Title:
MEMORY OPTIMIZATION METHOD AND APPARATUS USED FOR NEURAL NETWORK COMPILATION
Document Type and Number:
WIPO Patent Application WO/2024/065867
Kind Code:
A1
Abstract:
A memory optimization method and apparatus used for neural network compilation. The method comprises the following steps: step 1, compiling a neural network into a computing graph used for neural network computing; step 2, converting the computing graph into a topological graph; step 3, constructing an interval graph with respect to the life cycles of variables contained in the computing graph; and step 4, analyzing a relationship with respect to life cycles among tensor variables contained in computing graph nodes. A memory allocation optimization method for data streams in a computing graph generated by neural network compilation solves the problem of a deep learning operating system pre-allocating memories at the compilation stage to tensor variables flowing through nodes in the computing graph at runtime. An analysis method for a life cycle relationship among tensor variables contained in nodes of a computing graph; by means of analysis of the life cycle relationship of the tensor variables, an optimization method for allocating memories to the tensor variables contained in nodes of the computing graph is provided.
Inventors:
WANG HONGSHENG (CN)
CHEN GUANG (CN)
ZENG LINGFANG (CN)
CHEN GUANG (CN)
ZENG LINGFANG (CN)
Application Number:
PCT/CN2022/124003
Publication Date:
April 04, 2024
Filing Date:
October 09, 2022
Export Citation:
Assignee:
ZHEJIANG LAB (CN)
International Classes:
G06F9/50; G06N3/08
Foreign References:
| CN114186687A | 2022-03-15 | |||
| CN110597616A | 2019-12-20 | |||
| US10685295B1 | 2020-06-16 | |||
| CN111078395A | 2020-04-28 |
Attorney, Agent or Firm:
BEIJING ZHILIN HENGYUAN INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: