Shixiaowei02 · Shixiaowei02 · Mar 27, 2019 · Mar 14, 2019 · Mar 14, 2019 · Mar 14, 2019
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -24,6 +24,8 @@ message(STATUS "CXX compiler: ${CMAKE_CXX_COMPILER}, version: "
         "${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}")
 message(STATUS "C compiler: ${CMAKE_C_COMPILER}, version: "
         "${CMAKE_C_COMPILER_ID} ${CMAKE_C_COMPILER_VERSION}")
+message(STATUS "AR tools: ${CMAKE_AR}")
+
 if(WIN32)
     set(CMAKE_SUPPRESS_REGENERATION ON)
     set(CMAKE_STATIC_LIBRARY_PREFIX lib)
@@ -62,6 +64,7 @@ option(WITH_DISTRIBUTE  "Compile with distributed support"              OFF)
 option(WITH_PSLIB       "Compile with pslib support"                    OFF)
 option(WITH_CONTRIB     "Compile the third-party contributation"        OFF)
 option(REPLACE_ENFORCE_GLOG "Replace PADDLE_ENFORCE with glog/CHECK for better debug." OFF)
+# TODO(Superjomn) Remove WITH_ANAKIN option if not needed latter.
 option(WITH_ANAKIN      "Compile with Anakin library"                   OFF)
 option(ANAKIN_BUILD_FAT_BIN "Build anakin cuda fat-bin lib for all device plantform, ignored when WITH_ANAKIN=OFF" OFF)
 option(ANAKIN_BUILD_CROSS_PLANTFORM "Build anakin lib for any nvidia device plantform. ignored when WITH_ANAKIN=OFF" ON)
@@ -188,6 +191,7 @@ include(configure)          # add paddle env configuration
 if(WITH_GPU)
     include(cuda)
     include(tensorrt)
+    include(anakin_subgraph)
 endif()
 if(WITH_MKL OR WITH_MKLML)
     include(external/anakin)

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -156,7 +156,7 @@ python \
 
 This will enable VLOG messages generated by `buddy_allocator.{h,cc}` and in the verbose range of 0 to 3, so you will see above example VLOG message, which is in level 3.  This suggests that we output overall messages in lower verbose levels, so they display with higher probability.  When coding C++, please follow the verbose level convention as follows:
 
-- verbose level 1: [framework](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework)
-- verbose level 3: [operators](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators)
-- verbose level 5: [memory](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/memory), [platform](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/platform)
-- verbose level 7: [math](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/legacy/math)
+- verbose level 1: [framework](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/fluid/framework)
+- verbose level 3: [operators](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/fluid/operators)
+- verbose level 5: [memory](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/fluid/memory), [platform](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/fluid/platform)
+- verbose level 7: [math](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/fluid/operators/math/)
diff --git a/Dockerfile b/Dockerfile
@@ -75,8 +75,9 @@ RUN curl -s -q https://glide.sh/get | sh
 #    and its size is only one-third of the official one.
 # 2. Manually add ~IPluginFactory() in IPluginFactory class of NvInfer.h, otherwise, it couldn't work in paddle.
 #    See https://github.com/PaddlePaddle/Paddle/issues/10129 for details.
-RUN wget -qO- http://paddlepaddledeps.cdn.bcebos.com/TensorRT-4.0.0.3.Ubuntu-16.04.4.x86_64-gnu.cuda-8.0.cudnn7.0.tar.gz | \
-    tar -xz -C /usr/local && \
+
+RUN wget -q https://paddlepaddledeps.cdn.bcebos.com/TensorRT-4.0.1.6-ubuntu14.04.x86_64-gnu.cuda.8.0.cudnn7.0.tar.gz --no-check-certificate && \
+    tar -zxf TensorRT-4.0.1.6-ubuntu14.04.x86_64-gnu.cuda.8.0.cudnn7.0.tar.gz -C /usr/local && \
     cp -rf /usr/local/TensorRT/include /usr && \
     cp -rf /usr/local/TensorRT/lib /usr
 

diff --git a/benchmark/fluid/fluid_benchmark.py b/benchmark/fluid/fluid_benchmark.py
@@ -179,7 +179,6 @@ def train_parallel(train_args, test_args, args, train_prog, test_prog,
     else:
         build_strategy.reduce_strategy = fluid.BuildStrategy(
         ).ReduceStrategy.AllReduce
-    build_strategy.fuse_broadcast_op = args.fuse_broadcast_op
 
     avg_loss = train_args[0]
 

diff --git a/cmake/anakin_subgraph.cmake b/cmake/anakin_subgraph.cmake
@@ -0,0 +1,32 @@
+if(NOT WITH_GPU)
+    return()
+endif()
+
+set(ANAKIN_ROOT "/usr" CACHE PATH "ANAKIN ROOT")
+find_path(ANAKIN_INCLUDE_DIR anakin_config.h
+    PATHS ${ANAKIN_ROOT} ${ANAKIN_ROOT}/include
+    $ENV{ANAKIN_ROOT} $ENV{ANAKIN_ROOT}/include
+    NO_DEFAULT_PATH
+)
+
+find_library(ANAKIN_LIBRARY NAMES libanakin_saber_common.so libanakin.so
+    PATHS ${ANAKIN_ROOT}
+    $ENV{ANAKIN_ROOT} $ENV{ANAKIN_ROOT}/lib
+    NO_DEFAULT_PATH
+    DOC "Path to ANAKIN library.")
+
+if(ANAKIN_INCLUDE_DIR AND ANAKIN_LIBRARY)
+  if(WITH_DSO)
+    set(ANAKIN_FOUND ON)
+  endif(WITH_DSO)
+else()
+    set(ANAKIN_FOUND OFF)
+endif()
+
+if(ANAKIN_FOUND)
+    message(STATUS "Current ANAKIN header is ${ANAKIN_INCLUDE_DIR}/anakin_config.h. ")
+    include_directories(${ANAKIN_ROOT}/include)
+    include_directories(${ANAKIN_ROOT}/include/saber)
+    link_directories(${ANAKIN_ROOT})
+    add_definitions(-DPADDLE_WITH_ANAKIN)
+endif()
diff --git a/cmake/external/boost.cmake b/cmake/external/boost.cmake
@@ -24,7 +24,7 @@ set(BOOST_PROJECT       "extern_boost")
 # So we use 1.41.0 here.
 set(BOOST_VER           "1.41.0")
 set(BOOST_TAR "boost_1_41_0" CACHE STRING "" FORCE)
-set(BOOST_URL "http://paddlepaddledeps.cdn.bcebos.com/${BOOST_TAR}.tar.gz" CACHE STRING "" FORCE)
+set(BOOST_URL "http://paddlepaddledeps.bj.bcebos.com/${BOOST_TAR}.tar.gz" CACHE STRING "" FORCE)
 
 MESSAGE(STATUS "BOOST_TAR: ${BOOST_TAR}, BOOST_URL: ${BOOST_URL}")
 

diff --git a/cmake/external/grpc.cmake b/cmake/external/grpc.cmake
@@ -44,7 +44,7 @@ ExternalProject_Add(
     # 3. keep only zlib, cares, protobuf, boringssl under "third_party",
     #    checkout and clean other dirs under third_party
     # 4. remove .git, and package the directory.
-    URL "http://paddlepaddledeps.cdn.bcebos.com/grpc-v1.10.x.tar.gz"
+    URL "http://paddlepaddledeps.bj.bcebos.com/grpc-v1.10.x.tar.gz"
     URL_MD5  "1f268a2aff6759839dccd256adcc91cf"
     PREFIX          ${GRPC_SOURCES_DIR}
     UPDATE_COMMAND  ""

diff --git a/cmake/external/mkldnn.cmake b/cmake/external/mkldnn.cmake
@@ -31,9 +31,17 @@ IF(APPLE)
     return()
 ENDIF()
 
-MESSAGE(STATUS "Set ${MKLDNN_INSTALL_DIR}/lib to runtime path")
+# Introduce variables:
+# * CMAKE_INSTALL_LIBDIR
+INCLUDE(GNUInstallDirs)
+SET(LIBDIR "lib")
+if(CMAKE_INSTALL_LIBDIR MATCHES ".*lib64$")
+  SET(LIBDIR "lib64")
+endif()
+
+MESSAGE(STATUS "Set ${MKLDNN_INSTALL_DIR}/l${LIBDIR} to runtime path")
 SET(CMAKE_INSTALL_RPATH_USE_LINK_PATH TRUE)
-SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${MKLDNN_INSTALL_DIR}/lib")
+SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${MKLDNN_INSTALL_DIR}/${LIBDIR}")
 
 INCLUDE_DIRECTORIES(${MKLDNN_INC_DIR}) # For MKLDNN code to include internal headers.
 
@@ -58,7 +66,7 @@ ExternalProject_Add(
     ${EXTERNAL_PROJECT_LOG_ARGS}
     DEPENDS             ${MKLDNN_DEPENDS}
     GIT_REPOSITORY      "https://github.com/intel/mkl-dnn.git"
-    GIT_TAG             "830a10059a018cd2634d94195140cf2d8790a75a"
+    GIT_TAG             "863ff6e7042cec7d2e29897fe9f0872e0888b0fc"
     PREFIX              ${MKLDNN_SOURCES_DIR}
     UPDATE_COMMAND      ""
     CMAKE_ARGS          -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
@@ -79,9 +87,9 @@ ExternalProject_Add(
                         -DMKLROOT:PATH=${MKLML_ROOT}
 )
 if(WIN32)
-    SET(MKLDNN_LIB "${MKLDNN_INSTALL_DIR}/lib/mkldnn.lib" CACHE FILEPATH "mkldnn library." FORCE)
+    SET(MKLDNN_LIB "${MKLDNN_INSTALL_DIR}/${LIBDIR}/mkldnn.lib" CACHE FILEPATH "mkldnn library." FORCE)
 else(WIN32)
-    SET(MKLDNN_LIB "${MKLDNN_INSTALL_DIR}/lib/libmkldnn.so" CACHE FILEPATH "mkldnn library." FORCE)
+    SET(MKLDNN_LIB "${MKLDNN_INSTALL_DIR}/${LIBDIR}/libmkldnn.so" CACHE FILEPATH "mkldnn library." FORCE)
 endif(WIN32)
 
 ADD_LIBRARY(shared_mkldnn SHARED IMPORTED GLOBAL)
@@ -101,7 +109,7 @@ ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
 # copy the real so.0 lib to install dir
 # it can be directly contained in wheel or capi
 if(WIN32)
-    SET(MKLDNN_SHARED_LIB ${MKLDNN_INSTALL_DIR}/lib/mkldnn.dll)
+    SET(MKLDNN_SHARED_LIB ${MKLDNN_INSTALL_DIR}/bin/mkldnn.dll)
 else(WIN32)
     SET(MKLDNN_SHARED_LIB ${MKLDNN_INSTALL_DIR}/libmkldnn.so.0)
     ADD_CUSTOM_COMMAND(OUTPUT ${MKLDNN_SHARED_LIB}

diff --git a/cmake/external/mklml.cmake b/cmake/external/mklml.cmake
@@ -34,7 +34,7 @@ SET(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_RPATH}" "${MKLML_ROOT}/lib")
 SET(TIME_VERSION "2019.0.1.20181227")
 IF(WIN32)
     SET(MKLML_VER "mklml_win_${TIME_VERSION}" CACHE STRING "" FORCE)
-    SET(MKLML_URL "https://paddlepaddledeps.cdn.bcebos.com/${MKLML_VER}.zip" CACHE STRING "" FORCE)
+    SET(MKLML_URL "https://paddlepaddledeps.bj.bcebos.com/${MKLML_VER}.zip" CACHE STRING "" FORCE)
     SET(MKLML_LIB                 ${MKLML_LIB_DIR}/mklml.lib)
     SET(MKLML_IOMP_LIB            ${MKLML_LIB_DIR}/libiomp5md.lib)
     SET(MKLML_SHARED_LIB          ${MKLML_LIB_DIR}/mklml.dll)
@@ -43,7 +43,7 @@ ELSE()
     #TODO(intel-huying):
     #  Now enable Erf function in mklml library temporarily, it will be updated as offical version later.
     SET(MKLML_VER "Glibc225_vsErf_mklml_lnx_${TIME_VERSION}" CACHE STRING "" FORCE)
-    SET(MKLML_URL "http://paddlepaddledeps.cdn.bcebos.com/${MKLML_VER}.tgz" CACHE STRING "" FORCE)
+    SET(MKLML_URL "http://paddlepaddledeps.bj.bcebos.com/${MKLML_VER}.tgz" CACHE STRING "" FORCE)
     SET(MKLML_LIB                 ${MKLML_LIB_DIR}/libmklml_intel.so)
     SET(MKLML_IOMP_LIB            ${MKLML_LIB_DIR}/libiomp5.so)
     SET(MKLML_SHARED_LIB          ${MKLML_LIB_DIR}/libmklml_intel.so)

diff --git a/cmake/operators.cmake b/cmake/operators.cmake
@@ -110,7 +110,7 @@ function(op_library TARGET)
     # Define operators that don't need pybind here.
     foreach(manual_pybind_op "compare_op" "logical_op" "nccl_op"
 "tensor_array_read_write_op" "tensorrt_engine_op" "conv_fusion_op"
-"fusion_transpose_flatten_concat_op" "fusion_conv_inception_op")
+"fusion_transpose_flatten_concat_op" "fusion_conv_inception_op" "sync_batch_norm_op")
         if ("${TARGET}" STREQUAL "${manual_pybind_op}")
             set(pybind_flag 1)
         endif()

diff --git a/cmake/tensorrt.cmake b/cmake/tensorrt.cmake
@@ -33,5 +33,6 @@ if(TENSORRT_FOUND)
     message(STATUS "Current TensorRT header is ${TENSORRT_INCLUDE_DIR}/NvInfer.h. "
         "Current TensorRT version is v${TENSORRT_MAJOR_VERSION}. ")
     include_directories(${TENSORRT_INCLUDE_DIR})
+    link_directories(${TENSORRT_LIBRARY})
     add_definitions(-DPADDLE_WITH_TENSORRT)
 endif()