label="fc 0"]
+		fc1_1 [label="fc 1"]
+		fc1_2 [label="fc 2"]
+		m1 [label="memory"]
+		color=blue
+		fc1_0 -> fc1_1
+		fc1_1 -> fc1_2
+		fc1_1 -> m1
+		m1 -> fc1_1
+	}
+
+	subgraph cluster_timestep2 {
+		label="recurrent timestep i+1"
+		bgcolor=lightgray
+		node [style=filled,color=white]
+		fc2_0 [label="fc 0"]
+		fc2_1 [label="fc 1"]
+		fc2_2 [label="fc 2"]
+		m2 [label="memory"]
+		fc2_0 -> fc2_1
+		fc2_1 -> fc2_2
+		fc2_1 -> m2
+		m2 -> fc2_1
+	}
+	
+	
+	m0 -> m1 [style="dotted" constraint=false]
+	m1 -> m2 [style="dotted" constraint=false]
+
+}