Merge branch 'develop' into infer_mkl

.travis.yml

@@ -56,7 +56,7 @@ script:
     export DEPLOY_DOCS_SH=https://raw.githubusercontent.com/PaddlePaddle/PaddlePaddle.org/master/scripts/deploy/deploy_docs.sh
     export DOCS_DIR=`pwd`
     cd ..
-    curl $DEPLOY_DOCS_SH | bash -s $CONTENT_DEC_PASSWD $TRAVIS_BRANCH $DOCS_DIR $DOCS_DIR/build/doc/v2   
+    curl $DEPLOY_DOCS_SH | bash -s $CONTENT_DEC_PASSWD $TRAVIS_BRANCH $DOCS_DIR $DOCS_DIR/build/doc/
 notifications:
   email:
     on_success: change

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/cpp_data_feeding.md

@@ -20,9 +20,8 @@ class ReaderBase {
     PADDLE_ENFORCE(!shapes_.empty());
   }
   // Read the next batch of data. (A 'batch' can be only one instance)
+  // If the next batch doesn't exist, the '*out' will be an empty std::vector.
   virtual void ReadNext(std::vector<LoDTensor>* out) = 0;
-  // Show whether the next bacth exists.
-  virtual bool HasNext() const = 0;
   
   // Reinitialize the reader and read the file from the begin.
   virtual void ReInit() = 0;

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)