Merge pull request #296 from wangjiawei04/jiawei/dist_kv_benchmark

Dist KV Op, Cube Criteo Demo & Benchmark

core/configure/proto/server_configure.proto

@@ -52,9 +52,10 @@ message ModelToolkitConf { repeated EngineDesc engines = 1; };
 message ResourceConf {
   required string model_toolkit_path = 1;
   required string model_toolkit_file = 2;
-  optional string cube_config_file = 3;
-  optional string general_model_path = 4;
-  optional string general_model_file = 5;
+  optional string general_model_path = 3;
+  optional string general_model_file = 4;
+  optional string cube_config_path = 5;
+  optional string cube_config_file = 6;
 };
 
 // DAG node depency info

core/cube/cube-api/include/cube_api.h

@@ -99,6 +99,13 @@ class CubeAPI {
                std::function<void(DictValue*, size_t)> parse,
                std::string* version);
 
+  /**
+   * @brief: get all table names from cube server, thread safe.
+   * @param [out] vals: vector of table names
+   *
+   */
+  std::vector<std::string> get_table_names();
+
  public:
   static const char* error_msg(int error_code);
 

core/cube/cube-api/src/cube_api.cpp

@@ -682,5 +682,13 @@ int CubeAPI::opt_seek(const std::string& dict_name,
   return ret;
 }
 
+std::vector<std::string> CubeAPI::get_table_names() {
+  const std::vector<const MetaInfo*> metas = _meta->metas();
+  std::vector<std::string> table_names;
+  for (auto itr = metas.begin(); itr != metas.end(); ++itr) {
+    table_names.push_back((*itr)->dict_name);
+  }
+  return table_names;
+}
 }  // namespace mcube
 }  // namespace rec

core/general-server/CMakeLists.txt

+include_directories(SYSTEM  ${CMAKE_CURRENT_LIST_DIR}/../../)
 include(op/CMakeLists.txt)
 include(proto/CMakeLists.txt)
 add_executable(serving ${serving_srcs})

core/general-server/op/general_dist_kv_infer_op.cpp

+// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// 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.
+
+#include "core/general-server/op/general_dist_kv_infer_op.h"
+#include <algorithm>
+#include <iostream>
+#include <memory>
+#include <sstream>
+#include <unordered_map>
+#include <utility>
+#include "core/cube/cube-api/include/cube_api.h"
+#include "core/predictor/framework/infer.h"
+#include "core/predictor/framework/memory.h"
+#include "core/predictor/framework/resource.h"
+#include "core/util/include/timer.h"
+
+namespace baidu {
+namespace paddle_serving {
+namespace serving {
+
+using baidu::paddle_serving::Timer;
+using baidu::paddle_serving::predictor::MempoolWrapper;
+using baidu::paddle_serving::predictor::general_model::Tensor;
+using baidu::paddle_serving::predictor::general_model::Response;
+using baidu::paddle_serving::predictor::general_model::Request;
+using baidu::paddle_serving::predictor::general_model::FetchInst;
+using baidu::paddle_serving::predictor::InferManager;
+using baidu::paddle_serving::predictor::PaddleGeneralModelConfig;
+
+int GeneralDistKVInferOp::inference() {
+  VLOG(2) << "Going to run inference";
+  const GeneralBlob *input_blob = get_depend_argument<GeneralBlob>(pre_name());
+  VLOG(2) << "Get precedent op name: " << pre_name();
+  GeneralBlob *output_blob = mutable_data<GeneralBlob>();
+
+  if (!input_blob) {
+    LOG(ERROR) << "Failed mutable depended argument, op:" << pre_name();
+    return -1;
+  }
+
+  const TensorVector *in = &input_blob->tensor_vector;
+  TensorVector *out = &output_blob->tensor_vector;
+  int batch_size = input_blob->GetBatchSize();
+  VLOG(2) << "input batch size: " << batch_size;
+  std::vector<uint64_t> keys;
+  std::vector<rec::mcube::CubeValue> values;
+  int sparse_count = 0;
+  int dense_count = 0;
+  std::vector<std::pair<int64_t *, size_t>> dataptr_size_pairs;
+  size_t key_len = 0;
+  for (size_t i = 0; i < in->size(); ++i) {
+    if (in->at(i).dtype != paddle::PaddleDType::INT64) {
+      ++dense_count;
+      continue;
+    }
+    ++sparse_count;
+    size_t elem_num = 1;
+    for (size_t s = 0; s < in->at(i).shape.size(); ++s) {
+      elem_num *= in->at(i).shape[s];
+    }
+    key_len += elem_num;
+    int64_t *data_ptr = static_cast<int64_t *>(in->at(i).data.data());
+    dataptr_size_pairs.push_back(std::make_pair(data_ptr, elem_num));
+  }
+  keys.resize(key_len);
+  int key_idx = 0;
+  for (size_t i = 0; i < dataptr_size_pairs.size(); ++i) {
+    std::copy(dataptr_size_pairs[i].first,
+              dataptr_size_pairs[i].first + dataptr_size_pairs[i].second,
+              keys.begin() + key_idx);
+    key_idx += dataptr_size_pairs[i].second;
+  }
+  rec::mcube::CubeAPI *cube = rec::mcube::CubeAPI::instance();
+  std::vector<std::string> table_names = cube->get_table_names();
+  if (table_names.size() == 0) {
+    LOG(ERROR) << "cube init error or cube config not given.";
+    return -1;
+  }
+  int ret = cube->seek(table_names[0], keys, &values);
+
+  if (values.size() != keys.size() || values[0].buff.size() == 0) {
+    LOG(ERROR) << "cube value return null";
+  }
+  size_t EMBEDDING_SIZE = values[0].buff.size() / sizeof(float);
+  TensorVector sparse_out;
+  sparse_out.resize(sparse_count);
+  TensorVector dense_out;
+  dense_out.resize(dense_count);
+  int cube_val_idx = 0;
+  int sparse_idx = 0;
+  int dense_idx = 0;
+  std::unordered_map<int, int> in_out_map;
+  baidu::paddle_serving::predictor::Resource &resource =
+      baidu::paddle_serving::predictor::Resource::instance();
+  std::shared_ptr<PaddleGeneralModelConfig> model_config =
+      resource.get_general_model_config();
+  for (size_t i = 0; i < in->size(); ++i) {
+    if (in->at(i).dtype != paddle::PaddleDType::INT64) {
+      dense_out[dense_idx] = in->at(i);
+      ++dense_idx;
+      continue;
+    }
+
+    sparse_out[sparse_idx].lod.resize(in->at(i).lod.size());
+    for (size_t x = 0; x < sparse_out[sparse_idx].lod.size(); ++x) {
+      sparse_out[sparse_idx].lod[x].resize(in->at(i).lod[x].size());
+      std::copy(in->at(i).lod[x].begin(),
+                in->at(i).lod[x].end(),
+                sparse_out[sparse_idx].lod[x].begin());
+    }
+    sparse_out[sparse_idx].dtype = paddle::PaddleDType::FLOAT32;
+    sparse_out[sparse_idx].shape.push_back(
+        sparse_out[sparse_idx].lod[0].back());
+    sparse_out[sparse_idx].shape.push_back(EMBEDDING_SIZE);
+    sparse_out[sparse_idx].name = model_config->_feed_name[i];
+    sparse_out[sparse_idx].data.Resize(sparse_out[sparse_idx].lod[0].back() *
+                                       EMBEDDING_SIZE * sizeof(float));
+    float *dst_ptr = static_cast<float *>(sparse_out[sparse_idx].data.data());
+    for (int x = 0; x < sparse_out[sparse_idx].lod[0].back(); ++x) {
+      float *data_ptr = dst_ptr + x * EMBEDDING_SIZE;
+      memcpy(data_ptr,
+             values[cube_val_idx].buff.data(),
+             values[cube_val_idx].buff.size());
+      cube_val_idx++;
+    }
+    ++sparse_idx;
+  }
+  TensorVector infer_in;
+  infer_in.insert(infer_in.end(), dense_out.begin(), dense_out.end());
+  infer_in.insert(infer_in.end(), sparse_out.begin(), sparse_out.end());
+
+  output_blob->SetBatchSize(batch_size);
+
+  VLOG(2) << "infer batch size: " << batch_size;
+
+  Timer timeline;
+  int64_t start = timeline.TimeStampUS();
+  timeline.Start();
+
+  if (InferManager::instance().infer(
+          GENERAL_MODEL_NAME, &infer_in, out, batch_size)) {
+    LOG(ERROR) << "Failed do infer in fluid model: " << GENERAL_MODEL_NAME;
+    return -1;
+  }
+
+  int64_t end = timeline.TimeStampUS();
+  CopyBlobInfo(input_blob, output_blob);
+  AddBlobInfo(output_blob, start);
+  AddBlobInfo(output_blob, end);
+  return 0;
+}
+DEFINE_OP(GeneralDistKVInferOp);
+
+}  // namespace serving
+}  // namespace paddle_serving
+}  // namespace baidu

core/general-server/op/general_dist_kv_infer_op.h

+// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// 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.
+
+#pragma once
+#include <string>
+#include <vector>
+#ifdef BCLOUD
+#ifdef WITH_GPU
+#include "paddle/paddle_inference_api.h"
+#else
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
+#endif
+#else
+#include "paddle_inference_api.h"  // NOLINT
+#endif
+#include "core/general-server/general_model_service.pb.h"
+#include "core/general-server/op/general_infer_helper.h"
+
+namespace baidu {
+namespace paddle_serving {
+namespace serving {
+
+class GeneralDistKVInferOp
+    : public baidu::paddle_serving::predictor::OpWithChannel<GeneralBlob> {
+ public:
+  typedef std::vector<paddle::PaddleTensor> TensorVector;
+
+  DECLARE_OP(GeneralDistKVInferOp);
+
+  int inference();
+};
+
+}  // namespace serving
+}  // namespace paddle_serving
+}  // namespace baidu

core/predictor/CMakeLists.txt

@@ -3,7 +3,7 @@ include(common/CMakeLists.txt)
 include(op/CMakeLists.txt)
 include(mempool/CMakeLists.txt)
 include(framework/CMakeLists.txt)
-#include(plugin/CMakeLists.txt)
+include(tools/CMakeLists.txt)
 include(src/CMakeLists.txt)
 
 

core/predictor/common/CMakeLists.txt

-FILE(GLOB common_srcs ${CMAKE_CURRENT_LIST_DIR}/*.cpp)
+FILE(GLOB common_srcs ${CMAKE_CURRENT_LIST_DIR}/constant.cpp)
 LIST(APPEND pdserving_srcs ${common_srcs})

core/predictor/framework/resource.cpp

@@ -141,6 +141,17 @@ int Resource::initialize(const std::string& path, const std::string& file) {
     LOG(ERROR) << "unable to create tls_bthread_key of thrd_data";
     return -1;
   }
+  // init rocksDB or cube instance
+  if (resource_conf.has_cube_config_file() &&
+      resource_conf.has_cube_config_path()) {
+    LOG(INFO) << "init cube client, path[ " << resource_conf.cube_config_path()
+              << " ], config file [ " << resource_conf.cube_config_file()
+              << " ].";
+    rec::mcube::CubeAPI* cube = rec::mcube::CubeAPI::instance();
+    std::string cube_config_fullpath = "./" + resource_conf.cube_config_path() +
+                                       "/" + resource_conf.cube_config_file();
+    this->cube_config_fullpath = cube_config_fullpath;
+  }
 
   THREAD_SETSPECIFIC(_tls_bspec_key, NULL);
   return 0;
@@ -149,6 +160,15 @@ int Resource::initialize(const std::string& path, const std::string& file) {
 // model config
 int Resource::general_model_initialize(const std::string& path,
                                        const std::string& file) {
+  if (this->cube_config_fullpath.size() != 0) {
+    LOG(INFO) << "init cube by config file : " << this->cube_config_fullpath;
+    rec::mcube::CubeAPI* cube = rec::mcube::CubeAPI::instance();
+    int ret = cube->init(this->cube_config_fullpath.c_str());
+    if (ret != 0) {
+      LOG(ERROR) << "cube init error";
+      return -1;
+    }
+  }
   VLOG(2) << "general model path: " << path;
   VLOG(2) << "general model file: " << file;
   if (!FLAGS_enable_general_model) {

core/predictor/framework/resource.h

@@ -108,6 +108,7 @@ class Resource {
  private:
   int thread_finalize() { return 0; }
   std::shared_ptr<PaddleGeneralModelConfig> _config;
+  std::string cube_config_fullpath;
 
   THREAD_KEY_T _tls_bspec_key;
 };

core/predictor/proto/CMakeLists.txt

@@ -7,6 +7,7 @@ LIST(APPEND protofiles
         ${CMAKE_CURRENT_LIST_DIR}/./builtin_format.proto
         ${CMAKE_CURRENT_LIST_DIR}/./msg_data.proto
         ${CMAKE_CURRENT_LIST_DIR}/./xrecord_format.proto
+        ${CMAKE_CURRENT_LIST_DIR}/./framework.proto
 )
 
 PROTOBUF_GENERATE_SERVING_CPP(TRUE PROTO_SRCS PROTO_HDRS ${protofiles})

core/predictor/proto/framework.proto

+/* Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+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. */
+
+syntax = "proto2";
+option optimize_for = LITE_RUNTIME;
+package paddle.framework.proto;
+
+// Any incompatible changes to ProgramDesc and its dependencies should
+// raise the version defined version.h.
+//
+// Serailization and Deserialization codes should be modified in a way
+// that supports old versions following the version and compatibility policy.
+message Version { optional int64 version = 1 [ default = 0 ]; }
+
+enum AttrType {
+  INT = 0;
+  FLOAT = 1;
+  STRING = 2;
+  INTS = 3;
+  FLOATS = 4;
+  STRINGS = 5;
+  BOOLEAN = 6;
+  BOOLEANS = 7;
+  BLOCK = 8;
+  LONG = 9;
+  BLOCKS = 10;
+  LONGS = 11;
+}
+
+// OpDesc describes an instance of a C++ framework::OperatorBase
+// derived class type.
+message OpDesc {
+
+  message Attr {
+    required string name = 1;
+    required AttrType type = 2;
+    optional int32 i = 3;
+    optional float f = 4;
+    optional string s = 5;
+    repeated int32 ints = 6;
+    repeated float floats = 7;
+    repeated string strings = 8;
+    optional bool b = 10;
+    repeated bool bools = 11;
+    optional int32 block_idx = 12;
+    optional int64 l = 13;
+    repeated int32 blocks_idx = 14;
+    repeated int64 longs = 15;
+  };
+
+  message Var {
+    required string parameter = 1;
+    repeated string arguments = 2;
+  };
+
+  required string type = 3;
+  repeated Var inputs = 1;
+  repeated Var outputs = 2;
+  repeated Attr attrs = 4;
+  optional bool is_target = 5 [ default = false ];
+};
+
+// OpProto describes a C++ framework::OperatorBase derived class.
+message OpProto {
+
+  // VarProto describes the C++ type framework::Variable.
+  message Var {
+    required string name = 1;
+    required string comment = 2;
+
+    optional bool duplicable = 3 [ default = false ];
+    optional bool intermediate = 4 [ default = false ];
+    optional bool dispensable = 5 [ default = false ];
+  }
+
+  // AttrProto describes the C++ type Attribute.
+  message Attr {
+    required string name = 1;
+    required AttrType type = 2;
+    required string comment = 3;
+    // If that attribute is generated, it means the Paddle third
+    // language binding has responsibility to fill that
+    // attribute. End-User should not set that attribute.
+    optional bool generated = 4 [ default = false ];
+  }
+
+  required string type = 1;
+  repeated Var inputs = 2;
+  repeated Var outputs = 3;
+  repeated Attr attrs = 4;
+  required string comment = 5;
+}
+
+message VarType {
+  enum Type {
+    // Pod Types
+    BOOL = 0;
+    INT16 = 1;
+    INT32 = 2;
+    INT64 = 3;
+    FP16 = 4;
+    FP32 = 5;
+    FP64 = 6;
+    // Tensor<size_t> is used in C++.
+    SIZE_T = 19;
+    UINT8 = 20;
+    INT8 = 21;
+
+    // Other types that may need additional descriptions
+    LOD_TENSOR = 7;
+    SELECTED_ROWS = 8;
+    FEED_MINIBATCH = 9;
+    FETCH_LIST = 10;
+    STEP_SCOPES = 11;
+    LOD_RANK_TABLE = 12;
+    LOD_TENSOR_ARRAY = 13;
+    PLACE_LIST = 14;
+    READER = 15;
+    // Any runtime decided variable type is raw
+    // raw variables should manage their own allocations
+    // in operators like nccl_op
+    RAW = 17;
+    TUPLE = 18;
+  }
+
+  required Type type = 1;
+
+  message TensorDesc {
+    // Should only be PODType. Is enforced in C++
+    required Type data_type = 1;
+    repeated int64 dims = 2; // [UNK, 640, 480] is saved as [-1, 640, 480]
+  }
+  optional TensorDesc selected_rows = 2;
+
+  message LoDTensorDesc {
+    required TensorDesc tensor = 1;
+    optional int32 lod_level = 2 [ default = 0 ];
+  }
+  optional LoDTensorDesc lod_tensor = 3;
+
+  message LoDTensorArrayDesc {
+    required TensorDesc tensor = 1;
+    optional int32 lod_level = 2 [ default = 0 ];
+  }
+  optional LoDTensorArrayDesc tensor_array = 4;
+
+  message ReaderDesc { repeated LoDTensorDesc lod_tensor = 1; }
+  optional ReaderDesc reader = 5;
+
+  message Tuple { repeated Type element_type = 1; }
+  optional Tuple tuple = 7;
+}
+
+message VarDesc {
+  required string name = 1;
+  required VarType type = 2;
+  optional bool persistable = 3 [ default = false ];
+  // True if the variable is an input data and
+  // have to check the feed data shape and dtype
+  optional bool need_check_feed = 4 [ default = false ];
+}
+
+message BlockDesc {
+  required int32 idx = 1;
+  required int32 parent_idx = 2;
+  repeated VarDesc vars = 3;
+  repeated OpDesc ops = 4;
+  optional int32 forward_block_idx = 5 [ default = -1 ];
+}
+
+// CompatibleInfo is used to determine if a feature is compatible and
+// provides the information.
+message CompatibleInfo {
+  enum Type {
+    COMPATIBLE = 0;
+    DEFINITELY_NOT = 1;
+    POSSIBLE = 2;
+    BUG_FIX = 3;
+    PRECISION_CHANGE = 4;
+  }
+  required string version = 1;
+  required Type type = 2;
+}
+
+// In some cases, Paddle Fluid may perform operator definition iterations,
+// and the operator uses OpCompatibleMap for compatibility testing.
+message OpCompatibleMap {
+  message OpCompatiblePair {
+    required string op_name = 1;
+    required CompatibleInfo compatible_info = 2;
+  }
+  repeated OpCompatiblePair pair = 1;
+  optional string default_required_version = 2;
+}
+
+// Please refer to
+// https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/program.md
+// for more details.
+// TODO(panyx0718): A model can have multiple programs. Need a
+// way to distinguish them. Maybe ID or name?
+message ProgramDesc {
+  reserved 2; // For backward compatibility.
+  repeated BlockDesc blocks = 1;
+  optional Version version = 4;
+  optional OpCompatibleMap op_compatible_map = 3;
+}

core/predictor/tools/CMakeLists.txt

+set(seq_gen_src ${CMAKE_CURRENT_LIST_DIR}/seq_generator.cpp  ${CMAKE_CURRENT_LIST_DIR}/seq_file.cpp)
+LIST(APPEND seq_gen_src ${PROTO_SRCS})
+add_executable(seq_generator ${seq_gen_src})
+target_link_libraries(seq_generator protobuf -lpthread)

core/predictor/tools/seq_file.cpp

+// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// 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.
+
+#include <netinet/in.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <ctime>
+#include <iostream>
+
+#include <unistd.h>
+#include "seq_file.h"
+
+SeqFileWriter::SeqFileWriter(const char *file) {
+  _fs = new std::ofstream(file, std::ios::binary);
+  std::srand(std::time(0));
+  for (int i = 0; i < SYNC_MARKER_SIZE; ++i) {
+    _sync_marker[i] = std::rand() % 255;
+  }
+
+  _write_seq_header();
+
+  _bytes_to_prev_sync = 0;
+}
+
+void SeqFileWriter::close() {
+  _fs->close();
+  delete _fs;
+}
+
+SeqFileWriter::~SeqFileWriter() { close(); }
+
+void SeqFileWriter::_write_sync_marker() {
+  char begin[] = {'\xFF', '\xFF', '\xFF', '\xFF'};
+  _fs->write(begin, 4);
+
+  _fs->write(_sync_marker, SYNC_MARKER_SIZE);
+}
+
+void SeqFileWriter::_write_seq_header() {
+  _fs->write(SEQ_HEADER, sizeof(SEQ_HEADER) - 1);
+  _fs->write(_sync_marker, SYNC_MARKER_SIZE);
+}
+
+int SeqFileWriter::write(const char *key,
+                         size_t key_len,
+                         const char *value,
+                         size_t value_len) {
+  if (key_len != sizeof(int64_t)) {
+    std::cout << "Key length not equal to " << sizeof(int64_t) << std::endl;
+    return -1;
+  }
+
+  uint32_t record_len = key_len + value_len;
+  uint32_t b_record_len = htonl(record_len);
+  uint32_t b_key_len = htonl((uint32_t)key_len);
+  // std::cout << "b_record_len " << b_record_len << " record_len " <<
+  // record_len << std::endl;
+  _fs->write((char *)&b_record_len, sizeof(uint32_t));
+  _fs->write((char *)&b_key_len, sizeof(uint32_t));
+  _fs->write(key, key_len);
+  _fs->write(value, value_len);
+  _bytes_to_prev_sync += record_len;
+
+  if (_bytes_to_prev_sync >= SYNC_INTERVAL) {
+    _write_sync_marker();
+    _bytes_to_prev_sync = 0;
+  }
+
+  return 0;
+}
+
+/* vim: set expandtab ts=4 sw=4 sts=4 tw=100: */

core/predictor/tools/seq_file.h

+// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// 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.
+
+#ifndef __SEQ_FILE_H_
+#define __SEQ_FILE_H_
+
+#include <fstream>
+
+const int SYNC_MARKER_SIZE = 16;
+const char SEQ_HEADER[] =
+    "SEQ\x06"
+    "\"org.apache.hadoop.io.BytesWritable\""
+    "org.apache.hadoop.io.BytesWritable"
+    "\x00\x00\x00\x00\x00\x00";
+const int SYNC_INTERVAL = 2000;
+
+class SeqFileWriter {
+ public:
+  SeqFileWriter(const char *file);
+  ~SeqFileWriter();
+
+ public:
+  int write(const char *key,
+            size_t key_len,
+            const char *value,
+            size_t value_len);
+
+ private:
+  void close();
+  void _write_sync_marker();
+  void _write_seq_header();
+
+ private:
+  char _sync_marker[SYNC_MARKER_SIZE];
+  int _bytes_to_prev_sync;
+  std::ofstream *_fs;
+};
+
+#endif  //__SEQ_FILE_H_
+
+/* vim: set expandtab ts=4 sw=4 sts=4 tw=100: */

core/predictor/tools/seq_generator.cpp

+// Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// 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.
+
+#include <fstream>
+#include <iostream>
+#include <memory>
+#include "core/predictor/framework.pb.h"
+#include "seq_file.h"
+using paddle::framework::proto::VarType;
+std::map<int, size_t> var_type_size;
+void reg_var_types() {
+  var_type_size[static_cast<int>(VarType::FP16)] = sizeof(int16_t);
+  var_type_size[static_cast<int>(VarType::FP32)] = sizeof(float);
+  var_type_size[static_cast<int>(VarType::FP64)] = sizeof(double);
+  var_type_size[static_cast<int>(VarType::INT32)] = sizeof(int);
+  var_type_size[static_cast<int>(VarType::INT64)] = sizeof(int64_t);
+  var_type_size[static_cast<int>(VarType::BOOL)] = sizeof(bool);
+  var_type_size[static_cast<int>(VarType::SIZE_T)] = sizeof(size_t);
+  var_type_size[static_cast<int>(VarType::INT16)] = sizeof(int16_t);
+  var_type_size[static_cast<int>(VarType::UINT8)] = sizeof(uint8_t);
+  var_type_size[static_cast<int>(VarType::INT8)] = sizeof(int8_t);
+}
+int dump_parameter(const char *input_file, const char *output_file) {
+  std::ifstream is(input_file);
+  // the 1st field, unit32_t version for LoDTensor
+  uint32_t version;
+  is.read(reinterpret_cast<char *>(&version), sizeof(version));
+  if (version != 0) {
+    std::cout << "Version number " << version << " not supported" << std::endl;
+    return -1;
+  }
+  // the 2st field, LoD information
+  uint64_t lod_level;
+  is.read(reinterpret_cast<char *>(&lod_level), sizeof(lod_level));
+  std::vector<std::vector<size_t>> lod;
+  lod.resize(lod_level);
+  for (uint64_t i = 0; i < lod_level; ++i) {
+    uint64_t size;
+    is.read(reinterpret_cast<char *>(&size), sizeof(size));
+    std::vector<size_t> tmp(size / sizeof(size_t));
+    is.read(reinterpret_cast<char *>(tmp.data()),
+            static_cast<std::streamsize>(size));
+    lod[i] = tmp;
+  }
+  // the 3st filed, Tensor
+  // Note: duplicate version field
+  is.read(reinterpret_cast<char *>(&version), sizeof(version));
+  if (version != 0) {
+    std::cout << "Version number " << version << " not supported" << std::endl;
+    return -1;
+  }
+  // int32_t size
+  // proto buffer
+  VarType::TensorDesc desc;
+  int32_t size;
+  is.read(reinterpret_cast<char *>(&size), sizeof(size));
+  std::unique_ptr<char[]> buf(new char[size]);
+  is.read(reinterpret_cast<char *>(buf.get()), size);
+  if (!desc.ParseFromArray(buf.get(), size)) {
+    std::cout << "Cannot parse tensor desc" << std::endl;
+    return -1;
+  }
+  // read tensor
+  std::vector<int64_t> dims;
+  dims.reserve(static_cast<size_t>(desc.dims().size()));
+  std::copy(desc.dims().begin(), desc.dims().end(), std::back_inserter(dims));
+  std::cout << "Dims:";
+  for (auto x : dims) {
+    std::cout << " " << x;
+  }
+  std::cout << std::endl;
+  if (dims.size() != 2) {
+    std::cout << "Parameter dims not 2D" << std::endl;
+    return -1;
+  }
+  size_t numel = 1;
+  for (auto x : dims) {
+    numel *= x;
+  }
+  size_t buf_size = numel * var_type_size[desc.data_type()];
+  char *tensor_buf = new char[buf_size];
+  is.read(static_cast<char *>(tensor_buf), buf_size);
+  is.close();
+  SeqFileWriter seq_file_writer(output_file);
+  int value_buf_len = var_type_size[desc.data_type()] * dims[1];
+  char *value_buf = new char[value_buf_len];
+  size_t offset = 0;
+  for (int64_t i = 0; i < dims[0]; ++i) {
+    // std::cout << "key_len " << key_len << " value_len " << value_buf_len <<
+    // std::endl;
+    memcpy(value_buf, tensor_buf + offset, value_buf_len);
+    seq_file_writer.write((char *)&i, sizeof(i), value_buf, value_buf_len);
+    offset += value_buf_len;
+  }
+  return 0;
+}
+int main(int argc, char **argv) {
+  if (argc != 3) {
+    std::cout << "Usage: seq_generator PARAMETER_FILE OUTPUT_FILE" << std::endl;
+    return -1;
+  }
+  reg_var_types();
+  dump_parameter(argv[1], argv[2]);
+}
+/* vim: set expandtab ts=4 sw=4 sts=4 tw=100: */

python/examples/criteo_ctr_with_cube/README.md

+## 带稀疏参数服务器的CTR预测服务
+
+### 获取样例数据
+```
+sh get_data.sh
+```
+
+### 保存模型和配置文件
+```
+python local_train.py
+```
+执行脚本后会在当前目录生成ctr_server_model和ctr_client_config文件夹,以及ctr_server_model_kv, ctr_client_conf_kv。
+
+### 启动稀疏参数服务器
+```
+cp ../../../build_server/core/predictor/seq_generator seq_generator
+cp ../../../build_server/output/bin/cube* ./cube/
+sh cube_prepare.sh &
+```
+
+### 启动RPC预测服务，服务端线程数为4（可在test_server.py配置）
+
+```
+python test_server.py ctr_serving_model_kv 
+```
+
+### 执行预测
+
+```
+python test_client.py ctr_client_conf/serving_client_conf.prototxt ./raw_data
+```
+
+### Benchmark
+
+设备 ：Intel(R) Xeon(R) CPU 6148 @ 2.40GHz 
+
+模型 ：[Criteo CTR](https://github.com/PaddlePaddle/Serving/blob/develop/python/examples/ctr_criteo_with_cube/network_conf.py)
+
+server core/thread num ： 4/8
+
+执行
+```
+bash benchmark.sh
+```
+客户端每个线程会发送1000个batch
+
+| client  thread num | prepro | client infer | op0    | op1   | op2    | postpro | avg_latency | qps   |
+| ------------------ | ------ | ------------ | ------ | ----- | ------ | ------- | ----- | ----- |
+| 1                  | 0.035  | 1.596        | 0.021  | 0.518 | 0.0024 | 0.0025  | 6.774 | 147.7 |
+| 2                  | 0.034  | 1.780        | 0.027  | 0.463 | 0.0020 | 0.0023  | 6.931 | 288.3 |
+| 4                  | 0.038  | 2.954        | 0.025  | 0.455 | 0.0019 | 0.0027  | 8.378 | 477.5 |
+| 8                  | 0.044  | 8.230        | 0.028  | 0.464 | 0.0023 | 0.0034  | 14.191 | 563.8 |
+| 16                 | 0.048  | 21.037       | 0.028  | 0.455 | 0.0025 | 0.0041  | 27.236 | 587.5 |
+
+平均每个线程耗时图如下
+
+![avg cost](../../../doc/criteo-cube-benchmark-avgcost.png)
+
+每个线程QPS耗时如下
+
+![qps](../../../doc/criteo-cube-benchmark-qps.png)

python/examples/criteo_ctr_with_cube/args.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+import argparse
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="PaddlePaddle CTR example")
+    parser.add_argument(
+        '--train_data_path',
+        type=str,
+        default='./data/raw/train.txt',
+        help="The path of training dataset")
+    parser.add_argument(
+        '--sparse_only',
+        type=bool,
+        default=False,
+        help="Whether we use sparse features only")
+    parser.add_argument(
+        '--test_data_path',
+        type=str,
+        default='./data/raw/valid.txt',
+        help="The path of testing dataset")
+    parser.add_argument(
+        '--batch_size',
+        type=int,
+        default=1000,
+        help="The size of mini-batch (default:1000)")
+    parser.add_argument(
+        '--embedding_size',
+        type=int,
+        default=10,
+        help="The size for embedding layer (default:10)")
+    parser.add_argument(
+        '--num_passes',
+        type=int,
+        default=10,
+        help="The number of passes to train (default: 10)")
+    parser.add_argument(
+        '--model_output_dir',
+        type=str,
+        default='models',
+        help='The path for model to store (default: models)')
+    parser.add_argument(
+        '--sparse_feature_dim',
+        type=int,
+        default=1000001,
+        help='sparse feature hashing space for index processing')
+    parser.add_argument(
+        '--is_local',
+        type=int,
+        default=1,
+        help='Local train or distributed train (default: 1)')
+    parser.add_argument(
+        '--cloud_train',
+        type=int,
+        default=0,
+        help='Local train or distributed train on paddlecloud (default: 0)')
+    parser.add_argument(
+        '--async_mode',
+        action='store_true',
+        default=False,
+        help='Whether start pserver in async mode to support ASGD')
+    parser.add_argument(
+        '--no_split_var',
+        action='store_true',
+        default=False,
+        help='Whether split variables into blocks when update_method is pserver')
+    parser.add_argument(
+        '--role',
+        type=str,
+        default='pserver',  # trainer or pserver
+        help='The path for model to store (default: models)')
+    parser.add_argument(
+        '--endpoints',
+        type=str,
+        default='127.0.0.1:6000',
+        help='The pserver endpoints, like: 127.0.0.1:6000,127.0.0.1:6001')
+    parser.add_argument(
+        '--current_endpoint',
+        type=str,
+        default='127.0.0.1:6000',
+        help='The path for model to store (default: 127.0.0.1:6000)')
+    parser.add_argument(
+        '--trainer_id',
+        type=int,
+        default=0,
+        help='The path for model to store (default: models)')
+    parser.add_argument(
+        '--trainers',
+        type=int,
+        default=1,
+        help='The num of trianers, (default: 1)')
+    return parser.parse_args()

python/examples/criteo_ctr_with_cube/benchmark.py

+# -*- coding: utf-8 -*-
+#
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+from paddle_serving_client import Client
+import sys
+import os
+import criteo as criteo
+import time
+from paddle_serving_client.utils import MultiThreadRunner
+from paddle_serving_client.utils import benchmark_args
+from paddle_serving_client.metric import auc
+
+args = benchmark_args()
+
+
+def single_func(idx, resource):
+    client = Client()
+    client.load_client_config('ctr_client_conf/serving_client_conf.prototxt')
+    client.connect(['127.0.0.1:9292'])
+    batch = 1
+    buf_size = 100
+    dataset = criteo.CriteoDataset()
+    dataset.setup(1000001)
+    test_filelists = [
+        "./raw_data/part-%d" % x for x in range(len(os.listdir("./raw_data")))
+    ]
+    reader = dataset.infer_reader(test_filelists[len(test_filelists) - 40:],
+                                  batch, buf_size)
+    args.batch_size = 1
+    if args.request == "rpc":
+        fetch = ["prob"]
+        print("Start Time")
+        start = time.time()
+        itr = 1000
+        for ei in range(itr):
+            if args.batch_size == 1:
+                data = reader().next()
+                feed_dict = {}
+                feed_dict['dense_input'] = data[0][0]
+                for i in range(1, 27):
+                    feed_dict["embedding_{}.tmp_0".format(i - 1)] = data[0][i]
+                result = client.predict(feed=feed_dict, fetch=fetch)
+            else:
+                print("unsupport batch size {}".format(args.batch_size))
+
+    elif args.request == "http":
+        raise ("Not support http service.")
+    end = time.time()
+    qps = itr / (end - start)
+    return [[end - start, qps]]
+
+
+if __name__ == '__main__':
+    multi_thread_runner = MultiThreadRunner()
+    endpoint_list = ["127.0.0.1:9292"]
+    #result = single_func(0, {"endpoint": endpoint_list})
+    result = multi_thread_runner.run(single_func, args.thread,
+                                     {"endpoint": endpoint_list})
+    avg_cost = 0
+    qps = 0
+    for i in range(args.thread):
+        avg_cost += result[0][i * 2 + 0]
+        qps += result[0][i * 2 + 1]
+    avg_cost = avg_cost / args.thread
+    print("average total cost {} s.".format(avg_cost))
+    print("qps {} ins/s".format(qps))

python/examples/criteo_ctr_with_cube/benchmark.sh

+rm profile_log
+batch_size=1
+for thread_num in 1 2 4 8 16
+do
+    $PYTHONROOT/bin/python benchmark.py --thread $thread_num --model ctr_client_conf/serving_client_conf.prototxt --request rpc > profile 2>&1
+    echo "========================================"
+    echo "batch size : $batch_size" >> profile_log
+    $PYTHONROOT/bin/python ../util/show_profile.py profile $thread_num >> profile_log
+    tail -n 2 profile >> profile_log
+done

python/examples/criteo_ctr_with_cube/benchmark_batch.py

+# -*- coding: utf-8 -*-
+#
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+from paddle_serving_client import Client
+import sys
+import os
+import criteo as criteo
+import time
+from paddle_serving_client.utils import MultiThreadRunner
+from paddle_serving_client.utils import benchmark_args
+from paddle_serving_client.metric import auc
+
+args = benchmark_args()
+
+
+def single_func(idx, resource):
+    client = Client()
+    print([resource["endpoint"][idx % len(resource["endpoint"])]])
+    client.load_client_config('ctr_client_conf/serving_client_conf.prototxt')
+    client.connect(['127.0.0.1:9292'])
+    batch = 1
+    buf_size = 100
+    dataset = criteo.CriteoDataset()
+    dataset.setup(1000001)
+    test_filelists = [
+        "./raw_data/part-%d" % x for x in range(len(os.listdir("./raw_data")))
+    ]
+    reader = dataset.infer_reader(test_filelists[len(test_filelists) - 40:],
+                                  batch, buf_size)
+    if args.request == "rpc":
+        fetch = ["prob"]
+        start = time.time()
+        itr = 1000
+        for ei in range(itr):
+            if args.batch_size > 1:
+                feed_batch = []
+                for bi in range(args.batch_size):
+                    data = reader().next()
+                    feed_dict = {}
+                    feed_dict['dense_input'] = data[0][0]
+                    for i in range(1, 27):
+                        feed_dict["embedding_{}.tmp_0".format(i - 1)] = data[0][
+                            i]
+                    feed_batch.append(feed_dict)
+                result = client.batch_predict(
+                    feed_batch=feed_batch, fetch=fetch)
+            else:
+                print("unsupport batch size {}".format(args.batch_size))
+
+    elif args.request == "http":
+        raise ("Not support http service.")
+    end = time.time()
+    qps = itr * args.batch_size / (end - start)
+    return [[end - start, qps]]
+
+
+if __name__ == '__main__':
+    multi_thread_runner = MultiThreadRunner()
+    endpoint_list = ["127.0.0.1:9292"]
+    #result = single_func(0, {"endpoint": endpoint_list})
+    result = multi_thread_runner.run(single_func, args.thread,
+                                     {"endpoint": endpoint_list})
+    print(result)
+    avg_cost = 0
+    qps = 0
+    for i in range(args.thread):
+        avg_cost += result[0][i * 2 + 0]
+        qps += result[0][i * 2 + 1]
+    avg_cost = avg_cost / args.thread
+    print("average total cost {} s.".format(avg_cost))
+    print("qps {} ins/s".format(qps))

python/examples/criteo_ctr_with_cube/benchmark_batch.sh

+rm profile_log
+for thread_num in 1 2 4 8 16
+do
+for batch_size in 1 2 4 8 16 32 64 128 256 512
+do
+    $PYTHONROOT/bin/python benchmark_batch.py --thread $thread_num --batch_size $batch_size --model serving_client_conf/serving_client_conf.prototxt --request rpc > profile 2>&1
+    echo "========================================"
+    echo "batch size : $batch_size" >> profile_log
+    $PYTHONROOT/bin/python ../util/show_profile.py profile $thread_num >> profile_log
+    tail -n 2 profile >> profile_log
+done
+done

python/examples/criteo_ctr_with_cube/clean.sh

+ps -ef | grep cube | awk {'print $2'} | xargs kill -9
+ps -ef | grep SimpleHTTPServer | awk {'print $2'} | xargs kill -9
+rm -rf cube/cube_data cube/data cube/log* cube/nohup* cube/output/ cube/donefile cube/input cube/monitor cube/cube-builder.INFO
+ps -ef | grep test | awk {'print $2'} | xargs kill -9
+ps -ef | grep serving | awk {'print $2'} | xargs kill -9

python/examples/criteo_ctr_with_cube/criteo.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+
+import sys
+
+
+class CriteoDataset(object):
+    def setup(self, sparse_feature_dim):
+        self.cont_min_ = [0, -3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
+        self.cont_max_ = [
+            20, 600, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50
+        ]
+        self.cont_diff_ = [
+            20, 603, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50
+        ]
+        self.hash_dim_ = sparse_feature_dim
+        # here, training data are lines with line_index < train_idx_
+        self.train_idx_ = 41256555
+        self.continuous_range_ = range(1, 14)
+        self.categorical_range_ = range(14, 40)
+
+    def _process_line(self, line):
+        features = line.rstrip('\n').split('\t')
+        dense_feature = []
+        sparse_feature = []
+        for idx in self.continuous_range_:
+            if features[idx] == '':
+                dense_feature.append(0.0)
+            else:
+                dense_feature.append((float(features[idx]) - self.cont_min_[idx - 1]) / \
+                                     self.cont_diff_[idx - 1])
+        for idx in self.categorical_range_:
+            sparse_feature.append(
+                [hash(str(idx) + features[idx]) % self.hash_dim_])
+
+        return dense_feature, sparse_feature, [int(features[0])]
+
+    def infer_reader(self, filelist, batch, buf_size):
+        def local_iter():
+            for fname in filelist:
+                with open(fname.strip(), "r") as fin:
+                    for line in fin:
+                        dense_feature, sparse_feature, label = self._process_line(
+                            line)
+                        #yield dense_feature, sparse_feature, label
+                        yield [dense_feature] + sparse_feature + [label]
+
+        import paddle
+        batch_iter = paddle.batch(
+            paddle.reader.shuffle(
+                local_iter, buf_size=buf_size),
+            batch_size=batch)
+        return batch_iter
+
+    def generate_sample(self, line):
+        def data_iter():
+            dense_feature, sparse_feature, label = self._process_line(line)
+            feature_name = ["dense_input"]
+            for idx in self.categorical_range_:
+                feature_name.append("C" + str(idx - 13))
+            feature_name.append("label")
+            yield zip(feature_name, [dense_feature] + sparse_feature + [label])
+
+        return data_iter
+
+
+if __name__ == "__main__":
+    criteo_dataset = CriteoDataset()
+    criteo_dataset.setup(int(sys.argv[1]))
+    criteo_dataset.run_from_stdin()

python/examples/criteo_ctr_with_cube/criteo_reader.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+import sys
+import paddle.fluid.incubate.data_generator as dg
+
+
+class CriteoDataset(dg.MultiSlotDataGenerator):
+    def setup(self, sparse_feature_dim):
+        self.cont_min_ = [0, -3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
+        self.cont_max_ = [
+            20, 600, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50
+        ]
+        self.cont_diff_ = [
+            20, 603, 100, 50, 64000, 500, 100, 50, 500, 10, 10, 10, 50
+        ]
+        self.hash_dim_ = sparse_feature_dim
+        # here, training data are lines with line_index < train_idx_
+        self.train_idx_ = 41256555
+        self.continuous_range_ = range(1, 14)
+        self.categorical_range_ = range(14, 40)
+
+    def _process_line(self, line):
+        features = line.rstrip('\n').split('\t')
+        dense_feature = []
+        sparse_feature = []
+        for idx in self.continuous_range_:
+            if features[idx] == '':
+                dense_feature.append(0.0)
+            else:
+                dense_feature.append((float(features[idx]) - self.cont_min_[idx - 1]) / \
+                                     self.cont_diff_[idx - 1])
+        for idx in self.categorical_range_:
+            sparse_feature.append(
+                [hash(str(idx) + features[idx]) % self.hash_dim_])
+
+        return dense_feature, sparse_feature, [int(features[0])]
+
+    def infer_reader(self, filelist, batch, buf_size):
+        def local_iter():
+            for fname in filelist:
+                with open(fname.strip(), "r") as fin:
+                    for line in fin:
+                        dense_feature, sparse_feature, label = self._process_line(
+                            line)
+                        #yield dense_feature, sparse_feature, label
+                        yield [dense_feature] + sparse_feature + [label]
+
+        import paddle
+        batch_iter = paddle.batch(
+            paddle.reader.shuffle(
+                local_iter, buf_size=buf_size),
+            batch_size=batch)
+        return batch_iter
+
+    def generate_sample(self, line):
+        def data_iter():
+            dense_feature, sparse_feature, label = self._process_line(line)
+            feature_name = ["dense_input"]
+            for idx in self.categorical_range_:
+                feature_name.append("C" + str(idx - 13))
+            feature_name.append("label")
+            yield zip(feature_name, [dense_feature] + sparse_feature + [label])
+
+        return data_iter
+
+
+if __name__ == "__main__":
+    criteo_dataset = CriteoDataset()
+    criteo_dataset.setup(int(sys.argv[1]))
+    criteo_dataset.run_from_stdin()

python/examples/criteo_ctr_with_cube/cube/conf/cube.conf

+[{
+    "dict_name": "test_dict",
+    "shard": 1,
+    "dup": 1,
+    "timeout": 200,
+    "retry": 3,
+    "backup_request": 100,
+    "type": "ipport_list",
+    "load_balancer": "rr",
+    "nodes": [{
+        "ipport_list": "list://127.0.0.1:8027"
+    }]
+}]

python/examples/criteo_ctr_with_cube/cube/conf/gflags.conf

+--port=8027
+--dict_split=1
+--in_mem=true
+--log_dir=./log/

python/examples/criteo_ctr_with_cube/cube/keys

+1
+2
+3
+4
+5
+6
+7
+8
+9
+10

python/examples/criteo_ctr_with_cube/cube_prepare.sh

+
+mkdir -p cube_model
+mkdir -p cube/data
+./seq_generator ctr_serving_model/SparseFeatFactors ./cube_model/feature
+./cube/cube-builder -dict_name=test_dict -job_mode=base -last_version=0 -cur_version=0 -depend_version=0 -input_path=./cube_model -output_path=./cube/data -shard_num=1  -only_build=false
+mv ./cube/data/0_0/test_dict_part0/* ./cube/data/
+cd cube && ./cube 

python/examples/criteo_ctr_with_cube/get_data.sh

+wget --no-check-certificate https://paddle-serving.bj.bcebos.com/data/ctr_prediction/ctr_data.tar.gz
+tar -zxvf ctr_data.tar.gz

python/examples/criteo_ctr_with_cube/local_train.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+from __future__ import print_function
+
+from args import parse_args
+import os
+import paddle.fluid as fluid
+import sys
+from network_conf import dnn_model
+
+dense_feature_dim = 13
+
+
+def train():
+    args = parse_args()
+    sparse_only = args.sparse_only
+    if not os.path.isdir(args.model_output_dir):
+        os.mkdir(args.model_output_dir)
+    dense_input = fluid.layers.data(
+        name="dense_input", shape=[dense_feature_dim], dtype='float32')
+    sparse_input_ids = [
+        fluid.layers.data(
+            name="C" + str(i), shape=[1], lod_level=1, dtype="int64")
+        for i in range(1, 27)
+    ]
+    label = fluid.layers.data(name='label', shape=[1], dtype='int64')
+
+    #nn_input = None if sparse_only else dense_input
+    nn_input = dense_input
+    predict_y, loss, auc_var, batch_auc_var, infer_vars = dnn_model(
+        nn_input, sparse_input_ids, label, args.embedding_size,
+        args.sparse_feature_dim)
+
+    optimizer = fluid.optimizer.SGD(learning_rate=1e-4)
+    optimizer.minimize(loss)
+
+    exe = fluid.Executor(fluid.CPUPlace())
+    exe.run(fluid.default_startup_program())
+    dataset = fluid.DatasetFactory().create_dataset("InMemoryDataset")
+    dataset.set_use_var([dense_input] + sparse_input_ids + [label])
+
+    python_executable = "python"
+    pipe_command = "{} criteo_reader.py {}".format(python_executable,
+                                                   args.sparse_feature_dim)
+
+    dataset.set_pipe_command(pipe_command)
+    dataset.set_batch_size(128)
+    thread_num = 10
+    dataset.set_thread(thread_num)
+
+    whole_filelist = [
+        "raw_data/part-%d" % x for x in range(len(os.listdir("raw_data")))
+    ]
+
+    print(whole_filelist)
+    dataset.set_filelist(whole_filelist[:100])
+    dataset.load_into_memory()
+    fluid.layers.Print(auc_var)
+    epochs = 1
+    for i in range(epochs):
+        exe.train_from_dataset(
+            program=fluid.default_main_program(), dataset=dataset, debug=True)
+        print("epoch {} finished".format(i))
+
+    import paddle_serving_client.io as server_io
+    feed_var_dict = {}
+    feed_var_dict['dense_input'] = dense_input
+    for i, sparse in enumerate(sparse_input_ids):
+        feed_var_dict["embedding_{}.tmp_0".format(i)] = sparse
+    fetch_var_dict = {"prob": predict_y}
+
+    feed_kv_dict = {}
+    feed_kv_dict['dense_input'] = dense_input
+    for i, emb in enumerate(infer_vars):
+        feed_kv_dict["embedding_{}.tmp_0".format(i)] = emb
+    fetch_var_dict = {"prob": predict_y}
+
+    server_io.save_model("ctr_serving_model", "ctr_client_conf", feed_var_dict,
+                         fetch_var_dict, fluid.default_main_program())
+
+    server_io.save_model("ctr_serving_model_kv", "ctr_client_conf_kv",
+                         feed_kv_dict, fetch_var_dict,
+                         fluid.default_main_program())
+
+
+if __name__ == '__main__':
+    train()

python/examples/criteo_ctr_with_cube/network_conf.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+import paddle.fluid as fluid
+import math
+
+
+def dnn_model(dense_input, sparse_inputs, label, embedding_size,
+              sparse_feature_dim):
+    def embedding_layer(input):
+        emb = fluid.layers.embedding(
+            input=input,
+            is_sparse=True,
+            is_distributed=False,
+            size=[sparse_feature_dim, embedding_size],
+            param_attr=fluid.ParamAttr(
+                name="SparseFeatFactors",
+                initializer=fluid.initializer.Uniform()))
+        x = fluid.layers.sequence_pool(input=emb, pool_type='sum')
+        return emb, x
+
+    def mlp_input_tensor(emb_sums, dense_tensor):
+        #if isinstance(dense_tensor, fluid.Variable):
+        #    return fluid.layers.concat(emb_sums, axis=1)
+        #else:
+        return fluid.layers.concat(emb_sums + [dense_tensor], axis=1)
+
+    def mlp(mlp_input):
+        fc1 = fluid.layers.fc(input=mlp_input,
+                              size=400,
+                              act='relu',
+                              param_attr=fluid.ParamAttr(
+                                  initializer=fluid.initializer.Normal(
+                                      scale=1 / math.sqrt(mlp_input.shape[1]))))
+        fc2 = fluid.layers.fc(input=fc1,
+                              size=400,
+                              act='relu',
+                              param_attr=fluid.ParamAttr(
+                                  initializer=fluid.initializer.Normal(
+                                      scale=1 / math.sqrt(fc1.shape[1]))))
+        fc3 = fluid.layers.fc(input=fc2,
+                              size=400,
+                              act='relu',
+                              param_attr=fluid.ParamAttr(
+                                  initializer=fluid.initializer.Normal(
+                                      scale=1 / math.sqrt(fc2.shape[1]))))
+        pre = fluid.layers.fc(input=fc3,
+                              size=2,
+                              act='softmax',
+                              param_attr=fluid.ParamAttr(
+                                  initializer=fluid.initializer.Normal(
+                                      scale=1 / math.sqrt(fc3.shape[1]))))
+        return pre
+
+    emb_pair_sums = list(map(embedding_layer, sparse_inputs))
+    emb_sums = [x[1] for x in emb_pair_sums]
+    infer_vars = [x[0] for x in emb_pair_sums]
+    mlp_in = mlp_input_tensor(emb_sums, dense_input)
+    predict = mlp(mlp_in)
+    cost = fluid.layers.cross_entropy(input=predict, label=label)
+    avg_cost = fluid.layers.reduce_sum(cost)
+    accuracy = fluid.layers.accuracy(input=predict, label=label)
+    auc_var, batch_auc_var, auc_states = \
+        fluid.layers.auc(input=predict, label=label, num_thresholds=2 ** 12, slide_steps=20)
+    return predict, avg_cost, auc_var, batch_auc_var, infer_vars

python/examples/criteo_ctr_with_cube/test_client.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+from paddle_serving_client import Client
+import sys
+import os
+import criteo as criteo
+import time
+from paddle_serving_client.metric import auc
+
+client = Client()
+client.load_client_config(sys.argv[1])
+client.connect(["127.0.0.1:9292"])
+
+batch = 1
+buf_size = 100
+dataset = criteo.CriteoDataset()
+dataset.setup(1000001)
+test_filelists = ["{}/part-0".format(sys.argv[2])]
+reader = dataset.infer_reader(test_filelists, batch, buf_size)
+label_list = []
+prob_list = []
+start = time.time()
+for ei in range(10000):
+    data = reader().next()
+    feed_dict = {}
+    feed_dict['dense_input'] = data[0][0]
+    for i in range(1, 27):
+        feed_dict["embedding_{}.tmp_0".format(i - 1)] = data[0][i]
+    fetch_map = client.predict(feed=feed_dict, fetch=["prob"])
+    prob_list.append(fetch_map['prob'][1])
+    label_list.append(data[0][-1][0])
+
+print(auc(label_list, prob_list))
+end = time.time()
+print(end - start)

python/examples/criteo_ctr_with_cube/test_server.py

+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# 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.
+# pylint: disable=doc-string-missing
+
+import os
+import sys
+from paddle_serving_server import OpMaker
+from paddle_serving_server import OpSeqMaker
+from paddle_serving_server import Server
+
+op_maker = OpMaker()
+read_op = op_maker.create('general_reader')
+general_dist_kv_infer_op = op_maker.create('general_dist_kv_infer')
+response_op = op_maker.create('general_response')
+
+op_seq_maker = OpSeqMaker()
+op_seq_maker.add_op(read_op)
+op_seq_maker.add_op(general_dist_kv_infer_op)
+op_seq_maker.add_op(response_op)
+
+server = Server()
+server.set_op_sequence(op_seq_maker.get_op_sequence())
+server.set_num_threads(4)
+server.load_model_config(sys.argv[1])
+server.prepare_server(workdir="work_dir1", port=9292, device="cpu")
+server.run_server()

python/paddle_serving_server/__init__.py

@@ -32,6 +32,7 @@ class OpMaker(object):
             "general_text_reader": "GeneralTextReaderOp",
             "general_text_response": "GeneralTextResponseOp",
             "general_single_kv": "GeneralSingleKVOp",
+            "general_dist_kv_infer": "GeneralDistKVInferOp",
             "general_dist_kv": "GeneralDistKVOp",
             "general_copy": "GeneralCopyOp"
         }
@@ -82,6 +83,7 @@ class Server(object):
         self.infer_service_fn = "infer_service.prototxt"
         self.model_toolkit_fn = "model_toolkit.prototxt"
         self.general_model_config_fn = "general_model.prototxt"
+        self.cube_config_fn = "cube.conf"
         self.workdir = ""
         self.max_concurrency = 0
         self.num_threads = 4
@@ -157,6 +159,11 @@ class Server(object):
                       "w") as fout:
                 fout.write(str(self.model_conf))
             self.resource_conf = server_sdk.ResourceConf()
+            for workflow in self.workflow_conf.workflows:
+                for node in workflow.nodes:
+                    if "dist_kv" in node.name:
+                        self.resource_conf.cube_config_path = workdir
+                        self.resource_conf.cube_config_file = self.cube_config_fn
             self.resource_conf.model_toolkit_path = workdir
             self.resource_conf.model_toolkit_file = self.model_toolkit_fn
             self.resource_conf.general_model_path = workdir
@@ -295,6 +302,6 @@ class Server(object):
                       self.workdir,
                       self.workflow_fn,
                       self.num_threads)
-        print("Going to Run Comand")
+        print("Going to Run Command")
         print(command)
         os.system(command)

tools/serving_build.sh

@@ -49,7 +49,7 @@ function build_server() {
                   -DPYTHON_LIBRARIES=$PYTHONROOT/lib64/libpython2.7.so \
                   -DPYTHON_EXECUTABLE=$PYTHONROOT/bin/python \
                   -DCLIENT_ONLY=OFF ..
-            check_cmd "make -j2 >/dev/null"
+            check_cmd "make -j2 >/dev/null && make install -j2 >/dev/null"
             pip install python/dist/paddle_serving_server* >/dev/null
             ;;
         GPU)
@@ -58,7 +58,7 @@ function build_server() {
                   -DPYTHON_EXECUTABLE=$PYTHONROOT/bin/python \
                   -DCLIENT_ONLY=OFF \
                   -DWITH_GPU=ON ..
-            check_cmd "make -j2 >/dev/null"
+            check_cmd "make -j2 >/dev/null && make install -j2 >/dev/null"
             pip install python/dist/paddle_serving_server* >/dev/null
             ;;
         *)
@@ -68,13 +68,13 @@ function build_server() {
     esac
     echo "build server $TYPE part finished as expected."
     cd ..
-    rm -rf $DIRNAME
 }
 
 function python_test_fit_a_line() {
     cd fit_a_line
     sh get_data.sh
     local TYPE=$1
+    echo $TYPE
     case $TYPE in
         CPU)
             # test rpc
@@ -102,12 +102,43 @@ function python_test_fit_a_line() {
     cd ..
 }
 
+function python_run_criteo_ctr_with_cube() {
+    local TYPE=$1
+    yum install -y bc >/dev/null
+    cd criteo_ctr_with_cube
+    check_cmd "wget https://paddle-serving.bj.bcebos.com/unittest/ctr_cube_unittest.tar.gz"
+    check_cmd "tar xf ctr_cube_unittest.tar.gz"
+    check_cmd "mv models/ctr_client_conf ./"
+    check_cmd "mv models/ctr_serving_model_kv ./"
+    check_cmd "mv models/data ./cube/"
+    check_cmd "mv models/ut_data ./"
+    cp ../../../build-server-$TYPE/output/bin/cube* ./cube/ 
+    mkdir -p $PYTHONROOT/lib/python2.7/site-packages/paddle_serving_server/serving-cpu-avx-openblas-0.1.3/
+    yes | cp ../../../build-server-$TYPE/output/demo/serving/bin/serving $PYTHONROOT/lib/python2.7/site-packages/paddle_serving_server/serving-cpu-avx-openblas-0.1.3/
+
+    sh cube_prepare.sh &
+    check_cmd "mkdir work_dir1 && cp cube/conf/cube.conf ./work_dir1/"    
+    python test_server.py ctr_serving_model_kv &
+    check_cmd "python test_client.py ctr_client_conf/serving_client_conf.prototxt ./ut_data >score"
+    AUC=$(tail -n 2  score | awk 'NR==1')
+    VAR2="0.70"
+    RES=$( echo "$AUC>$VAR2" | bc )
+    if [[ $RES -eq 0 ]]; then
+        echo "error with criteo_ctr_with_cube inference auc test, auc should > 0.70"
+        exit 1
+    fi
+    echo "criteo_ctr_with_cube inference auc test success"
+    ps -ef | grep "paddle_serving_server" | grep -v grep | awk '{print $2}' | xargs kill
+    ps -ef | grep "cube" | grep -v grep | awk '{print $2}' | xargs kill
+}
+
 function python_run_test() {
     cd python/examples
     local TYPE=$1
     # Frist time run, downloading PaddleServing components ...
     python -c "from paddle_serving_server import Server; server = Server(); server.download_bin()"
     python_test_fit_a_line $TYPE
+    python_run_criteo_ctr_with_cube $TYPE
     echo "test python $TYPE part finished as expected."
     cd ../..
 }
@@ -117,6 +148,7 @@ function main() {
     init
     build_client $TYPE
     build_server $TYPE
+    cd Serving/
     python_run_test $TYPE
     echo "serving $TYPE part finished as expected."
 }