device_context.cc

/* Copyright (c) 2016 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 "paddle/fluid/platform/device_context.h"
#include <set>

#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
#include "paddle/fluid/memory/allocation/cuda_device_context_allocator.h"
#include "paddle/fluid/platform/cuda_device_guard.h"
#endif

#include "glog/logging.h"

namespace paddle {
namespace memory {

AllocationPtr Alloc(const platform::DeviceContext& dev_ctx, size_t size) {
  auto place = dev_ctx.GetPlace();
  if (size == 0) {
    return Alloc(place, size);
  }

  if (platform::is_gpu_place(place)) {
#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
    auto* default_dev_ctx = static_cast<platform::CUDADeviceContext*>(
        platform::DeviceContextPool::Instance().Get(place));
    auto& desired_dev_ctx =
        static_cast<const platform::CUDADeviceContext&>(dev_ctx);
    if (default_dev_ctx->stream() == desired_dev_ctx.stream()) {
      return Alloc(place, size);
    } else {
      return allocation::CUDADeviceContextAllocatorPool::Instance().Alloc(
          desired_dev_ctx, size);
    }
#else
    PADDLE_THROW(platform::errors::PermissionDenied(
        "Paddle can't use CUDA device since it's not compiled with CUDA,"
        "Please recompile or reinstall Paddle with GPU support."));
#endif
  } else if (platform::is_xpu_place(place)) {
#ifdef PADDLE_WITH_XPU
    // TODO(liuyuhui): Consider xpu stream later
    return Alloc(place, size);
#else
    PADDLE_THROW(platform::errors::PermissionDenied(
        "Paddle can't use XPU device since it's not compiled with XPU,"
        "Please recompile or reinstall Paddle with XPU support."));
#endif
  } else {
    return Alloc(place, size);
  }
}

}  // namespace memory
}  // namespace paddle

namespace paddle {
namespace platform {

#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
bool allow_tf32_cublas = true;
void SetAllowTF32Cublas(bool active) { allow_tf32_cublas = active; }
bool AllowTF32Cublas() { return allow_tf32_cublas; }

bool allow_tf32_cudnn = true;
void SetAllowTF32Cudnn(bool active) { allow_tf32_cudnn = active; }
bool AllowTF32Cudnn() { return allow_tf32_cudnn; }
#endif  // PADDLE_WITH_CUDA

DeviceContextPool* DeviceContextPool::pool = nullptr;

platform::DeviceContext* DeviceContextPool::Get(const platform::Place& place) {
  auto it = device_contexts_.find(place);
  if (it == device_contexts_.end()) {
    PADDLE_THROW(platform::errors::Unimplemented(
        "Place %s is not supported. Please check that your paddle compiles "
        "with WITH_GPU or WITH_XPU option or check that your train process "
        "hold the "
        "correct gpu_id if you use Executor.",
        place));
  }
  return it->second.get().get();
}

template <typename DevCtx, typename PlaceType>
inline void EmplaceDeviceContext(
    std::map<Place, std::shared_future<std::unique_ptr<DeviceContext>>>*
        map_ptr,
    platform::Place p) {
  using PtrType = std::unique_ptr<DeviceContext>;
  map_ptr->emplace(p, std::async(std::launch::deferred, [=] {
                     // lazy evaluation. i.e., only create device context at
                     // first `Get`
                     return PtrType(new DevCtx(BOOST_GET_CONST(PlaceType, p)));
                   }));
}

DeviceContextPool::DeviceContextPool(
    const std::vector<platform::Place>& places) {
  PADDLE_ENFORCE_GT(
      places.size(), 0,
      platform::errors::InvalidArgument("The number of platform places should "
                                        "be larger than 0. But received %d.",
                                        places.size()));
  std::set<Place> set;
  for (auto& p : places) {
    set.insert(p);
  }
  for (auto& p : set) {
    if (platform::is_cpu_place(p)) {
#ifdef PADDLE_WITH_MKLDNN
      EmplaceDeviceContext<MKLDNNDeviceContext, CPUPlace>(&device_contexts_, p);
#else
      EmplaceDeviceContext<CPUDeviceContext, CPUPlace>(&device_contexts_, p);
#endif
    } else if (platform::is_gpu_place(p)) {
#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
      EmplaceDeviceContext<CUDADeviceContext, CUDAPlace>(&device_contexts_, p);
#else
      PADDLE_THROW(
          platform::errors::Unimplemented("CUDAPlace is not supported. Please "
                                          "re-compile with WITH_GPU option."));
#endif
    } else if (platform::is_cuda_pinned_place(p)) {
#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
      EmplaceDeviceContext<CUDAPinnedDeviceContext, CUDAPinnedPlace>(
          &device_contexts_, p);
#else
      PADDLE_THROW(platform::errors::Unimplemented(
          "CUDAPlace is not supported. Please re-compile with WITH_GPU "
          "option."));
#endif
    } else if (platform::is_xpu_place(p)) {
#ifdef PADDLE_WITH_XPU
      EmplaceDeviceContext<XPUDeviceContext, XPUPlace>(&device_contexts_, p);
#else
      PADDLE_THROW(
          platform::errors::Unimplemented("XPUPlace is not supported. Please "
                                          "re-compile with WITH_XPU option."));
#endif
    }
  }
}

CPUDeviceContext::CPUDeviceContext() {
  eigen_device_.reset(new Eigen::DefaultDevice());
}

CPUDeviceContext::CPUDeviceContext(CPUPlace place) : place_(place) {
  eigen_device_.reset(new Eigen::DefaultDevice());
}

Eigen::DefaultDevice* CPUDeviceContext::eigen_device() const {
  return eigen_device_.get();
}

Place CPUDeviceContext::GetPlace() const { return place_; }

#ifdef PADDLE_WITH_XPU
XPUDeviceContext::XPUDeviceContext() { context_ = xpu::create_context(); }

XPUDeviceContext::~XPUDeviceContext() {}

XPUDeviceContext::XPUDeviceContext(XPUPlace place) : place_(place) {
  int dev_id = -1;
  int ret = xpu_current_device(&dev_id);
  PADDLE_ENFORCE_EQ(ret, XPU_SUCCESS,
                    platform::errors::External(
                        "XPU API return wrong value[%d], please check whether "
                        "Baidu Kunlun Card is properly installed.",
                        ret));
  ret = xpu_set_device(place.device);
  PADDLE_ENFORCE_EQ(ret, XPU_SUCCESS,
                    platform::errors::External(
                        "XPU API return wrong value[%d], please check whether "
                        "Baidu Kunlun Card is properly installed.",
                        ret));

  LOG_FIRST_N(WARNING, 1) << "Please NOTE: xpu device: " << place_.device;

  context_ = xpu::create_context();
  const int MAX_XPU_NUM = 16;
  const int l3_size = 13.5 * 1024 * 1024;
  static void* l3ptrs[MAX_XPU_NUM] = {nullptr};

  auto selected_xpus = GetXPUSelectedDevices();
  for (unsigned int i = 0; i < selected_xpus.size(); i++) {
    if (place.device == selected_xpus[i]) {
      if (l3ptrs[place.device] == nullptr) {
        xpu_malloc(static_cast<void**>(&l3ptrs[place.device]), l3_size,
                   XPU_MEM_L3);
      }
      if (l3ptrs[place.device] != nullptr) {
        context_->_l3_mgr.set(l3ptrs[place.device], l3_size);
        VLOG(3) << "xpu place " << place.device << " set l3 size " << l3_size;
      }
      break;
    }
  }

  ret = xpu_set_device(dev_id);
  PADDLE_ENFORCE_EQ(ret, XPU_SUCCESS,
                    platform::errors::External(
                        "XPU API return wrong value[%d], please check whether "
                        "Baidu Kunlun Card is properly installed.",
                        ret));
}

void XPUDeviceContext::Wait() const {
  int ret = xpu_set_device(place_.device);
  PADDLE_ENFORCE_EQ(ret, XPU_SUCCESS,
                    platform::errors::External(
                        "XPU API return wrong value[%d], please check whether "
                        "Baidu Kunlun Card is properly installed.",
                        ret));
  xpu_wait(context_->xpu_stream);
}

Place XPUDeviceContext::GetPlace() const { return place_; }

xpu::Context* XPUDeviceContext::x_context() const { return context_; }
#endif

#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)

class EigenCudaStreamDevice : public Eigen::StreamInterface {
 public:
  EigenCudaStreamDevice() : scratch_(nullptr), semaphore_(nullptr) {
    Eigen::initializeDeviceProp();
  }
  ~EigenCudaStreamDevice() override {}

  void Reinitialize(const gpuStream_t* cuda_stream, CUDAPlace place) {
    stream_ = cuda_stream;
    place_ = place;
    device_prop_ = &Eigen::m_deviceProperties[place.device];
  }

  const gpuStream_t& stream() const override { return *stream_; }

#ifdef PADDLE_WITH_HIP
  const hipDeviceProp_t& deviceProperties() const override {
#else
  const cudaDeviceProp& deviceProperties() const override {
#endif
    return *device_prop_;
  }

  void* allocate(size_t num_bytes) const override {
    if (UNLIKELY(num_bytes == 0)) {
      return nullptr;
    }
    auto buf = memory::Alloc(place_, num_bytes);
    VLOG(4) << "Eigen allocated at " << buf->ptr() << ", size" << buf->size()
            << " requested " << num_bytes;
    void* retv = buf->ptr();
    {
      std::lock_guard<std::mutex> lock(mtx_);
      allocations_.emplace(retv, std::move(buf));
    }
    return retv;
  }

  void deallocate(void* buffer) const override {
    if (LIKELY(buffer)) {
      std::lock_guard<std::mutex> lock(mtx_);
      allocations_.erase(buffer);
    }
  }

  void* scratchpad() const override {
    if (scratch_ == NULL) {
// windows use an old version of eigen that uses kCudaScratchSize,
// once windows updates eigen to a recent version, the following code
// can use kGpuScratchSize uniformly
#ifdef _WIN32
      scratch_ = allocate(Eigen::kCudaScratchSize + sizeof(unsigned int));
#else
      scratch_ = allocate(Eigen::kGpuScratchSize + sizeof(unsigned int));
#endif
    }
    return scratch_;
  }

  unsigned int* semaphore() const override {
    if (semaphore_ == NULL) {
#ifdef _WIN32
      char* scratch =
          static_cast<char*>(scratchpad()) + Eigen::kCudaScratchSize;
#else
      char* scratch = static_cast<char*>(scratchpad()) + Eigen::kGpuScratchSize;
#endif
      semaphore_ = reinterpret_cast<unsigned int*>(scratch);
#ifdef PADDLE_WITH_HIP
      PADDLE_ENFORCE_CUDA_SUCCESS(
          hipMemsetAsync(semaphore_, 0, sizeof(unsigned int), *stream_));
#else
      PADDLE_ENFORCE_CUDA_SUCCESS(
          cudaMemsetAsync(semaphore_, 0, sizeof(unsigned int), *stream_));
#endif
    }
    return semaphore_;
  }

 private:
  CUDAPlace place_;
  const gpuStream_t* stream_;  // not owned;
#ifdef PADDLE_WITH_HIP
  const hipDeviceProp_t* device_prop_;
#else
  const cudaDeviceProp* device_prop_;  // not owned;
#endif
  mutable void* scratch_;
  mutable unsigned int* semaphore_;
  mutable std::mutex mtx_;  // to protect allocations_
  mutable std::unordered_map<void*, memory::AllocationPtr> allocations_;
};

void CudnnWorkspaceHandle::ReallocWorkspace(size_t required_workspace_bytes) {
  if (required_workspace_bytes <= WorkspaceSize()) {
    return;
  }
  // reset allocation first before re-allocate to save memory
  allocation_.reset();
  allocation_ = memory::Alloc(device_context_, required_workspace_bytes);
}

thread_local std::unordered_map<const CUDADeviceContext*,
                                std::shared_ptr<CUDAContext>>
    CUDADeviceContext::thread_ctx_;
thread_local std::mutex CUDADeviceContext::ctx_mtx_;

void CUDAContext::InitEigenContext() {
  eigen_stream_.reset(new EigenCudaStreamDevice());
  eigen_stream_->Reinitialize(&RawStream(), place_);
  eigen_device_.reset(new Eigen::GpuDevice(eigen_stream_.get()));
}

CUDAContext::CUDAContext(const CUDAPlace& place,
                         const stream::Priority& priority) {
  place_ = place;
  CUDADeviceGuard guard(place_.device);
  stream_.reset(new stream::CUDAStream(place, priority));
  InitEigenContext();
  InitCuBlasContext();
  InitCuDNNContext();
#ifndef PADDLE_WITH_HIP
  InitCuSolverContext();
#endif
}

CUDAContext::~CUDAContext() {
  CUDADeviceGuard guard(place_.device);
  DestoryCuDNNContext();
  DestoryCuBlasContext();
#ifndef PADDLE_WITH_HIP
  DestoryCuSolverContext();
#endif
}

CUDADeviceContext::CUDADeviceContext(CUDAPlace place) : place_(place) {
  CUDADeviceGuard guard(place_.device);
  compute_capability_ = GetCUDAComputeCapability(place_.device);
  multi_process_ = GetCUDAMultiProcessors(place_.device);
  max_threads_per_mp_ = GetCUDAMaxThreadsPerMultiProcessor(place_.device);
  max_grid_dim_size_ = GetGpuMaxGridDimSize(place_.device);
  max_threads_per_block_ = GetCUDAMaxThreadsPerBlock(place_.device);

  driver_version_ = GetCUDADriverVersion(place_.device);
  runtime_version_ = GetCUDARuntimeVersion(place_.device);

  LOG_FIRST_N(WARNING, 1) << "Please NOTE: device: " << place_.device
                          << ", GPU Compute Capability: "
                          << compute_capability_ / 10 << "."
                          << compute_capability_ % 10
                          << ", Driver API Version: " << driver_version_ / 1000
                          << "." << (driver_version_ % 100) / 10
                          << ", Runtime API Version: "
                          << runtime_version_ / 1000 << "."
                          << (runtime_version_ % 100) / 10;
#ifdef PADDLE_WITH_HIP
  size_t version_major, version_minor, version_patch;
  PADDLE_ENFORCE_CUDA_SUCCESS(dynload::miopenGetVersion(
      &version_major, &version_minor, &version_patch));
  LOG_FIRST_N(WARNING, 1) << "device: " << place_.device
                          << ", MIOpen Version: " << version_major << "."
                          << version_minor << "." << version_patch;
#else
  size_t cudnn_dso_ver = dynload::cudnnGetVersion();
  LOG_FIRST_N(WARNING, 1) << "device: " << place_.device
                          << ", cuDNN Version: " << cudnn_dso_ver / 1000 << "."
                          << (cudnn_dso_ver % 1000) / 100 << ".";
#endif
  {
    // Check CUDA/CUDNN version compatiblity
    auto local_cuda_version =
        (driver_version_ / 1000) * 10 + (driver_version_ % 100) / 10;
#ifdef PADDLE_WITH_HIP
    auto compile_cuda_version = (HIP_VERSION / 100) * 10 + (HIP_VERSION % 10);
#else
    auto compile_cuda_version =
        (CUDA_VERSION / 1000) * 10 + (CUDA_VERSION % 100) / 10;
#endif
    if (local_cuda_version < compile_cuda_version) {
      LOG_FIRST_N(WARNING, 1)
          << "WARNING: device: " << place_.device
          << ". The installed Paddle is compiled with CUDA "
          << compile_cuda_version / 10 << "." << compile_cuda_version % 10
          << ", but CUDA runtime version in your machine is "
          << local_cuda_version / 10 << "." << local_cuda_version % 10
          << ", which may cause serious incompatible bug. "
          << "Please recompile or reinstall Paddle with compatible CUDA "
             "version.";
    }
  }
  default_ctx_.reset(new CUDAContext(place_));
}

CUDADeviceContext::~CUDADeviceContext() {
  SetDeviceId(place_.device);
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
  if (nccl_comm_) {
    PADDLE_ENFORCE_CUDA_SUCCESS(dynload::ncclCommDestroy(nccl_comm_));
  }
#endif
}

Place CUDADeviceContext::GetPlace() const { return place_; }

void CUDADeviceContext::Wait() const { context()->Stream()->Wait(); }

int CUDADeviceContext::GetComputeCapability() const {
  return compute_capability_;
}

int CUDADeviceContext::GetMaxPhysicalThreadCount() const {
  return multi_process_ * max_threads_per_mp_;
}

int CUDADeviceContext::GetSMCount() const { return multi_process_; }

int CUDADeviceContext::GetMaxThreadsPerBlock() const {
  return max_threads_per_block_;
}

Eigen::GpuDevice* CUDADeviceContext::eigen_device() const {
  return context()->EigenDevice().get();
}

bool CUDADeviceContext::tensor_core_available() const {
  return context()->CublasTensorCoreHandle() != nullptr;
}

dim3 CUDADeviceContext::GetCUDAMaxGridDimSize() const {
  return max_grid_dim_size_;
}

#ifdef PADDLE_WITH_HIP
miopenHandle_t CUDADeviceContext::cudnn_handle() const {
#else
cudnnHandle_t CUDADeviceContext::cudnn_handle() const {
#endif
  return context()->CudnnHandle();
}

CudnnWorkspaceHandle CUDADeviceContext::cudnn_workspace_handle() const {
  return CudnnWorkspaceHandle(*this, &cudnn_handle_mtx_);
}

#ifndef PADDLE_WITH_HIP
cusolverDnHandle_t CUDADeviceContext::cusolver_dn_handle() const {
  return context()->CusolverDnHandle();
}
#endif

gpuStream_t CUDADeviceContext::stream() const { return context()->RawStream(); }

CUDAPinnedDeviceContext::CUDAPinnedDeviceContext() {
  eigen_device_.reset(new Eigen::DefaultDevice());
}

CUDAPinnedDeviceContext::CUDAPinnedDeviceContext(CUDAPinnedPlace place)
    : place_(place) {
  eigen_device_.reset(new Eigen::DefaultDevice());
}

Eigen::DefaultDevice* CUDAPinnedDeviceContext::eigen_device() const {
  return eigen_device_.get();
}

Place CUDAPinnedDeviceContext::GetPlace() const { return place_; }
#endif

#ifdef PADDLE_WITH_MKLDNN
MKLDNNDeviceContext::MKLDNNDeviceContext(CPUPlace place)
    : CPUDeviceContext(place), p_blobmap_() {
  p_blobmap_.reset(new BlobMap());
  p_mutex_.reset(new std::mutex());
}

MKLDNNDeviceContextThreadLocals::Body::Body()
    : cur_engine(mkldnn::engine::kind::cpu, 0), cur_stream(cur_engine) {
  cur_mkldnn_session_id = kMKLDNNSessionID_Default;
  cur_input_shape_str = "";
  cur_input_shape_cache_capacity = 1;
  cur_paddle_data_layout = paddle::framework::DataLayout::kNCHW;
}

// When Thread finish we clear oneDNN cache
// This is needed when we have one executor used by many threads
// e.g. test_analyzer_detect. Thread ID is not part of caching key
// (for naive executor) so we need to clear cache when one thread finish
// and other is to start inference
// TODO(jczaja): Ideally it would be good to clear only part of cache
// related to thread that is to be terminated
MKLDNNDeviceContextThreadLocals::Body::~Body() {
  auto cpu_place = paddle::platform::CPUPlace();
  platform::DeviceContextPool& pool = platform::DeviceContextPool::Instance();
  platform::MKLDNNDeviceContext* dev_ctx =
      (platform::MKLDNNDeviceContext*)pool.Get(cpu_place);
  dev_ctx->ResetBlobMap();
}

void MKLDNNDeviceContextThreadLocals::Body::set_cur_mkldnn_session_id(
    size_t sid) {
  cur_mkldnn_session_id = sid;
}
size_t MKLDNNDeviceContextThreadLocals::Body::get_cur_mkldnn_session_id(void) {
  return cur_mkldnn_session_id;
}

void MKLDNNDeviceContextThreadLocals::Body::set_cur_input_shape_str(
    std::string input_shape_str) {
  cur_input_shape_str = input_shape_str;
}
void MKLDNNDeviceContextThreadLocals::Body::set_cur_input_shape_cache_capacity(
    int input_shape_cache_capacity) {
  cur_input_shape_cache_capacity = input_shape_cache_capacity;
}

void MKLDNNDeviceContextThreadLocals::Body::set_cur_paddle_data_layout(
    framework::DataLayout dl) {
  cur_paddle_data_layout = dl;
}

framework::DataLayout
MKLDNNDeviceContextThreadLocals::Body::get_cur_paddle_data_layout(void) {
  return cur_paddle_data_layout;
}

void MKLDNNDeviceContextThreadLocals::Body::log_lib_version(void) {
  if (!said_once) {
    said_once = true;
    auto dv = dnnl::version();
    LOG(INFO) << "oneDNN v" << dv->major << "." << dv->minor << "."
              << dv->patch;
  }
}

const mkldnn::engine& MKLDNNDeviceContextThreadLocals::Body::get_engine(void) {
  return cur_engine;
}

mkldnn::stream& MKLDNNDeviceContextThreadLocals::Body::get_stream(void) {
  return cur_stream;
}

void MKLDNNDeviceContext::ResetBlobMap() {
  std::lock_guard<decltype(*p_mutex_)> lock(*p_mutex_);
  if (!block_next_cache_clearing_) {
    VLOG(3) << "Clearing DNNL cache.";
    p_blobmap_->clear();
  } else {
    VLOG(3) << "Prevented Clearing DNNL cache.";
    block_next_cache_clearing_ = false;
  }
}

void MKLDNNDeviceContext::BlockNextCacheClearing() {
  std::lock_guard<decltype(*p_mutex_)> lock(*p_mutex_);
  VLOG(3) << "Next DNNL cache clearing has been blocked.";
  block_next_cache_clearing_ = true;
}

size_t MKLDNNDeviceContext::GetShapeBlobSize() const {
  std::lock_guard<decltype(*p_mutex_)> lock(*p_mutex_);
  BlobMap* pMap = p_blobmap_.get();
  auto map_it = pMap->find(tls().cur_mkldnn_session_id);
  if (map_it == pMap->end()) {
    PADDLE_THROW(platform::errors::NotFound(
        "MKLDNNDeviceContext don't find cur_mkldnn_session_id: %d.",
        tls().cur_mkldnn_session_id));
  }
  return map_it->second->size();
}

void MKLDNNDeviceContext::SetBlob(const std::string& name,
                                  BlobPtr_t<void> data) const {
  BlobMap* pMap = p_blobmap_.get();
  BlobPtr_t<ShapeBlob> sBlob = nullptr;
  BlobPtr_t<KeyBlob> pBlob = nullptr;

  int sid = tls().get_cur_mkldnn_session_id();

  std::lock_guard<decltype(*p_mutex_)> lock(*p_mutex_);

  // Find ShapeBlob for current mkldnn session id.
  auto map_it = pMap->find(sid);

  if (map_it == pMap->end()) {
    // 1st time to set blob in current thread
    sBlob = std::make_shared<ShapeBlob>();
    (*pMap)[sid] = sBlob;
    VLOG(2) << "SetBlob: sid=" << sid << ", add new sid\n";
  } else {
    sBlob = map_it->second;
  }

  // Find KeyBlob for current input shape
  auto key_it = sBlob->find(tls().cur_input_shape_str);

  if (key_it == sBlob->end()) {
    // In cache clearing mode, cur_input_shape_cache_capacity defines
    // max pblob capacity
    if ((static_cast<size_t>(sid) ==
         MKLDNNDeviceContextThreadLocals::kMKLDNNSessionID_CacheClearing) &&
        sBlob->size() &&
        (sBlob->size() >=
         static_cast<size_t>(tls().cur_input_shape_cache_capacity))) {
      VLOG(2) << "sid=" << sid
              << ", remove all blobs of shape: " << sBlob->begin()->first;
      sBlob->erase(sBlob->begin()->first);
    }
    pBlob = std::make_shared<KeyBlob>();
    (*sBlob)[tls().cur_input_shape_str] = pBlob;
  } else {
    pBlob = key_it->second;
  }

  // Find Blob via name
  auto blob_it = pBlob->find(name);
  if (blob_it == pBlob->end()) {
    (*pBlob)[name] = data;
  } else {
    blob_it->second = data;  // set data to existing blob
  }
  VLOG(2) << "SetBlob: sid=" << sid << ", add blob=" << name << "\n";
  // lock will be automatically released when out of scope
  return;
}

unsigned int MKLDNNDeviceContext::GetCachedObjectsNumber(void) {
  unsigned int num_entries = 0;
  for (auto const& l3 : *p_blobmap_) {
    for (auto const& l2 : *(l3.second)) {
      num_entries += (l2.second)->size();
    }
  }
  return num_entries;
}

MKLDNNDeviceContext::BlobPtr_t<void> MKLDNNDeviceContext::GetBlob(
    const std::string& name) const {
  BlobMap* pMap = p_blobmap_.get();
  BlobPtr_t<ShapeBlob> sBlob = nullptr;
  BlobPtr_t<KeyBlob> pBlob = nullptr;

  int sid = tls().get_cur_mkldnn_session_id();

  std::lock_guard<decltype(*p_mutex_)> lock(*p_mutex_);

  // Find ShapeBlob for current mkldnn session id firstly
  auto map_it = pMap->find(sid);
  if (map_it == pMap->end()) {
    VLOG(2) << "GetBlob: sid=" << sid << ", miss sid\n";
    return nullptr;
  }
  sBlob = map_it->second;

  // Find KeyBlob for current input shape secondly
  auto sBlob_it = sBlob->find(tls().cur_input_shape_str);
  if (sBlob_it == sBlob->end()) {
    VLOG(2) << "GetBlob: sid=" << tls().cur_input_shape_str
            << ", miss input_shape_str\n";
    return nullptr;
  }
  pBlob = sBlob_it->second;

  // Find Blob via name
  auto key_it = pBlob->find(name);

  if (key_it == pBlob->end()) {
    VLOG(2) << "GetBlob sid=" << sid << ", miss blob=" << name << "\n";
    return nullptr;
  }

  VLOG(2) << "GetBlob sid=" << sid << ", get blob=" << name << "\n";
  // lock will be automatically released when out of scope
  return key_it->second;
}

#endif

}  // namespace platform
}  // namespace paddle