You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

memory_api.cc 2.8 kB

5 years ago
5 years ago
5 years ago
5 years ago
5 years ago
5 years ago
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475
  1. /**
  2. * Copyright 2020 Huawei Technologies Co., Ltd
  3. *
  4. * Licensed under the Apache License, Version 2.0 (the "License");
  5. * you may not use this file except in compliance with the License.
  6. * You may obtain a copy of the License at
  7. *
  8. * http://www.apache.org/licenses/LICENSE-2.0
  9. *
  10. * Unless required by applicable law or agreed to in writing, software
  11. * distributed under the License is distributed on an "AS IS" BASIS,
  12. * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13. * See the License for the specific language governing permissions and
  14. * limitations under the License.
  15. */
  16. #include "framework/memory/memory_api.h"
  17. #include <memory>
  18. #include "common/ge/plugin_manager.h"
  19. #include "graph/manager/graph_mem_allocator.h"
  20. #include "graph/manager/host_mem_manager.h"
  21. #include "graph/manager/rdma_pool_allocator.h"
  22. #include "graph/utils/type_utils.h"
  23. #include "hccl/base.h"
  24. #include "hccl/hccl_types.h"
  25. namespace ge {
  26. Status InitRdmaPool(size_t size, rtMemType_t mem_type) {
  27. GELOGD("InitRdmaPool in");
  28. return MemManager::Instance().RdmaPoolInstance(mem_type).InitMemory(size);
  29. }
  30. Status RdmaRemoteRegister(const std::vector<HostVarInfo> &var_info, rtMemType_t mem_type) {
  31. GELOGD("Start to register rdma memory with host var size %zu", var_info.size());
  32. uint64_t device_base = 0;
  33. uint64_t device_size = 0;
  34. GE_CHK_STATUS_RET(MemManager::Instance().RdmaPoolInstance(mem_type).GetBaseAddr(device_base, device_size));
  35. return SUCCESS;
  36. }
  37. Status MallocSharedMemory(const TensorInfo &tensor_info, uint64_t &dev_addr, uint64_t &memory_size) {
  38. GELOGD("MallocSharedMemory in");
  39. uint32_t type_size = 0;
  40. bool result = TypeUtils::GetDataTypeLength(tensor_info.data_type, type_size);
  41. if (!result) {
  42. GELOGE(GRAPH_FAILED, "GetDataTypeLength failed, data_type=(%s).",
  43. TypeUtils::DataTypeToSerialString(tensor_info.data_type).c_str());
  44. return GRAPH_FAILED;
  45. }
  46. memory_size = type_size;
  47. for (auto dim : tensor_info.dims) {
  48. if (dim <= 0) {
  49. GELOGE(GRAPH_FAILED, "Tensor dims should be positive");
  50. return GRAPH_FAILED;
  51. }
  52. memory_size *= dim;
  53. }
  54. SharedMemInfo mem_info(tensor_info.var_name, memory_size);
  55. Status ret = HostMemManager::Instance().MallocSharedMemory(mem_info);
  56. if (ret != SUCCESS) {
  57. GELOGE(GRAPH_FAILED, "MallocSharedMemory failed op name [%s]", tensor_info.var_name.c_str());
  58. return GRAPH_FAILED;
  59. }
  60. dev_addr = reinterpret_cast<uint64_t>(reinterpret_cast<uintptr_t>(mem_info.device_address));
  61. GELOGD("MallocSharedMemory Succeeded");
  62. return SUCCESS;
  63. }
  64. Status GetVarBaseAddrAndSize(const string &var_name, uint64_t &base_addr, uint64_t &var_size) {
  65. GELOGD("GetVarBaseAddrAndSize in");
  66. return HostMemManager::Instance().QueryVarMemInfo(var_name, base_addr, var_size);
  67. }
  68. } // namespace ge

图引擎模块(GE)是MindSpore的一个子模块,其代码由C++实现,位于前端模块ME和底层硬件之间,起到承接作用。图引擎模块以ME下发的图作为输入,然后进行一系列的深度图优化操作,最后输出一张可以在底层硬件上高效运行的图。GE针对昇腾AI处理器的硬件结构特点,做了特定的优化工作,以此来充分发挥出昇腾AI处理器的强大算力。在进行模型训练/推理时,GE会被自动调用而用户并不感知。GE主要由GE API和GE Core两部分组成,详细的架构图如下所示