!3300 Optimize server code

Merge pull request !3300 from ZPaC/new-feature-of-ps

mindspore/ccsrc/backend/kernel_compiler/CMakeLists.txt

@@ -42,6 +42,7 @@ if (NOT (ENABLE_CPU AND (ENABLE_D OR ENABLE_GPU)))
     list(REMOVE_ITEM CPU_SRC_LIST "cpu/ps/push_kernel.cc")
     list(REMOVE_ITEM CPU_SRC_LIST "cpu/ps/sparse_apply_adam_ps_kernel.cc")
     list(REMOVE_ITEM CPU_SRC_LIST "cpu/ps/sparse_apply_ftrl_ps_kernel.cc")
+    list(REMOVE_ITEM CPU_SRC_LIST "cpu/ps/sparse_apply_lazy_adam_ps_kernel.cc")
 endif()
 
 if (ENABLE_GPU)

mindspore/ccsrc/backend/kernel_compiler/cpu/ps/sparse_apply_adam_ps_kernel.h

@@ -13,7 +13,7 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  */
-#ifndef MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_ADAM_CPU_KERNEL_H_
+#ifndef MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_ADAM_PS_KERNEL_H_
 #define MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_ADAM_PS_KERNEL_H_
 
 #include <vector>

mindspore/ccsrc/backend/kernel_compiler/cpu/ps/sparse_apply_lazy_adam_ps_kernel.cc

@@ -0,0 +1,104 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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 "backend/kernel_compiler/cpu/ps/sparse_apply_lazy_adam_ps_kernel.h"
+#include <memory>
+#include "backend/kernel_compiler/common_utils.h"
+#include "runtime/device/cpu/cpu_device_address.h"
+#include "frontend/parallel/ps/util.h"
+
+namespace mindspore {
+namespace kernel {
+namespace ps {
+void SparseApplyLazyAdamPSKernel::InitKernel(
+  const std::shared_ptr<std::vector<std::shared_ptr<std::vector<size_t>>>> &shapes) {
+  const std::vector<std::shared_ptr<std::vector<size_t>>> &shape_vec = *shapes;
+  std::vector<size_t> &var_shape = *(shape_vec[0]);
+  std::vector<size_t> &m_shape = *(shape_vec[1]);
+  std::vector<size_t> &v_shape = *(shape_vec[2]);
+  const std::vector<size_t> &grad_shape = *(shape_vec[9]);
+  const std::vector<size_t> &indices_shape = *(shape_vec[10]);
+
+  Shard(&var_shape, 0);
+  Shard(&m_shape, 0);
+  Shard(&v_shape, 0);
+
+  if (!IsSameShape(var_shape, m_shape)) {
+    MS_LOG(EXCEPTION) << "var and m should have the same shape";
+  }
+  if (!IsSameShape(var_shape, v_shape)) {
+    MS_LOG(EXCEPTION) << "var and v should have the same shape";
+  }
+  var_first_dim_size_ = var_shape[0];
+  for (size_t i = 1; i < var_shape.size(); ++i) {
+    if (var_shape[i] != grad_shape[i]) {
+      MS_LOG(EXCEPTION) << "The shape of var and grad must equal in dimension " << i;
+    }
+    var_outer_dim_size_ *= var_shape[i];
+  }
+  if (indices_shape.size() != 1) {
+    MS_LOG(EXCEPTION) << "indices must be 1D";
+  }
+  indices_size_ = indices_shape[0];
+  if (grad_shape[0] != indices_size_) {
+    MS_LOG(ERROR) << "The first dimension of grad shape must be equal to indices";
+  }
+  /*
+  if (AnfAlgo::HasNodeAttr(USE_NESTEROV, kernel_node)) {
+    use_nesterov_ = AnfAlgo::GetNodeAttr<bool>(kernel_node, "use_nesterov");
+  }
+  */
+  workspace_size_list_.emplace_back(indices_size_ * var_outer_dim_size_ * sizeof(float));
+  workspace_size_list_.emplace_back(indices_size_ * sizeof(int));
+  workspace_size_list_.emplace_back(indices_size_ * var_outer_dim_size_ * sizeof(float));
+  workspace_size_list_.emplace_back(indices_size_ * sizeof(int));
+  workspace_size_list_.emplace_back(var_first_dim_size_ * var_outer_dim_size_ * sizeof(float));
+}
+
+void SparseApplyLazyAdamPSKernel::ReInit(
+  const std::shared_ptr<std::vector<std::shared_ptr<std::vector<size_t>>>> &shapes) {
+  const std::vector<std::shared_ptr<std::vector<size_t>>> &shape_vec = *shapes;
+  const std::vector<size_t> &indices_shape = *(shape_vec[0]);
+  indices_size_ = indices_shape[0];
+  workspace_size_list_[0] = indices_size_ * var_outer_dim_size_ * sizeof(float);
+  workspace_size_list_[1] = indices_size_ * sizeof(int);
+}
+
+void SparseApplyLazyAdamPSKernel::ReInit(const std::vector<AddressPtr> &inputs) {
+  const auto &indices_addr = inputs[10];
+  indices_size_ = indices_addr->size / sizeof(int);
+  workspace_size_list_[0] = indices_size_ * var_outer_dim_size_ * sizeof(float);
+  workspace_size_list_[1] = indices_size_ * sizeof(int);
+}
+
+bool SparseApplyLazyAdamPSKernel::Execute(const std::vector<AddressPtr> &inputs,
+                                          const std::vector<AddressPtr> &workspace,
+                                          const std::vector<AddressPtr> &outputs) {
+  ReInit(inputs);
+  int *indices = reinterpret_cast<int *>(inputs[10]->addr);
+  for (size_t i = 0; i < inputs[10]->size / sizeof(int); i++) {
+    indices[i] -= rank_id_ * var_first_dim_size_;
+  }
+  return Launch(inputs, workspace, outputs);
+}
+
+const std::vector<size_t> &SparseApplyLazyAdamPSKernel::input_sizes() const { return GetInputSizeList(); }
+
+const std::vector<size_t> &SparseApplyLazyAdamPSKernel::output_sizes() const { return GetOutputSizeList(); }
+
+const std::vector<size_t> &SparseApplyLazyAdamPSKernel::workspace_sizes() const { return GetWorkspaceSizeList(); }
+}  // namespace ps
+}  // namespace kernel
+}  // namespace mindspore

mindspore/ccsrc/backend/kernel_compiler/cpu/ps/sparse_apply_lazy_adam_ps_kernel.h

@@ -0,0 +1,49 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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 MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_LAZY_ADAM_PS_KERNEL_H_
+#define MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_LAZY_ADAM_PS_KERNEL_H_
+
+#include <vector>
+#include <memory>
+#include "backend/kernel_compiler/cpu/ps/pserver_kernel.h"
+#include "backend/kernel_compiler/cpu/sparse_apply_lazy_adam_cpu_kernel.h"
+
+namespace mindspore {
+namespace kernel {
+namespace ps {
+using mindspore::kernel::SparseApplyLazyAdamCPUKernel;
+class SparseApplyLazyAdamPSKernel : public SparseApplyLazyAdamCPUKernel, public PServerKernel {
+ public:
+  SparseApplyLazyAdamPSKernel(size_t rank_id, size_t pserver_num) : PServerKernel(rank_id, pserver_num) {}
+  ~SparseApplyLazyAdamPSKernel() override = default;
+
+  void InitKernel(const std::shared_ptr<std::vector<std::shared_ptr<std::vector<size_t>>>> &) override;
+  void ReInit(const std::shared_ptr<std::vector<std::shared_ptr<std::vector<size_t>>>> &) override;
+  bool Execute(const std::vector<AddressPtr> &inputs, const std::vector<AddressPtr> &workspace,
+               const std::vector<AddressPtr> &outputs) override;
+
+  const std::vector<size_t> &input_sizes() const override;
+  const std::vector<size_t> &output_sizes() const override;
+  const std::vector<size_t> &workspace_sizes() const override;
+
+ protected:
+  void ReInit(const std::vector<AddressPtr> &) override;
+};
+}  // namespace ps
+}  // namespace kernel
+}  // namespace mindspore
+
+#endif  // MINDSPORE_CCSRC_BACKEND_KERNEL_COMPILER_CPU_SPARSE_APPLY_LAZY_ADAM_PS_KERNEL_H_

mindspore/ccsrc/backend/kernel_compiler/cpu/sparse_apply_lazy_adam_cpu_kernel.h

@@ -33,7 +33,7 @@ class SparseApplyLazyAdamCPUKernel : public CPUKernel {
   bool Launch(const std::vector<AddressPtr> &inputs, const std::vector<AddressPtr> &workspace,
               const std::vector<AddressPtr> &outputs) override;
 
- private:
+ protected:
   size_t indices_size_{0};
   size_t var_first_dim_size_{0};
   size_t var_outer_dim_size_{1};

mindspore/ccsrc/frontend/parallel/ps/common.h

@@ -63,6 +63,7 @@ constexpr int kInitWeightToOptimIdCmd = 11;
 constexpr int kInitOptimInputsShapeCmd = 12;
 constexpr int kInitEmbeddingsCmd = 20;
 constexpr int kEmbeddingLookupCmd = 30;
+constexpr int kFinalizeCmd = 40;
 
 constexpr size_t kInvalidKey = UINT64_MAX;
 

mindspore/ccsrc/frontend/parallel/ps/optimizer_info.cc

@@ -57,6 +57,16 @@ void DenseOptimInfo::Accumulate(const Values &values, const Lengths &lengths) {
   }
 }
 
+void DenseOptimInfo::ComputeMean(size_t n) {
+  if (n > 1) {
+    float *accum_grad_data = reinterpret_cast<float *>(gradient()->addr);
+    size_t size = gradient()->size / sizeof(float);
+    for (size_t i = 0; i < size; i++) {
+      accum_grad_data[i] /= n;
+    }
+  }
+}
+
 void DenseOptimInfo::Reset() { memset_s(gradient()->addr, gradient()->size, 0x00, gradient()->size); }
 
 void SparseOptimInfo::Accumulate(const Values &values, const Lengths &lengths) {

mindspore/ccsrc/frontend/parallel/ps/optimizer_info.h

@@ -33,6 +33,7 @@ class OptimizerInfo {
   virtual void Update(const Values &values, const Lengths &lengths) {}
   virtual void UpdateWeight(const WeightPtr &weight);
   virtual void Accumulate(const Values &values, const Lengths &lengths) = 0;
+  virtual void ComputeMean(size_t n) {}
   virtual void Reset() {}
   void AddWorkspace(const AddressPtr &workspace);
 
@@ -58,6 +59,7 @@ class DenseOptimInfo : public OptimizerInfo {
   ~DenseOptimInfo() override = default;
 
   void Accumulate(const Values &values, const Lengths &lens) override;
+  void ComputeMean(size_t n) override;
   void Reset() override;
 };
 

mindspore/ccsrc/frontend/parallel/ps/parameter_server.h

@@ -41,7 +41,7 @@
 #include "backend/kernel_compiler/kernel.h"
 #include "backend/kernel_compiler/cpu/cpu_kernel_factory.h"
 #include "backend/kernel_compiler/cpu/ps/pserver_kernel.h"
-#include "backend/kernel_compiler/cpu/ps/sparse_apply_adam_ps_kernel.h"
+#include "backend/kernel_compiler/cpu/ps/sparse_apply_lazy_adam_ps_kernel.h"
 #include "backend/kernel_compiler/cpu/ps/sparse_apply_ftrl_ps_kernel.h"
 #include "backend/kernel_compiler/cpu/ps/apply_momentum_ps_kernel.h"
 #include "backend/kernel_compiler/cpu/ps/embedding_look_up_ps_kernel.h"
@@ -90,6 +90,7 @@ class ParameterServer {
     void HandleInitInputsShape(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data, ::ps::KVPairs<T> *res);
     void HandleInitEmbeddings(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data, ::ps::KVPairs<T> *res);
     void HandleEmbeddingLookup(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data, ::ps::KVPairs<T> *res);
+    void HandleFinalize(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data, ::ps::KVPairs<T> *res);
 
     ParameterServer *ps_;
     typedef void (ServerHandler::*RequestHandler)(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data,
@@ -165,6 +166,7 @@ void ParameterServer<T>::ServerHandler::Init() {
   handlers_[kInitOptimInputsShapeCmd] = &ServerHandler::HandleInitInputsShape;
   handlers_[kInitEmbeddingsCmd] = &ServerHandler::HandleInitEmbeddings;
   handlers_[kEmbeddingLookupCmd] = &ServerHandler::HandleEmbeddingLookup;
+  handlers_[kFinalizeCmd] = &ServerHandler::HandleFinalize;
 }
 
 template <typename T>
@@ -256,16 +258,16 @@ void ParameterServer<T>::ServerHandler::HandleEmbeddingLookup(const ::ps::KVMeta
   ps_->DoEmbeddingLookup(key, req_data.keys.segment(1, req_data.keys.size()), res);
 }
 
+template <typename T>
+void ParameterServer<T>::ServerHandler::HandleFinalize(const ::ps::KVMeta &req_meta, const ::ps::KVPairs<T> &req_data,
+                                                       ::ps::KVPairs<T> *res) {
+  ::ps::Finalize(0, false);
+}
+
 template <typename T>
 bool ParameterServer<T>::Init(const FuncGraphPtr &func_graph) {
-  const char *server_num = getenv(kEnvPServerNum);
+  pserver_num_ = ::ps::NumServers();
-  const char *worker_num = getenv(kEnvWorkerNum);
+  worker_num_ = ::ps::NumWorkers();
-  if (server_num != nullptr) {
-    pserver_num_ = *server_num - '0';
-  }
-  if (worker_num != nullptr) {
-    worker_num_ = *worker_num - '0';
-  }
   func_graph_ = func_graph;
   rank_id_ = ::ps::MyRank();
   handler_.reset(new ServerHandler(this));
@@ -319,7 +321,7 @@ void ParameterServer<T>::InitOptimInputsShape(const Keys &keys, const Values &va
     if (optimizers_.count(key) == 0 && optim_inputs_shape_.count(key) > 0) {
       if (optim_name == kSparseAdam) {
         std::shared_ptr<PServerKernel> optimizer =
-          std::make_shared<kernel::ps::SparseApplyAdamPSKernel>(rank_id_, pserver_num_);
+          std::make_shared<kernel::ps::SparseApplyLazyAdamPSKernel>(rank_id_, pserver_num_);
         optimizer->InitKernel(optim_inputs_shape_[key]);
         optimizers_[key] = optimizer;
       } else if (optim_name == kApplyMomentum) {
@@ -368,10 +370,11 @@ void ParameterServer<T>::InitEmbeddingTable(
   }
 
   WeightPtr embedding = std::make_shared<Weight>(total_dims, 0);
+  T *embedding_data = embedding->data();
   std::default_random_engine engine;
   std::normal_distribution<float> random(0, 0.01);
   for (size_t i = 0; i < total_dims; i++) {
-    (*embedding)[i] = random(engine);
+    embedding_data[i] = random(engine);
   }
   weights_[key] = embedding;
 
@@ -402,6 +405,7 @@ void ParameterServer<T>::UpdateWeights() {
       const std::vector<kernel::AddressPtr> &workspaces = optim_info->workspaces();
       const std::vector<kernel::AddressPtr> &outputs = optim_info->outputs();
 
+      optim_info->ComputeMean(worker_num_);
       optimizer->Execute(inputs, workspaces, outputs);
       optim_info->Reset();
     }

mindspore/ccsrc/frontend/parallel/ps/worker.h

@@ -50,6 +50,7 @@ class Worker {
   void InitPSParamAndOptim(const std::string &param_name, void *param_data, size_t param_size);
   void DoPSEmbeddingLookup(const ::ps::SArray<::ps::Key> &keys, const ::ps::SArray<int> &lookup_ids,
                            const ::ps::SArray<int> &lens, ::ps::SArray<T> *lookup_result, int cmd);
+  void Finalize();
 
  private:
   Worker() : kv_worker_(nullptr), running_(false), key_cnt_(0) {}
@@ -118,6 +119,11 @@ void Worker<T>::DoPSEmbeddingLookup(const ::ps::SArray<::ps::Key> &keys, const :
   kv_worker_->EmbeddingLookup(keys, lookup_ids, lens, lookup_result, cmd);
 }
 
+template <typename T>
+void Worker<T>::Finalize() {
+  kv_worker_->Finalize();
+}
+
 template <typename T>
 void Worker<T>::InitPSParamData(const std::vector<size_t> &keys, void *origin_addr, size_t size) {
   ::ps::SArray<T> addr(reinterpret_cast<T *>(origin_addr), size / sizeof(T));

mindspore/ccsrc/frontend/parallel/ps/worker_proxy.h

@@ -58,6 +58,7 @@ class WorkerProxy : public ::ps::KVWorker<T> {
                          const ::ps::SArray<int> &lens = {}, const Callback &cb = nullptr, int priority = 0);
   void PushData(const ::ps::SArray<::ps::Key> &keys, const ::ps::SArray<T> &vals, const ::ps::SArray<int> &lens = {},
                 int cmd = 0, int priority = 0);
+  void Finalize();
 
  private:
   template <typename C>
@@ -146,6 +147,17 @@ void WorkerProxy<T>::PushData(const ::ps::SArray<::ps::Key> &keys, const ::ps::S
   obj_->WaitRequest(ts);
 }
 
+template <typename T>
+void WorkerProxy<T>::Finalize() {
+  int ts = obj_->NewRequest(::ps::kServerGroup);
+  ::ps::KVPairs<T> kvs;
+  kvs.keys.push_back(0);
+  kvs.vals.push_back(0.0f);
+  Send(obj_, ts, true, false, kFinalizeCmd, kvs, broadcast_slicer_);
+  obj_->WaitRequest(ts);
+  ::ps::Finalize(0, false);
+}
+
 template <typename T>
 template <typename C>
 int WorkerProxy<T>::AddLookupCB(const ::ps::SArray<::ps::Key> &keys, const ::ps::SArray<int> &lookup_ids,

tests/st/host_device/test_host_device_lenet.py

@@ -75,7 +75,7 @@ def train(net, data, label):
     print(res)
     print("+++++++++++++++++++++++++++")
     diff = res.asnumpy()[0] - 2.3025851
-    assert np.all(diff < 1.e-7)
+    assert np.all(diff < 1.e-6)
 
 
 @pytest.mark.level0