merge develop

test=develop

CMakeLists.txt

@@ -130,6 +130,21 @@ if (APPLE OR WIN32)
         "Disable MKL for building on mac and windows" FORCE)
 endif()
 
+if (WIN32)
+    set(WITH_AVX OFF CACHE STRING
+            "Disable AVX when compiling for Windows" FORCE)
+    set(WITH_DSO OFF CACHE STRING
+            "Disable DSO when compiling for Windows" FORCE)
+    set(WITH_MKL OFF CACHE STRING
+            "Disable MKL when compiling for Windows" FORCE)
+    set(WITH_DISTRIBUTE OFF CACHE STRING
+            "Disable DISTRIBUTE when compiling for Windows" FORCE)
+    set(WITH_C_API OFF CACHE STRING
+            "Disable C_API when compiling for Windows" FORCE)
+    set(WITH_FLUID_ONLY ON CACHE STRING
+            "Enable FLUID_ONLY when compiling for Windows" FORCE)
+endif()
+
 set(THIRD_PARTY_PATH "${CMAKE_BINARY_DIR}/third_party" CACHE STRING
   "A path setting third party libraries download & build directories.")
 
@@ -191,11 +206,11 @@ include(external/cares)
 include(external/cub)
 include(external/xxhash)    # download xxhash
 include(external/dlpack)
-
-if (NOT WIN32)
-# there is no official support of snappystream, warpctc, nccl, cupti in windows
 include(external/snappy)    # download snappy
 include(external/snappystream) # download snappystream
+
+if (NOT WIN32)
+# there is no official support of warpctc, nccl, cupti in windows
 include(external/warpctc)   # download, build, install warpctc
 include(cupti)
 endif (NOT WIN32)

cmake/external/gtest.cmake

@@ -50,7 +50,11 @@ IF(WITH_TESTING)
         CMAKE_ARGS      -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
                         -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
                         -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
+                        -DCMAKE_CXX_FLAGS_RELEASE=${CMAKE_CXX_FLAGS_RELEASE}
+                        -DCMAKE_CXX_FLAGS_DEBUG=${CMAKE_CXX_FLAGS_DEBUG}
                         -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
+                        -DCMAKE_C_FLAGS_DEBUG=${CMAKE_C_FLAGS_DEBUG}
+                        -DCMAKE_C_FLAGS_RELEASE=${CMAKE_C_FLAGS_RELEASE}
                         -DCMAKE_INSTALL_PREFIX=${GTEST_INSTALL_DIR}
                         -DCMAKE_POSITION_INDEPENDENT_CODE=ON
                         -DBUILD_GMOCK=ON

cmake/external/snappy.cmake

@@ -24,7 +24,11 @@ set(SNAPPY_SOURCES_DIR ${THIRD_PARTY_PATH}/snappy)
 set(SNAPPY_INSTALL_DIR ${THIRD_PARTY_PATH}/install/snappy)
 set(SNAPPY_INCLUDE_DIR "${SNAPPY_INSTALL_DIR}/include" CACHE PATH "snappy include directory." FORCE)
 
-set(SNAPPY_LIBRARIES "${SNAPPY_INSTALL_DIR}/lib/libsnappy.a")
+if (WIN32)
+    set(SNAPPY_LIBRARIES "${SNAPPY_INSTALL_DIR}/lib/snappy.lib")
+else(WIN32)
+    set(SNAPPY_LIBRARIES "${SNAPPY_INSTALL_DIR}/lib/libsnappy.a")
+endif (WIN32)
 
 ExternalProject_Add(
     extern_snappy
@@ -34,8 +38,12 @@ ExternalProject_Add(
     UPDATE_COMMAND  ""
     CMAKE_ARGS      -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
                     -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
-                    -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
                     -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
+                    -DCMAKE_C_FLAGS_DEBUG=${CMAKE_C_FLAGS_DEBUG}
+                    -DCMAKE_C_FLAGS_RELEASE=${CMAKE_C_FLAGS_RELEASE}
+                    -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
+                    -DCMAKE_CXX_FLAGS_RELEASE=${CMAKE_CXX_FLAGS_RELEASE}
+                    -DCMAKE_CXX_FLAGS_DEBUG=${CMAKE_CXX_FLAGS_DEBUG}
                     -DCMAKE_INSTALL_PREFIX=${SNAPPY_INSTALL_DIR}
                     -DCMAKE_INSTALL_LIBDIR=${SNAPPY_INSTALL_DIR}/lib
                     -DCMAKE_POSITION_INDEPENDENT_CODE=ON

cmake/external/snappystream.cmake

@@ -18,36 +18,45 @@ ENDIF()
 
 include (ExternalProject)
 
-# NOTE: snappy is needed when linking with recordio
-
 set(SNAPPYSTREAM_SOURCES_DIR ${THIRD_PARTY_PATH}/snappy_stream)
 set(SNAPPYSTREAM_INSTALL_DIR ${THIRD_PARTY_PATH}/install/snappy_stream)
 set(SNAPPYSTREAM_INCLUDE_DIR "${SNAPPYSTREAM_INSTALL_DIR}/include" CACHE PATH "snappy stream include directory." FORCE)
 
-set(SNAPPYSTREAM_LIBRARIES "${SNAPPYSTREAM_INSTALL_DIR}/lib/libsnappystream.a")
-
-ExternalProject_Add(
-        extern_snappystream
-        GIT_REPOSITORY "https://github.com/hoxnox/snappystream.git"
-        GIT_TAG "0.2.8"
-        PREFIX          ${SNAPPYSTREAM_SOURCES_DIR}
-        UPDATE_COMMAND  ""
-        CMAKE_ARGS      -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
-                        -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
-                        -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
-                        -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
-                        -DCMAKE_INSTALL_PREFIX=${SNAPPY_INSTALL_DIR}
-                        -DCMAKE_INSTALL_LIBDIR=${SNAPPY_INSTALL_DIR}/lib
-                        -DCMAKE_POSITION_INDEPENDENT_CODE=ON
-                        -DCMAKE_BUILD_TYPE=${THIRD_PARTY_BUILD_TYPE}
-                        -DSNAPPY_ROOT=${SNAPPY_INSTALL_DIR}
-                        ${EXTERNAL_OPTIONAL_ARGS}
-                        CMAKE_CACHE_ARGS
-                        -DCMAKE_INSTALL_PREFIX:PATH=${SNAPPYSTREAM_INSTALL_DIR}
-                        -DCMAKE_INSTALL_LIBDIR:PATH=${SNAPPYSTREAM_INSTALL_DIR}/lib
-                        -DCMAKE_BUILD_TYPE:STRING=${THIRD_PARTY_BUILD_TYPE}
-        DEPENDS snappy
-)
+if(WIN32)
+    # Fix me, VS2015 come without VLA support
+    set(SNAPPYSTREAM_LIBRARIES "${SNAPPYSTREAM_INSTALL_DIR}/lib/snappystream.lib")
+    MESSAGE(WARNING, "In windows, snappystream has no compile support for windows,
+    please build it manually and put it at " ${SNAPPYSTREAM_INSTALL_DIR})
+else(WIN32)
+    set(SNAPPYSTREAM_LIBRARIES "${SNAPPYSTREAM_INSTALL_DIR}/lib/libsnappystream.a")
+
+    ExternalProject_Add(
+            extern_snappystream
+            GIT_REPOSITORY "https://github.com/hoxnox/snappystream.git"
+            GIT_TAG "0.2.8"
+            PREFIX          ${SNAPPYSTREAM_SOURCES_DIR}
+            UPDATE_COMMAND  ""
+            CMAKE_ARGS      -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
+                            -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
+                            -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
+                            -DCMAKE_C_FLAGS_DEBUG=${CMAKE_C_FLAGS_DEBUG}
+                            -DCMAKE_C_FLAGS_RELEASE=${CMAKE_C_FLAGS_RELEASE}
+                            -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
+                            -DCMAKE_CXX_FLAGS_RELEASE=${CMAKE_CXX_FLAGS_RELEASE}
+                            -DCMAKE_CXX_FLAGS_DEBUG=${CMAKE_CXX_FLAGS_DEBUG}
+                            -DCMAKE_INSTALL_PREFIX=${SNAPPY_INSTALL_DIR}
+                            -DCMAKE_INSTALL_LIBDIR=${SNAPPY_INSTALL_DIR}/lib
+                            -DCMAKE_POSITION_INDEPENDENT_CODE=ON
+                            -DCMAKE_BUILD_TYPE=${THIRD_PARTY_BUILD_TYPE}
+                            -DSNAPPY_ROOT=${SNAPPY_INSTALL_DIR}
+                            ${EXTERNAL_OPTIONAL_ARGS}
+                            CMAKE_CACHE_ARGS
+                            -DCMAKE_INSTALL_PREFIX:PATH=${SNAPPYSTREAM_INSTALL_DIR}
+                            -DCMAKE_INSTALL_LIBDIR:PATH=${SNAPPYSTREAM_INSTALL_DIR}/lib
+                            -DCMAKE_BUILD_TYPE:STRING=${THIRD_PARTY_BUILD_TYPE}
+            DEPENDS snappy
+    )
+endif(WIN32)
 
 add_library(snappystream STATIC IMPORTED GLOBAL)
 set_property(TARGET snappystream PROPERTY IMPORTED_LOCATION ${SNAPPYSTREAM_LIBRARIES})

cmake/generic.cmake

@@ -351,6 +351,9 @@ function(cc_test TARGET_NAME)
     cmake_parse_arguments(cc_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
     add_executable(${TARGET_NAME} ${cc_test_SRCS})
     target_link_libraries(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main lod_tensor memory gtest gflags glog)
+    if(WIN32)
+      target_link_libraries(${TARGET_NAME} shlwapi)
+    endif(WIN32)
     add_dependencies(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main lod_tensor memory gtest gflags glog)
     add_test(NAME ${TARGET_NAME}
              COMMAND ${TARGET_NAME} ${cc_test_ARGS}

cmake/operators.cmake

@@ -84,9 +84,7 @@ function(op_library TARGET)
     endif()
     if (WIN32)
     # remove windows unsupported op, because windows has no nccl, no warpctc such ops.
-    foreach(windows_unsupport_op "nccl_op" "gen_nccl_id_op" "warpctc_op" "hierarchical_sigmoid_op"
-     "crf_decoding_op" "select_op" "lstmp_op" "gru_op" "fusion_gru_op" "lstm_op" "fusion_lstm_op" "cumsum_op"
-      "fusion_seqconv_eltadd_relu_op" "channel_send_op" "channel_create_op" "channel_close_op" "channel_recv_op")
+    foreach(windows_unsupport_op "nccl_op" "gen_nccl_id_op" "warpctc_op")
         if ("${TARGET}" STREQUAL "${windows_unsupport_op}")
           return()
         endif()

cmake/simd.cmake

@@ -57,43 +57,46 @@ int main()
     return 0;
 }" SSE3_FOUND)
 
-# Check AVX
-set(CMAKE_REQUIRED_FLAGS ${AVX_FLAG})
-set(AVX_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
-CHECK_CXX_SOURCE_RUNS("
-#include <immintrin.h>
-int main()
-{
-    __m256 a = _mm256_set_ps (-1.0f, 2.0f, -3.0f, 4.0f, -1.0f, 2.0f, -3.0f, 4.0f);
-    __m256 b = _mm256_set_ps (1.0f, 2.0f, 3.0f, 4.0f, 1.0f, 2.0f, 3.0f, 4.0f);
-    __m256 result = _mm256_add_ps (a, b);
-    return 0;
-}" AVX_FOUND)
+# disable AVX by default on windows
+if(NOT WIN32)
+    # Check AVX
+    set(CMAKE_REQUIRED_FLAGS ${AVX_FLAG})
+    set(AVX_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
+    CHECK_CXX_SOURCE_RUNS("
+    #include <immintrin.h>
+    int main()
+    {
+        __m256 a = _mm256_set_ps (-1.0f, 2.0f, -3.0f, 4.0f, -1.0f, 2.0f, -3.0f, 4.0f);
+        __m256 b = _mm256_set_ps (1.0f, 2.0f, 3.0f, 4.0f, 1.0f, 2.0f, 3.0f, 4.0f);
+        __m256 result = _mm256_add_ps (a, b);
+        return 0;
+    }" AVX_FOUND)
 
-# Check AVX 2
-set(CMAKE_REQUIRED_FLAGS ${AVX2_FLAG})
-set(AVX2_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
-CHECK_CXX_SOURCE_RUNS("
-#include <immintrin.h>
-int main()
-{
-    __m256i a = _mm256_set_epi32 (-1, 2, -3, 4, -1, 2, -3, 4);
-    __m256i result = _mm256_abs_epi32 (a);
-    return 0;
-}" AVX2_FOUND)
+    # Check AVX 2
+    set(CMAKE_REQUIRED_FLAGS ${AVX2_FLAG})
+    set(AVX2_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
+    CHECK_CXX_SOURCE_RUNS("
+    #include <immintrin.h>
+    int main()
+    {
+        __m256i a = _mm256_set_epi32 (-1, 2, -3, 4, -1, 2, -3, 4);
+        __m256i result = _mm256_abs_epi32 (a);
+        return 0;
+    }" AVX2_FOUND)
 
-# Check AVX512F
-set(CMAKE_REQUIRED_FLAGS ${AVX512F_FLAG})
-set(AVX512F_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
-CHECK_CXX_SOURCE_RUNS("
-#include <immintrin.h>
-int main()
-{
-    __m512i a = _mm512_set_epi32 (-1, 2, -3, 4, -1, 2, -3, 4,
-                                  13, -5, 6, -7, 9, 2, -6, 3);
-    __m512i result = _mm512_abs_epi32 (a);
-    return 0;
-}" AVX512F_FOUND)
+    # Check AVX512F
+    set(CMAKE_REQUIRED_FLAGS ${AVX512F_FLAG})
+    set(AVX512F_FOUND_EXITCODE 1 CACHE STRING "Result from TRY_RUN" FORCE)
+    CHECK_CXX_SOURCE_RUNS("
+    #include <immintrin.h>
+    int main()
+    {
+        __m512i a = _mm512_set_epi32 (-1, 2, -3, 4, -1, 2, -3, 4,
+                                      13, -5, 6, -7, 9, 2, -6, 3);
+        __m512i result = _mm512_abs_epi32 (a);
+        return 0;
+    }" AVX512F_FOUND)
+endif(NOT WIN32)
 
 set(CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS_RETAINED})
 mark_as_advanced(MMX_FOUND SSE2_FOUND SSE3_FOUND AVX_FOUND AVX2_FOUND AVX512F_FOUND)

paddle/fluid/CMakeLists.txt

@@ -3,13 +3,9 @@ add_subdirectory(platform)
 add_subdirectory(framework)
 add_subdirectory(operators)
 add_subdirectory(string)
-
-add_subdirectory(pybind)
-if (NOT WIN32)
 add_subdirectory(recordio)
-endif(NOT WIN32)
+add_subdirectory(pybind)
 
 # NOTE: please add subdirectory inference at last.
 add_subdirectory(inference)
-
 add_subdirectory(train)

paddle/fluid/framework/CMakeLists.txt

@@ -31,9 +31,7 @@ function(windows_symbolic TARGET)
 endfunction()
 
 add_subdirectory(ir)
-if (NOT WIN32)
 add_subdirectory(details)
-endif (NOT WIN32)
 # ddim lib
 proto_library(framework_proto SRCS framework.proto)
 
@@ -68,11 +66,7 @@ if(WITH_GPU)
 else()
   cc_test(mixed_vector_test SRCS mixed_vector_test.cc DEPS place memory device_context tensor)
 endif()
-if (NOT WIN32)
-  cc_library(lod_tensor SRCS lod_tensor.cc DEPS ddim place tensor framework_proto recordio version)
-else()
-  cc_library(lod_tensor SRCS lod_tensor.cc DEPS ddim place tensor framework_proto version)
-endif (NOT WIN32)
+cc_library(lod_tensor SRCS lod_tensor.cc DEPS ddim place tensor framework_proto recordio version)
 
 cc_test(lod_tensor_test SRCS lod_tensor_test.cc DEPS lod_tensor memory)
 nv_test(lod_tensor_gpu_test SRCS lod_tensor_test.cu DEPS lod_tensor)
@@ -122,13 +116,8 @@ cc_test(op_proto_maker_test SRCS op_proto_maker_test.cc DEPS op_proto_maker)
 cc_library(op_info SRCS op_info.cc DEPS attribute framework_proto)
 cc_library(shape_inference SRCS shape_inference.cc DEPS ddim attribute device_context)
 
-if (NOT WIN32)
 cc_library(operator SRCS operator.cc DEPS op_info device_context tensor scope glog
     shape_inference data_transform lod_tensor profiler)
-else()
-cc_library(operator SRCS operator.cc DEPS op_info device_context tensor scope glog
-    shape_inference data_transform lod_tensor)
-endif(NOT WIN32)
 
 cc_test(operator_test SRCS operator_test.cc DEPS operator op_registry device_context)
 
@@ -183,12 +172,10 @@ else()
   cc_test(test_naive_executor SRCS naive_executor_test.cc DEPS naive_executor elementwise_add_op)
 endif()
 
-if (NOT WIN32)
 cc_library(parallel_executor SRCS parallel_executor.cc DEPS
         threaded_ssa_graph_executor scope_buffered_ssa_graph_executor
         graph build_strategy
         fast_threaded_ssa_graph_executor)
-endif() # NOT WIN32
 
 cc_library(prune SRCS prune.cc DEPS framework_proto)
 cc_test(prune_test SRCS prune_test.cc DEPS op_info prune recurrent_op device_context)

paddle/fluid/framework/details/fast_threaded_ssa_graph_executor.h

@@ -13,9 +13,9 @@
 // limitations under the License.
 
 #pragma once
+#include <ThreadPool.h>
 #include <string>
 #include <vector>
-#include "ThreadPool.h"
 #include "paddle/fluid/framework/blocking_queue.h"
 #include "paddle/fluid/framework/details/exception_holder.h"
 #include "paddle/fluid/framework/details/execution_strategy.h"

paddle/fluid/framework/eigen.h

@@ -13,11 +13,6 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #pragma once
-// logging.h and windows.h conflict
-#define GLOG_NO_ABBREVIATED_SEVERITIES
-// solve static linking error in windows
-// https://github.com/google/glog/issues/301
-#define GOOGLE_GLOG_DLL_DECL
 
 #include "paddle/fluid/framework/tensor.h"
 #include "unsupported/Eigen/CXX11/Tensor"

paddle/fluid/framework/op_registry.h

@@ -23,11 +23,6 @@ limitations under the License. */
 #include <unordered_map>
 #include <unordered_set>
 
-#if defined(_WIN32)
-#define GLOG_NO_ABBREVIATED_SEVERITIES  // msvc conflict logging with windows.h
-#define GOOGLE_GLOG_DLL_DECL
-#endif
-
 #include "glog/logging.h"  // For VLOG()
 #include "paddle/fluid/framework/attribute.h"
 #include "paddle/fluid/framework/details/op_registry.h"

paddle/fluid/framework/operator.cc

@@ -11,8 +11,6 @@ distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
-#define GLOG_NO_ABBREVIATED_SEVERITIES
-#define GOOGLE_GLOG_DLL_DECL
 
 #include <gflags/gflags.h>
 #include <glog/logging.h>

paddle/fluid/framework/operator.h

@@ -20,8 +20,6 @@ limitations under the License. */
 #include <tuple>
 #include <unordered_map>
 #include <vector>
-#define GLOG_NO_ABBREVIATED_SEVERITIES
-#define GOOGLE_GLOG_DLL_DECL
 
 #include "glog/logging.h"  // For VLOG
 #include "paddle/fluid/framework/attribute.h"

paddle/fluid/inference/api/api_impl.h

@@ -14,12 +14,6 @@ limitations under the License. */
 
 #pragma once
 
-// logging.h and windows.h conflict
-#define GLOG_NO_ABBREVIATED_SEVERITIES
-// solve static linking error in windows
-// https://github.com/google/glog/issues/301
-#define GOOGLE_GLOG_DLL_DECL
-
 #include <glog/logging.h>
 #include <map>
 #include <memory>

paddle/fluid/memory/allocation/cpu_allocator.h

@@ -15,6 +15,12 @@
 #pragma once
 #include "paddle/fluid/memory/allocation/allocator.h"
 
+#ifdef _WIN32
+#define posix_memalign_free _aligned_free
+#define posix_memalign(p, a, s) \
+  (((*(p)) = _aligned_malloc((s), (a))), *(p) ? 0 : errno)
+#endif
+
 namespace paddle {
 namespace memory {
 namespace allocation {

paddle/fluid/operators/CMakeLists.txt

@@ -22,9 +22,7 @@ if(WITH_DISTRIBUTE)
     add_subdirectory(distributed_ops)
 endif()
 
-if (NOT WIN32)
-    add_subdirectory(reader)
-endif()
+add_subdirectory(reader)
 
 if (NOT WIN32)
     add_subdirectory(nccl)
@@ -42,7 +40,7 @@ endif()
 register_operators(EXCLUDES warpctc_op conv_fusion_op DEPS ${OP_HEADER_DEPS})
 
 # warpctc_op needs cudnn 7 above
-if (WITH_GPU)
+if (WITH_GPU AND NOT WIN32)
     if (${CUDNN_MAJOR_VERSION} VERSION_LESS 7)
         op_library(warpctc_op DEPS dynload_warpctc sequence_padding sequence_scale SRCS warpctc_op.cc warpctc_op.cu.cc)
     else()
@@ -59,10 +57,12 @@ endif()
 
 set(COMMON_OP_DEPS ${OP_HEADER_DEPS})
 
-set(COMMON_OP_DEPS ${COMMON_OP_DEPS} selected_rows_functor selected_rows lod_tensor maxouting unpooling pooling lod_rank_table context_project sequence_pooling executor dynload_warpctc sequence_padding sequence_scale cos_sim_functor memory jit_kernel concat_and_split cross_entropy softmax vol2col im2col sampler)
+set(COMMON_OP_DEPS ${COMMON_OP_DEPS} selected_rows_functor selected_rows lod_tensor maxouting unpooling pooling lod_rank_table context_project sequence_pooling executor)
 if (NOT WIN32)
-  set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence2batch lstm_compute matrix_bit_code gru_compute activation_functions)
+    set(COMMON_OP_DEPS ${COMMON_OP_DEPS} dynload_warpctc)
 endif()
+set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence_padding sequence_scale cos_sim_functor memory jit_kernel concat_and_split cross_entropy softmax vol2col im2col sampler)
+set(COMMON_OP_DEPS ${COMMON_OP_DEPS} sequence2batch lstm_compute matrix_bit_code gru_compute activation_functions)
 if (WITH_GPU)
   set(COMMON_OP_DEPS ${COMMON_OP_DEPS} depthwise_conv)
 endif()

paddle/fluid/operators/hierarchical_sigmoid_op.h

@@ -111,7 +111,7 @@ class HierarchicalSigmoidGradOpKernel : public framework::OpKernel<T> {
     auto pre_out_mat = EigenMatrix<T>::From(*pre_out);
     auto pre_out_grad_mat = EigenMatrix<T>::From(pre_out_grad);
     auto out_grad_mat = EigenMatrix<T>::From(*out_grad);
-    Eigen::array<int, 2> bcast({{1, static_cast<int>(pre_out_grad.dims()[1])}});
+    Eigen::array<int, 2> bcast{1, static_cast<int>(pre_out_grad.dims()[1])};
 
     // softrelu derivative
     pre_out_grad_mat.device(place) =

paddle/fluid/operators/math/CMakeLists.txt

@@ -1,6 +1,4 @@
-if (NOT WIN32)
-    add_subdirectory(detail)
-endif(NOT WIN32)
+add_subdirectory(detail)
 
 function(math_library TARGET)
     # math_library is a function to create math library.
@@ -43,10 +41,8 @@ math_library(depthwise_conv)
 math_library(im2col)
 math_library(sampler)
 
-if (NOT WIN32) # windows do not support avx functions yet.
-    math_library(gru_compute DEPS activation_functions math_function)
-    math_library(lstm_compute DEPS activation_functions)
-endif (NOT WIN32)
+math_library(gru_compute DEPS activation_functions math_function)
+math_library(lstm_compute DEPS activation_functions)
 
 cc_library(blas SRCS blas.cc DEPS cblas framework_proto device_context)
 math_library(math_function DEPS blas)
@@ -58,9 +54,9 @@ math_library(sequence_padding)
 math_library(sequence_pooling DEPS math_function)
 math_library(sequence_scale)
 math_library(softmax DEPS math_function)
-if (NOT WIN32)
-    math_library(matrix_bit_code)
-endif (NOT WIN32)
+
+math_library(matrix_bit_code)
+
 math_library(unpooling)
 math_library(vol2col)
 
@@ -76,13 +72,12 @@ if(WITH_GPU)
 endif()
 cc_test(concat_test SRCS concat_test.cc DEPS concat_and_split)
 cc_test(cpu_vec_test SRCS cpu_vec_test.cc DEPS blas cpu_info)
-if (NOT WIN32)
-    set(JIT_KERNEL_SRCS jit_kernel.cc jit_kernel_blas.cc jit_kernel_exp.cc jit_kernel_rnn.cc jit_kernel_crf_decode.cc jit_kernel_layer_norm.cc)
-    set(JIT_KERNEL_DEPS cpu_info cblas gflags enforce)
-    if(WITH_XBYAK)
-        list(APPEND JIT_KERNEL_SRCS jit_gen.cc jit_code.cc)
-        list(APPEND JIT_KERNEL_DEPS xbyak)
-    endif()
-    cc_library(jit_kernel SRCS ${JIT_KERNEL_SRCS} DEPS ${JIT_KERNEL_DEPS})
-    cc_test(jit_kernel_test SRCS jit_kernel_test.cc DEPS jit_kernel)
-endif (NOT WIN32)
+
+set(JIT_KERNEL_SRCS jit_kernel.cc jit_kernel_blas.cc jit_kernel_exp.cc jit_kernel_rnn.cc jit_kernel_crf_decode.cc jit_kernel_layer_norm.cc)
+set(JIT_KERNEL_DEPS cpu_info cblas gflags enforce)
+if(WITH_XBYAK)
+    list(APPEND JIT_KERNEL_SRCS jit_gen.cc jit_code.cc)
+    list(APPEND JIT_KERNEL_DEPS xbyak)
+endif()
+cc_library(jit_kernel SRCS ${JIT_KERNEL_SRCS} DEPS ${JIT_KERNEL_DEPS})
+cc_test(jit_kernel_test SRCS jit_kernel_test.cc DEPS jit_kernel)

paddle/fluid/operators/math/detail/activation_functions.h

@@ -15,6 +15,7 @@ limitations under the License. */
 #pragma once
 #include <math.h>
 #include <string>
+
 #include "paddle/fluid/platform/enforce.h"
 #include "paddle/fluid/platform/hostdevice.h"
 

paddle/fluid/operators/math/matrix_bit_code.h

@@ -67,7 +67,7 @@ inline constexpr size_t FindLastSet(size_t x) {
              : (std::is_same<size_t, unsigned long>::value  // NOLINT
                     ? (x ? 8 * sizeof(x) - __builtin_clzl(x) : 0)
                     : (x ? 8 * sizeof(x) - __builtin_clzll(x) : 0));
-
+}
 #else
 // windows don't have built-in clz, ctz function
 template <typename T>
@@ -92,7 +92,6 @@ inline int clz(const T& value) {
 
 inline size_t FindLastSet(size_t x) { return sizeof(size_t) * 8 - clz(x); }
 #endif  // !_WIN32
-}
 
 struct SimpleCode {
   SimpleCode(size_t code, size_t num_classes) : c_(code + num_classes) {}

paddle/fluid/operators/reader/create_py_reader_op.cc

@@ -74,7 +74,7 @@ class CreatePyReaderOpMaker : public FileReaderMakerBase {
              "Name of the `LoDTensorBlockingQueueHolder` variable");
 
     AddComment(R"DOC(
-			Create PyReader to support LoDTensor data feeding in Python side.
+      Create PyReader to support LoDTensor data feeding in Python side.
       )DOC");
   }
 };

paddle/fluid/operators/roi_align_op.cc

@@ -35,10 +35,10 @@ class ROIAlignOp : public framework::OperatorWithKernel {
                    "The format of input tensor is NCHW.");
     PADDLE_ENFORCE(rois_dims.size() == 2,
                    "ROIs should be a 2-D LoDTensor of shape (num_rois, 4)"
-                   "given as [[x1, y1, x2, y2], …].");
+                   "given as [[x1, y1, x2, y2], ...].");
     PADDLE_ENFORCE(rois_dims[1] == 4,
                    "ROIs should be a 2-D LoDTensor of shape (num_rois, 4)"
-                   "given as [[x1, y1, x2, y2], …].");
+                   "given as [[x1, y1, x2, y2], ...].");
     int pooled_height = ctx->Attrs().Get<int>("pooled_height");
     int pooled_width = ctx->Attrs().Get<int>("pooled_width");
     float spatial_scale = ctx->Attrs().Get<float>("spatial_scale");
@@ -103,7 +103,7 @@ class ROIAlignOpMaker : public framework::OpProtoAndCheckerMaker {
              "(LoDTensor), "
              "ROIs (Regions of Interest) to pool over. "
              "should be a 2-D LoDTensor of shape (num_rois, 4)"
-             "given as [[x1, y1, x2, y2], …]. "
+             "given as [[x1, y1, x2, y2], ...]. "
              "(x1, y1) is the top left coordinates, and "
              "(x2, y2) is the bottom right coordinates.");
     AddOutput("Out",

paddle/fluid/operators/roi_pool_op.cc

@@ -40,10 +40,10 @@ class ROIPoolOp : public framework::OperatorWithKernel {
                    "The format of input tensor is NCHW.");
     PADDLE_ENFORCE(rois_dims.size() == 2,
                    "ROIs should be a 2-D LoDTensor of shape (num_rois, 4)"
-                   "given as [[x1, y1, x2, y2], …].");
+                   "given as [[x1, y1, x2, y2], ...].");
     PADDLE_ENFORCE(rois_dims[1] == kROISize,
                    "ROIs should be a 2-D LoDTensor of shape (num_rois, 4)"
-                   "given as [[x1, y1, x2, y2], …].");
+                   "given as [[x1, y1, x2, y2], ...].");
 
     int pooled_height = ctx->Attrs().Get<int>("pooled_height");
     int pooled_width = ctx->Attrs().Get<int>("pooled_width");
@@ -110,7 +110,7 @@ class ROIPoolOpMaker : public framework::OpProtoAndCheckerMaker {
              "(LoDTensor), "
              "ROIs (Regions of Interest) to pool over. "
              "should be a 2-D LoDTensor of shape (num_rois, 4)"
-             "given as [[x1, y1, x2, y2], …]. "
+             "given as [[x1, y1, x2, y2], ...]. "
              "Where batch_id is the id of the data, "
              "(x1, y1) is the top left coordinates, and "
              "(x2, y2) is the bottom right coordinates.");

paddle/fluid/operators/space_to_depth_op.cc

@@ -86,7 +86,7 @@ class SpaceToDepthOpMaker : public framework::OpProtoAndCheckerMaker {
         .GreaterThan(1);
     AddComment(R"DOC(
         reorg operator used in Yolo v2.
-        The equation is: C2 = C1/blocksize * blocksize, W2 = W1 * blocksize + offset % blocksize, H2 = H1 * blocksize + offset / blocksize,   
+        The equation is: C2 = C1/blocksize * blocksize, W2 = W1 * blocksize + offset % blocksize, H2 = H1 * blocksize + offset / blocksize,
 
         Reshape Input(X) into the shape according to Attr(blocksize). The
         data in Input(X) are unchanged.