Scheduling method, scheduling device, electronic device, and storage medium

The scheduling method optimizes AI chip architectures by replicating valid data rows between computing units, addressing inefficiencies in data parallelism and model parallelism to enhance computing power utilization in multilayer neural networks.

JP7876064B2Active Publication Date: 2026-06-18BEIJING YOUZHUJU NETWORK TECH CO LTD
2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
BEIJING YOUZHUJU NETWORK TECH CO LTD
Filing Date
2023-09-18
Publication Date
2026-06-18

AI Technical Summary

⚠Technical Problem

Current AI chip architectures face challenges in efficiently processing dynamic computational graphs due to limitations in data parallelism and model parallelism, leading to repetitive data calculations and suboptimal utilization of computing power, especially in multilayer convolutional neural networks.

⚗Method used

A scheduling method that involves multiple computing units performing convolution operations, determining data replication transmission modes based on placement rules, and replicating valid data rows to optimize computing power utilization, reducing repetitive calculations.

🎯Benefits of technology

The method enhances computing power utilization by minimizing redundant operations and improving efficiency in processing multilayer convolutional neural networks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The present invention provides a scheduling method, a scheduling device, an electronic device, and a storage medium. The scheduling method includes: a plurality of computing units each performing a first convolution calculation on a corresponding plurality of data sets to obtain a corresponding plurality of first calculation result sets, the plurality of first calculation result sets being for constituting a first convolution layer obtained by the first convolution calculation; determining a data duplication and transmission mode corresponding to the plurality of first calculation result sets in the plurality of computing units according to an arrangement rule for the plurality of computing units of a second convolution layer obtained by the plurality of computing units performing a second convolution calculation on the first convolution layer; and obtaining a first intermediate data row required for padding in the second convolution calculation process by the first computing unit from the first calculation result set in the second computing unit based on the data duplication and transmission mode corresponding to the first computing unit among the plurality of computing units that is to pad a valid data row. The scheduling method can effectively reduce repeated data calculations and improve the utilization rate of chip computing capacity.
Need to check novelty before this filing date? Find Prior Art