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

cmake/generic.cmake

@@ -244,11 +244,11 @@ function(cc_test TARGET_NAME)
     cmake_parse_arguments(cc_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
     add_executable(${TARGET_NAME} ${cc_test_SRCS})
     # Support linking flags: --whole-archive (Linux) / -force_load (MacOS)
-    target_circle_link_libraries(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
+    target_circle_link_libraries(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags glog)
     if("${cc_test_DEPS}" MATCHES "ARCHIVE_START")
       list(REMOVE_ITEM cc_test_DEPS ARCHIVE_START ARCHIVE_END)
     endif()
-    add_dependencies(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
+    add_dependencies(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags glog)
     add_test(NAME ${TARGET_NAME}
              COMMAND ${TARGET_NAME} ${cc_test_ARGS}
              WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
@@ -311,8 +311,8 @@ function(nv_test TARGET_NAME)
     set(multiValueArgs SRCS DEPS)
     cmake_parse_arguments(nv_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
     cuda_add_executable(${TARGET_NAME} ${nv_test_SRCS})
-    target_link_libraries(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
-    add_dependencies(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
+    target_link_libraries(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags glog)
+    add_dependencies(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags glog)
     add_test(${TARGET_NAME} ${TARGET_NAME})
   endif()
 endfunction(nv_test)

doc/design/fluid_dist/parameter_server.md

@@ -59,6 +59,17 @@ After converting:
      queue. It will block until the queue has the required number of
      tensors.
 
+### Sparse Update
+
+For embedding layers, the gradient may have many rows containing only 0 when training,
+if the gradient uses a dense tensor to do parameter optimization,
+it could spend unnecessary memory, slow down the calculations and waste
+the bandwidth while doing distributed training.
+In Fluid, we introduce [SelectedRows](../selected_rows.md) to represent a list of rows containing
+non-zero gradient data. So when we do parameter optimization both locally and remotely,
+we only need to send those non-zero rows to the optimizer operators:
+
+<img src="src/sparse_update.png" width="700" />
 
 ### Benefits
 
@@ -91,6 +102,6 @@ After converting:
   `min_count` attribute), does our current design support it? (similar
   question for the *Add* OP)
 
+### References
 
-### References:
 [1] [TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems](https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45166.pdf)

doc/v2/dev/index_cn.rst

@@ -1,9 +1,24 @@
 开发标准
 ========
+PaddlePaddle遵守如下三个部分的代码和文档规范。
+
+PaddlePaddle使用git做版本管理，docker作为构建和测试环境。代码中包含了Cuda, C++, Python, Shell等多种编程语言。语言规范遵守Google C++ Style, Pep-8, 代码库中包含自动化检查工具做风格检查。代码注释需要遵守Doxygen规范，不满足风格要求的代码会编译失败。关于如何使用git, 构建测试及代码开发, 我们提供了如下指南。
 
 ..  toctree::
   :maxdepth: 1
 
   contribute_to_paddle_cn.md
+
+PaddlePaddle面向国内外用户，包含了中文和英文两部分的文档。设计文档和issue问题描述都推荐使用英文。对于设计文档，重在问题描述，背景阐述，然后才是解决方案。文档由Sphinx生成，因此代码注释也需要符合Sphinx文档标准。推荐本地使用paddlepaddle.org工具编译生成和预览文档，请参阅如下文档。
+
+..  toctree::
+  :maxdepth: 1
+
   write_docs_cn.rst
+
+PaddlePaddle V2 使用新增Layer方式定义新的操作。组合基础API可以实现多种复杂Layer, 满足绝大多数应用。如需要定制Layer，请参阅如下文档，欢迎提交patch。
+
+..  toctree::
+  :maxdepth: 1
+
   new_layer_cn.rst

doc/v2/howto/cluster/cmd_argument_cn.md

@@ -71,6 +71,13 @@ paddle.init(
 - trainer_id：**必选，默认0**，每个trainer的唯一ID，从0开始的整数
 - pservers：**必选，默认127.0.0.1**，当前训练任务启动的pserver的IP列表，多个IP使用“,”隔开
 
+```python
+trainer = paddle.trainer.SGD(..., is_local=False)
+```
+
+参数说明
+
+- is_local: **必选, 默认True**, 是否使用PServer更新参数
 
 ## 准备数据集