Merge pull request #12738 from luotao1/anakin_cpu

support anakin for only-cpu environment

CMakeLists.txt

@@ -213,9 +213,11 @@ include(configure)          # add paddle env configuration
 if(WITH_GPU)
     include(cuda)
     include(tensorrt)
+endif()
+if(WITH_MKL OR WITH_MKLML)
     include(external/anakin)
 elseif()
-    set(WITH_ANAKIN OFF CACHE STRING "Anakin is used in GPU only now." FORCE)
+    set(WITH_ANAKIN OFF CACHE STRING "Anakin is used in MKL only now." FORCE)
 endif()
 
 include(generic)            # simplify cmake module

cmake/external/anakin.cmake

@@ -16,16 +16,6 @@ set(ANAKIN_LIBRARY     ${ANAKIN_INSTALL_DIR})
 set(ANAKIN_SHARED_LIB  ${ANAKIN_LIBRARY}/libanakin.so)
 set(ANAKIN_SABER_LIB   ${ANAKIN_LIBRARY}/libanakin_saber_common.so)
 
-# TODO(luotao): ANAKIN_MODLE_URL etc will move to demo ci later.
-set(INFERENCE_URL "http://paddle-inference-dist.bj.bcebos.com")
-set(ANAKIN_MODLE_URL "${INFERENCE_URL}/mobilenet_v2.anakin.bin")
-set(ANAKIN_RNN_MODLE_URL "${INFERENCE_URL}/anakin_test%2Fditu_rnn.anakin2.model.bin")
-set(ANAKIN_RNN_DATA_URL "${INFERENCE_URL}/anakin_test%2Fditu_rnn_data.txt")
-execute_process(COMMAND bash -c "mkdir -p ${ANAKIN_SOURCE_DIR}")
-execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_MODLE_URL} -N")
-execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_RNN_MODLE_URL} -N")
-execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_RNN_DATA_URL} -N")
-
 include_directories(${ANAKIN_INCLUDE})
 include_directories(${ANAKIN_INCLUDE}/saber/)
 include_directories(${ANAKIN_INCLUDE}/saber/core/)
@@ -48,6 +38,11 @@ set(ANAKIN_COMPILE_EXTRA_FLAGS
     -Wno-reorder
     -Wno-error=cpp)
 
+if(WITH_GPU)
+    set(CMAKE_ARGS_PREFIX -DUSE_GPU_PLACE=YES -DCUDNN_ROOT=${CUDNN_ROOT} -DCUDNN_INCLUDE_DIR=${CUDNN_INCLUDE_DIR})
+else()
+    set(CMAKE_ARGS_PREFIX -DUSE_GPU_PLACE=NO)
+endif()
 ExternalProject_Add(
     extern_anakin
     ${EXTERNAL_PROJECT_LOG_ARGS}
@@ -56,13 +51,11 @@ ExternalProject_Add(
     GIT_TAG             "9424277cf9ae180a14aff09560d3cd60a49c76d2"
     PREFIX              ${ANAKIN_SOURCE_DIR}
     UPDATE_COMMAND      ""
-    CMAKE_ARGS          -DUSE_GPU_PLACE=YES
+    CMAKE_ARGS          ${CMAKE_ARGS_PREFIX}
                         -DUSE_X86_PLACE=YES
                         -DBUILD_WITH_UNIT_TEST=NO
                         -DPROTOBUF_ROOT=${THIRD_PARTY_PATH}/install/protobuf
                         -DMKLML_ROOT=${THIRD_PARTY_PATH}/install/mklml
-                        -DCUDNN_ROOT=${CUDNN_ROOT}
-                        -DCUDNN_INCLUDE_DIR=${CUDNN_INCLUDE_DIR}
                         -DENABLE_OP_TIMER=${ANAKIN_ENABLE_OP_TIMER}
                         ${EXTERNAL_OPTIONAL_ARGS}
     CMAKE_CACHE_ARGS    -DCMAKE_INSTALL_PREFIX:PATH=${ANAKIN_INSTALL_DIR}

cmake/inference_lib.cmake

@@ -145,7 +145,7 @@ copy(memory_lib
 set(inference_deps paddle_fluid_shared paddle_fluid)
 
 set(module "inference/api")
-if (WITH_ANAKIN AND WITH_GPU)
+if (WITH_ANAKIN AND WITH_MKL)
     copy(anakin_inference_lib DEPS paddle_inference_api inference_anakin_api
         SRCS
         ${PADDLE_BINARY_DIR}/paddle/fluid/inference/api/libinference_anakin_api* # compiled anakin api

paddle/fluid/inference/api/CMakeLists.txt

@@ -61,7 +61,7 @@ cc_library(paddle_inference_tensorrt_subgraph_engine
 inference_api_test(test_api_tensorrt_subgraph_engine SRC api_tensorrt_subgraph_engine_tester.cc ARGS test_word2vec)
 endif()
 
-if (WITH_ANAKIN AND WITH_GPU) # only needed in CI
+if (WITH_ANAKIN AND WITH_MKL) # only needed in CI
     # compile the libinference_anakin_api.a and anakin.so.
     cc_library(inference_anakin_api SRCS api.cc api_anakin_engine.cc DEPS anakin_shared anakin_saber mklml)
     cc_library(inference_anakin_api_shared SHARED SRCS api.cc api_anakin_engine.cc DEPS anakin_shared anakin_saber)
@@ -71,12 +71,24 @@ if (WITH_ANAKIN AND WITH_GPU) # only needed in CI
     anakin_target(inference_anakin_api)
     anakin_target(inference_anakin_api_shared)
     if (WITH_TESTING)
-        cc_test(api_anakin_engine_tester SRCS api_anakin_engine_tester.cc 
-                ARGS --model=${ANAKIN_SOURCE_DIR}/mobilenet_v2.anakin.bin
-                DEPS inference_anakin_api_shared dynload_cuda SERIAL)
+        # TODO(luotao): ANAKIN_MODLE_URL etc will move to demo ci later.
+        set(INFERENCE_URL "http://paddle-inference-dist.bj.bcebos.com")
+        set(ANAKIN_RNN_MODLE_URL "${INFERENCE_URL}/anakin_test%2Fditu_rnn.anakin2.model.bin")
+        set(ANAKIN_RNN_DATA_URL "${INFERENCE_URL}/anakin_test%2Fditu_rnn_data.txt")
+        execute_process(COMMAND bash -c "mkdir -p ${ANAKIN_SOURCE_DIR}")
+        execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_RNN_MODLE_URL} -N")
+        execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_RNN_DATA_URL} -N")
+        if(WITH_GPU)
+            set(anakin_test_extra_deps dynload_cuda)
+            set(ANAKIN_MODLE_URL "${INFERENCE_URL}/mobilenet_v2.anakin.bin")
+            execute_process(COMMAND bash -c "cd ${ANAKIN_SOURCE_DIR}; wget -q --no-check-certificate ${ANAKIN_MODLE_URL} -N")
+            cc_test(api_anakin_engine_tester SRCS api_anakin_engine_tester.cc 
+                    ARGS --model=${ANAKIN_SOURCE_DIR}/mobilenet_v2.anakin.bin
+                    DEPS inference_anakin_api_shared ${anakin_test_extra_deps} SERIAL)
+        endif()
         cc_test(api_anakin_engine_rnn_tester SRCS api_anakin_engine_rnn_tester.cc 
                 ARGS --model=${ANAKIN_SOURCE_DIR}/anakin_test%2Fditu_rnn.anakin2.model.bin 
                      --datapath=${ANAKIN_SOURCE_DIR}/anakin_test%2Fditu_rnn_data.txt
-                DEPS inference_anakin_api_shared dynload_cuda SERIAL)
+                DEPS inference_anakin_api_shared ${anakin_test_extra_deps} SERIAL)
     endif(WITH_TESTING)
 endif()

paddle/fluid/inference/api/api_anakin_engine.cc

@@ -193,7 +193,9 @@ PaddleInferenceAnakinPredictor<Target>::Clone() {
   return std::move(cls);
 }
 
+#ifdef PADDLE_WITH_CUDA
 template class PaddleInferenceAnakinPredictor<anakin::NV>;
+#endif
 template class PaddleInferenceAnakinPredictor<anakin::X86>;
 
 // A factory to help create difference predictor.
@@ -202,10 +204,15 @@ std::unique_ptr<PaddlePredictor> CreatePaddlePredictor<
     AnakinConfig, PaddleEngineKind::kAnakin>(const AnakinConfig &config) {
   VLOG(3) << "Anakin Predictor create.";
   if (config.target_type == AnakinConfig::NVGPU) {
+#ifdef PADDLE_WITH_CUDA
     VLOG(3) << "Anakin Predictor create on [ NVIDIA GPU ].";
     std::unique_ptr<PaddlePredictor> x(
         new PaddleInferenceAnakinPredictor<anakin::NV>(config));
     return x;
+#else
+    LOG(ERROR) << "AnakinConfig::NVGPU could not used in ONLY-CPU environment";
+    return nullptr;
+#endif
   } else if (config.target_type == AnakinConfig::X86) {
     VLOG(3) << "Anakin Predictor create on [ Intel X86 ].";
     std::unique_ptr<PaddlePredictor> x(