hummingbird
/
graphengine
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
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.
1731 Commits
28 Branches
21 MB
0 Download
 
 
 
 
 
 
Tree: d5849832d3
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'd5849832d3'
${ noResults }
graphengine/ge/graph/manager
lianghao24@hisilicon.com eeac447869 warn check 3 years ago
..
model_manager update cmake 4 years ago
util hcomrecieve 3 years ago
block_memory.h update ge folder 4 years ago
graph_caching_allocator.cc Add statics log in graph caching allocator 3 years ago
graph_caching_allocator.h Add statics log in graph caching allocator 3 years ago
graph_context.cc update ge folder 4 years ago
graph_context.h update ge folder 4 years ago
graph_manager.cc warn check 3 years ago
graph_manager.h transfer to error_message context 3 years ago
graph_manager_utils.cc update cmake 4 years ago
graph_manager_utils.h Online infer support dynamic dims 3 years ago
graph_mem_allocator.cc rectify error codes 3 years ago
graph_mem_allocator.h fixed sc warning 3 years ago
graph_var_manager.cc Fix printf like format 3 years ago
graph_var_manager.h hide character 3 years ago
host_mem_allocator.cc GeTensor aligned addr & zero copy support 3 years ago
host_mem_allocator.h rm compile macro 3 years ago
host_mem_manager.cc rm compile macro 3 years ago
host_mem_manager.h rm compile macro 3 years ago
memory_api.cc change hccl inference 3 years ago
rdma_pool_allocator.cc rectify error codes 3 years ago
rdma_pool_allocator.h cache support 3 years ago
trans_var_data_utils.cc log rectify 3 years ago
trans_var_data_utils.h update cmake 4 years ago
Delete task

Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.

cancel
confirm

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

C++ Text Makefile Shell CMake other