Efficient data sharing and compression extension across processing systems
By utilizing surface library technology and dedicated circuitry, efficient data sharing and compression model expansion across processing systems are achieved, solving the problems of inefficient data sharing and insufficient compression model expansion in existing technologies, and improving system performance and communication efficiency.
CN122415313APending Publication Date: 2026-07-17INTEL CORP
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INTEL CORP
- Filing Date
- 2018-04-23
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies are inefficient and cumbersome for data sharing across processing systems, and have failed to effectively achieve data compression and the expansion of compression models.
Method used
The use of surface library technology facilitates data sharing across processing systems and enables efficient data transfer and expansion of compression models by processing commands/instructions through dedicated circuit systems.
Benefits of technology
It improves the efficiency of data sharing and communication performance, supports a wider range of operations, including graphics data and machine learning operations, and enhances the overall performance of the processing system.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122415313A_ABST
Abstract
本申请公开了跨处理系统进行的高效数据共享和压缩扩展。描述了一种用于促进在自主机器处进行的数据共享和模型压缩扩展的机制。如本文所述的实施例的一种方法包括：检测第一处理器处理与第一计算装置处的神经网络相关的信息，其中，所述第一处理器包括第一图形处理器，并且所述第一计算装置包括第一自主机器。所述方法进一步包括：促进所述第一处理器将所述信息的一个或多个部分存储在数据库处的库中，其中，所述一个或多个部分可由计算装置的第二处理器访问。
Need to check novelty before this filing date? Find Prior Art